• The strongest evidence was comparative: Nu’s Brazilian 90+ delinquency rates stayed flat to slightly better while peer cohorts deteriorated sharply, widening the gap across income bands.

• Primacy looks increasingly like the real moat: the primary-account relationship generates richer information and payment priority; Nu’s decisioning layer then converts that into better risk-adjusted economics.

• One quarter is still only evidence: RA-NIM reached 12.4% and credit economics improved materially, but the next two quarters need to show that Q2 was the start of a pattern rather than a favorable print.

Nu Holdings crossed an obvious milestone in 2Q26: quarterly net income exceeded $1 billion for the first time. The company now serves 139 million customers, ARPAC reached roughly $17, activity rose to 83.5%, and Brazil crossed 86% activity. Risk-adjusted NIM, the number that mattered most after 1Q26’s credit scare, jumped from 9.5% to 12.4%, while ROE returned to 33%.

Those numbers are excellent. They are also not the most important part of the quarter.

In our 1Q26 piece, the fundamental question was deliberately simple: “Is Nubank’s scale making its underwriting engine smarter, or merely making its credit book larger?” If more customers created more data, and more data created better decisions, then scale should improve the economics of the system. If scale merely enabled Nu to push more unsecured credit, the company could still be a very good bank, but it would remain a bank in the conventional sense: cyclical, capital-intensive, and ultimately constrained by credit losses.

2Q26 did not settle that debate. It did, however, produce the first evidence that lets us ask a better version of the question: does owning the primary customer relationship make Nubank’s economic advantage larger as the system scales, rather than gradually diluting it?

The Gap That Shouldn’t Be Widening

The key disclosure sits on slide 20. Nubank showed Brazilian credit-card 90+ delinquencies by income cohort alongside comparable Brazilian financial institutions. In the Mass Market cohort, Nu’s NPL rate moved from roughly 9.1% in July 2025 to 8.7% in May 2026. The S1-S2 peer group excluding Nu moved from 13.9% to 20.6%. The absolute gap therefore widened from 4.8 percentage points to 11.9 points. The same directional pattern appears in Super Core and High Income: Nu remained stable to improving while peers deteriorated.

The level is striking; the widening is more interesting. A persistent gap could simply mean Nu selected better borrowers in the first place. A widening gap suggests that explanation may be incomplete, although it does not eliminate selection: credit losses are nonlinear, and superior borrower selection can itself become more valuable as weaker customers cross default thresholds.

There are at least three overlapping mechanisms. The first is information. A primary institution sees salary, balances, spending and cash-flow deterioration as they happen rather than through a delayed bureau snapshot. The second is decisioning. Nu must convert those signals into better limits, pricing, collections and product decisions; having better information without using it well creates little value. The third is payment priority. A primary relationship may sit higher in a stressed customer’s hierarchy of obligations.

Management increasingly describes the advantage in exactly these terms. It says customers for whom Nu is the primary relationship have delinquency around half the portfolio average, and connects primacy to richer behavioral data, stronger underwriting and a higher position in the customer’s payment hierarchy. Vélez was even more explicit in Q&A:

“Over 60% of our mass market customers use us as their primary bank account… [that] positions us effectively as being senior in the credit stack of a customer.”

That does not prove which mechanism causes the gap. But slide 20 is the first strong comparative evidence that Nu’s primary-account system is producing different credit outcomes, not simply different growth rates.

Primacy Is the Asset; Intelligence Is the Multiplier

This changes how I think about the architecture we introduced in 3Q25. At the time, we described Nubank as three reinforcing layers: Trust & Distribution, Data, and Intelligence. NuFormer was the new “brain,” converting a proprietary dataset into better decisions.

That framework still works, but 2Q26 suggests a clearer hierarchy. Primacy is the scarce customer position. Information is what primacy generates. Decisioning is the capability that converts information into economic action. NuFormer is therefore an amplifier of the moat, not the moat by itself.

AI architectures will diffuse. Compute can be purchased. Competitors can hire engineers. What is much harder to purchase quickly is the position inside a customer’s financial life that generates years of proprietary behavioral context.

The flywheel is better expressed as:

Primacy → Information → Decisioning → Better Economics → Deeper Primacy.

Ecosystem breadth still matters, but as a complement rather than a separate thesis. In 4Q25 we argued that Nu’s architecture resembles proprietary aggregation: additional services such as NuCel, travel and commerce can create more interactions and feed information back into the financial decision engine. That remains plausible. The test for every adjacency, however, should now be stricter: does it deepen the primary relationship or materially improve the decisions made from it?

NuFormer makes this loop more powerful. Management says improvements to its common backbone can propagate across downstream applications; the latest generation quadrupled context length, training and inference speed, and can reach comparable predictive performance with 20 million fine-tuning rows where older approaches required more than 400 million. It now supports credit in Brazil and Mexico, unsecured lending, customer service and more than 100 marketing campaigns.

What management still does not disclose is equally important: we do not know the incremental approval rate, loss rate or risk-adjusted return of a NuFormer cohort versus an equivalent previous-model cohort. 2Q26 shows differentiated system outcomes. It does not yet isolate AI attribution.

Taking More Risk, Making More Money

That distinction matters because the most impressive economic result this quarter was not lower risk. Nu intentionally added risk.

Seasonality reduced the 15–90 day delinquency ratio by 37 basis points; deliberate expansion into higher-risk, higher-return cohorts added 24 basis points back. Those same decisions added approximately $170 million to the credit-loss allowance. Yet early delinquencies still improved to 4.8%, and risk-adjusted NIM reached 12.4%.

This is the progression we wanted to see after 4Q25. Then, Nu expanded credit availability and IFRS 9 forced expected losses to be recognized before the associated balances had generated their full revenue. 1Q26 absorbed much of the cost. 2Q26 began to show the other side: credit income contributed 178 basis points to sequential RA-NIM expansion and lower credit cost another 115 basis points.

The relevant objective of a lender is not to minimize defaults. It is to identify which expected defaults are worth accepting because the yield more than compensates for them. 2Q26 is the strongest evidence yet that Nu may be moving the efficient frontier outward.

The caveat is important. Desenrola helped cost of credit; seasonality helped; and Livingston explicitly refused to call 12% RA-NIM a floor. Management is effectively giving investors permission to contemplate a higher normalized regime while warning them not to annualize 12.4%. The next two quarters will tell us whether the durable number begins with 11 or falls back toward the old 9–10% range.

Mexico Becomes the Second Experiment

Mexico is now useful for a different reason. In 2Q25, the critical experiment was funding: Nu cut deposit yields and customers largely stayed. Trust had become an economic asset; Nu did not have to permanently buy deposits with the highest rate.

The full banking license now unlocks the next experiment: can trust become primacy?

Mexico already reaches 16.5% of adults, approximately the penetration Brazil had in 2020, but generates $12.3 of ARPAC versus Brazil’s $5.6 at the comparable stage. Management attributes that to higher income, stronger credit-card unit economics and larger earning balances. Mexico also has only a 35% loan-to-deposit ratio, leaving substantial room to convert funding into earning assets.

The banking license matters because payroll direct deposits and a fuller product set make it possible for Nu to become the account around which everyday financial activity organizes. Meanwhile, Mexico is standardizing payment interfaces in a way management explicitly compares with the regulatory conditions that helped Pix accelerate Brazilian digital banking.

But the bar has not moved. In 1Q26, we said Mexico becomes valuation-relevant when it moves from evidence of transferability to a material second earnings engine. 2Q26 does not clear that bar. What changed is that the mechanisms required to clear it are increasingly in place.

A Better Bank, or a Different Kind of Bank?

This is now the variant perception. The market does not need to be wrong about Nubank’s credit exposure for the stock to work. The base case only requires the market eventually to conclude that Nu is a premium financial compounder whose high ROE and superior post-loss economics are durable. A full platform classification is additional upside, not a prerequisite.

Using 2029 as the three-year earnings horizon, I would frame the outcomes as follows:

The bear case is not that Nubank fails. It is that higher-risk growth eventually behaves like higher-risk growth, RA-NIM compresses, Mexico fails to become material and Nu remains valued as a cyclical emerging-market bank. The base case requires no software multiple: roughly 20% revenue compounding, around 30% ROE and post-loss margins above 10% are enough. The bull case requires something structurally different, Mexico reproduces the Brazilian loop, primary-account density continues widening credit economics, and the decision engine scales those advantages across markets.

Evidence, Then Pattern

The next four quarters should be judged against hard signposts, not narrative. RA-NIM averaging above roughly 10.5–11% would strengthen the higher-regime argument; a fall below 9.5–10% without a strategic explanation would challenge it. Early NPL should remain around or below 5% outside normal 1Q seasonality, while 90+ NPL moving beyond roughly 7.5% alongside worsening early buckets would be a serious warning. Most importantly, the Nu-versus-peer cohort gap should hold or widen; compression because Nu deteriorates would attack the primacy thesis directly. Brazil ARPAC should move beyond $18 toward $19–20, while Mexico should show rising payroll/primary-account adoption, LDR and ARPAC without sacrificing funding discipline. Efficiency around 20–21% would demonstrate that these investments are still occurring inside a structurally low-cost model.

1Q26 ended with a useful discipline: one quarter can be explained; a pattern gets capitalized. 2Q26 supplies the first evidence that Nubank’s left tail may genuinely be narrower than that of the institutions from which it is taking share. More importantly, it offers a mechanism: primary relationships generate richer information and potentially greater payment priority; Nu’s decisioning infrastructure converts those advantages into profitable risk-taking; better economics allow Nu to deepen the relationship further.

That is a more demanding thesis than “Nubank has better AI,” and a more valuable one if it is true. 3Q26 and 4Q26 now have a simple job: show that 2Q26 was not merely a very good quarter, but the first visible consequence of a compounding primacy advantage.

Subscribe now

• Wallet share is rising, but payment yield is falling: Adyen processed 24% more volume while constant-currency revenue grew 21%, with take rate falling to 16.2 bps as larger merchants moved into lower pricing tiers.

• The thesis has shifted from transactions to relationships: the real opportunity is to deepen merchant economics through Unified Commerce, Platforms, loyalty, billing, issuing and money movement not simply process more volume.

• The missing proof is monetization: Adyen increasingly demonstrates that its platform creates merchant value but still does not clearly disclose how much incremental revenue it captures outside core processing.

“We are no longer just a payments company. We are the complete financial operating system for modern commerce.”

There is one number in Adyen’s first-half results that complicates that statement: 16.2 basis points.

Adyen processed €803.8 billion in H1, 24% more than a year ago, while net revenue increased 21% in constant currency. Digital revenue grew 15% against 17% volume growth, while even Platforms grew 40% against 42% volume. Adyen is winning more business, but the revenue attached to each euro of payment volume is declining as large merchants move into lower pricing tiers.

Six months ago we concluded that the old “architecture dividend” thesis needed an upgrade. Adyen’s single platform was still differentiated, but we had confused evidence that the platform creates value with evidence that Adyen captures that value. As we wrote after Q1, “The single platform cannot merely process more volume. It has to help Adyen earn more around that volume.”

H1 is the first report that makes clear management has reached essentially the same conclusion.

The fundamental question is therefore no longer whether Adyen has superior payment infrastructure. It is whether Adyen can increase the total economics of a merchant relationship faster than scale commoditizes the underlying payment.

That is a harder question. It is also a much more interesting one.

The Merchant, Not the Transaction

The most important disclosure in H1 was not revenue or EBITDA. It was Adyen’s newly disclosed customer maturity curve.

Roughly two-thirds of H1 growth came from merchants onboarded in 2024 or earlier. Typical wallet share rises from below 20% in years three through seven to above 40% after a decade, while the 300 merchants accounting for roughly 60% of growth represented more than 70% three years ago. That disclosure suggests a different way of measuring Adyen.

The traditional payment equation is straightforward: processed volume × take rate = revenue. The emerging Adyen equation is closer to merchant relationship × wallet share × workflows × products = economic value.

Consider how that might work. A merchant starts with online payments, adds more countries, then physical locations through Unified Commerce. Those transactions make Adyen’s understanding of the merchant and shopper richer; Dynamic Identification and Uplift use that information to improve authorization and fraud outcomes. Talon.One can add loyalty and incentive decisioning before the transaction, Orb can meter and bill usage, while accounts, issuing and Intelligent Money Movement extend Adyen’s role after the transaction.

The payment becomes the foundation rather than the entire product. There is, however, a crucial piece of information management did not provide.

Bryan Bergin asked where Adyen’s existing customer base sits on the newly disclosed maturity curve: how much volume already comes from merchants above 40% wallet share versus customers earlier in their penetration journey. Management talked about diversification and continued expansion but did not give the requested mix.

That matters. Adyen has shown investors the slope of the curve but not where the installed base sits on it. If most volume remains early in the curve, the disclosure is a powerful forward indicator. If much of the business is already mature, it is more descriptive of the past. The non-answer does not prove the latter, but neither should we treat the former as established.

The Paradox of Winning Wallet Share

There is another complication embedded in this maturity curve.

More wallet share is good for Adyen, but it also pushes large merchants through volume tiers that reduce payment pricing. Management is categorical that it sees no evidence of competitive pricing pressure in its data; Pieter van der Does said the take-rate movement reflects existing merchants giving Adyen more volume and therefore moving into lower tiers.

This means Adyen’s original flywheel has an opposing force inside it. More wallet share creates more payments, more data and deeper integration, but it can simultaneously reduce the economics of basic processing.

This offers one possible interpretation of EMEA’s slowdown. EMEA revenue grew 15% in H1, versus materially faster growth previously, while North America grew 30% constant currency. Management argues geographic attribution is inherently noisy because global merchants sell across regions and points to the payoff from sustained US investment.

A hypothesis worth testing is that EMEA also contains a relatively more mature set of Adyen relationships, making tiering more visible there. We do not have enough disclosure to establish that. But if it proves correct, the implications are important: Talon.One, Orb, money movement and financial products are not simply exciting incremental growth opportunities layered onto payments. They become a required offset to the natural economics of merchant maturity.

In other words, the question is not how much upside the operating-system strategy adds. It may increasingly be how much operating-system monetization Adyen needs to sustain the growth rate investors already expect.

Purity Was Never the Point

This also explains what otherwise looks like a striking reversal in strategy. The company we originally liked was defined partly by what it refused to do. While payment incumbents accumulated acquisitions and integration debt, Adyen built a single system from scratch. The single platform produced consistent data, lower complexity and faster deployment. Now Adyen has bought Talon.One and Orb.

I no longer think the right interpretation is that Adyen abandoned its architectural philosophy. Rather, management has revealed the deeper principle underneath it. Architectural purity was valuable because it gave Adyen greater control over merchant complexity; purity itself was never the objective.

Pieter made the distinction unusually explicit on the call: for the core, Adyen still prefers to build; for adjacent capabilities, management evaluates both internal development and acquisitions. It simply happened that the answer was “build” for nineteen years and “buy” twice in H1.

The optimization function has changed from build everything ourselves to control the merchant workflow without sacrificing the architecture.

That is sensible. It also creates a new test. Talon.One and Orb cannot merely become software products sold beside Adyen payments; they need to become native participants in the same data loop. Management says merchants already using both Talon.One and Adyen prove the combination works, and the intention is to deepen that integration.

If loyalty, billing, payments and money movement produce one shared intelligence layer, the acquisitions strengthen Adyen’s moat. If they remain separate applications connected through APIs, Adyen has merely broadened the catalogue while introducing the architectural complexity it once avoided.

The Missing Arrow

There is now considerable evidence for most of the Adyen flywheel. Unified Commerce grew 27% constant currency; Platforms grew 40%; Adyen has 486 merchants processing across channels at scale and 293,000 active Platform business customers. More importantly, Uplift increased merchant conversion by an average 0.9 percentage points by the end of H1.

That establishes several arrows: more transactions create richer data; richer data creates better decisions; better decisions create measurable merchant value; merchant relationships continue deepening over long periods.

One arrow remains conspicuously missing:

better merchant economics → materially more Adyen revenue outside processing.

In Q1 we said we wanted visible value-added-services revenue, continued >mid-20s Unified Commerce growth, >mid-30s Platforms growth, take-rate stabilization toward 16.5 basis points and a real Talon.One operating proof point. H1 passes the UC and Platforms tests, provides one useful Uplift outcome metric, and demonstrates management’s strategic commitment. It misses the take-rate test and largely defers the monetization test.

Even financial products, which management says should contribute roughly one percentage point of growth this year, remain largely unbroken out.

This is the central analytical tension: Adyen increasingly shows us that its platform creates value, but still does not show us clearly how much of that value Adyen captures.

The Narrative and the Numbers

This distinction also clarifies the valuation argument.

Bloomberg consensus already assumes a significant part of the operational bull case: approximately €3.49 billion of FY27 revenue, €4.19 billion in FY28, and an FY28 EBITDA margin around 55%. At the supplied €1,061.60 reference price, however, Adyen trades at roughly 22.5x FY27 and 18.2x FY28 adjusted earnings.

So consensus may already expect strong execution, but the stock does not obviously carry a full “commerce operating system” multiple. This creates two different bull cases. The first is that estimates are too low. The more interesting one is that estimates may be broadly right, but Adyen proves that their composition and durability deserve a different valuation category.

That gives us three futures:

The bear case does not require Adyen to become a bad business; it merely remains an excellent processor whose adjacent products never materially alter merchant economics. The base case assumes those products become meaningful enough to preserve high-teens growth and mid-50s margins. The bull case requires something qualitatively different: revenue per merchant begins compounding independently of payment yield, making processed volume an increasingly poor denominator for the company.

The evidence required from here is therefore quite specific. We need actual monetization or attach-rate disclosure from Talon.One, Orb, IMM or financial products; more information about where customers sit on the maturity curve; evidence that the acquired products become part of Adyen’s unified data architecture; and CapEx returning toward historical levels while the >55% 2028 margin objective survives the expansion.

In May, we asked whether Adyen could sell more than payments. H1 tells us that management has decided it must.

Adyen has shown us the wallet-share slope. It has shown that unified data can produce better merchant outcomes. It has bought capabilities before and after the payment that could expand the economics of the relationship.

What it has not shown is the number connecting those things: how much incremental revenue the expanded platform earns.

That is the missing arrow. If it appears, 16.2 basis points will increasingly look like the wrong denominator. If it does not, it may turn out to have been the most important number all along.

Subscribe now

• The thesis has moved beyond peak earnings: contracts now provide multi-year demand visibility, while Sandisk is deliberately keeping actual bit growth below technological productivity.

• The real compounding mechanism is per share: technology lowers capital needs, contracts stabilize cash flow, supply discipline limits reinvestment, and buybacks permanently shrink the denominator.

• The key risk is time: Sandisk does not need NAND cyclicality to vanish. It needs current economics to persist for three to four years long enough for buybacks to permanently reshape per-share earnings before the next downturn.

This is my fifth piece on Sandisk in nine months. Each previous piece revised the last as the business kept transforming faster than the framework could hold, and I am wary of doing that again on the strength of an Investor Day that has not yet been tested by a difficult quarter. What follows is not a re-modelled forecast. It extends the argument I made eight days ago in The Earnings That Stay, which was that the debate had moved from peak earnings to duration. Investor Day did not invalidate that thesis. It showed the machinery management believes sits between the contracts we had already identified and the shareholder value we were still unwilling to assume.

How Long Is Long Enough?

The bear argument on Sandisk is that eighty percent gross margins cannot persist because memory always mean-reverts. The bull argument is that management has guided to eighty percent, so the stock should trade at ten times three hundred dollars of earnings. Both arguments treat the level of earnings as the only thing that matters. Neither engages with the question that Investor Day actually posed.

The real question is not whether current economics are sustainable forever. The real question is whether they need to be. Suppose Sandisk remains highly profitable for three or four years, even if margins ultimately settle below management’s targets. During those years, contracts make cash generation predictable, technology supplies most of the required bit growth, capital intensity stays low, and management returns the surplus through repurchases. The next downturn would then strike a company with materially fewer shares outstanding. It could reduce corporate earnings. It could not recreate the shares already retired.

That is the consequence I underweighted after fiscal fourth quarter. The question is not merely how much of today’s earnings stay. It is what those earnings can permanently change while they are here. The mechanism has four moving parts and they reinforce each other. Technology continues to reduce the capital required to produce each bit. Contracts convert the demand side into forecastable cash flow. Supply discipline lets management hold actual bit growth below what the technology could deliver. And buybacks compound the resulting cash into a permanently smaller share count. Any one of these in isolation is corporate finance. The four together are how a cyclical business quietly restructures its per-share economics before the cycle has a chance to reset them.

Why Sandisk, Why Now

The strategic question worth asking is why Sandisk is the company attempting this and not one of the four larger memory manufacturers. Samsung has more scale, more capital, and more customer relationships. SK Hynix has better balance sheet flexibility and an integrated HBM business. Micron is larger and more diversified. Any of them could theoretically have moved first on multi-year contracts with financial guarantees. None of them did.

Sandisk moved first because Sandisk had to. The company came out of the Western Digital separation with a market capitalisation that was, at the low, less than the replacement cost of a single leading-edge NAND fab. In that context, management had two options: run the business as a smaller version of its competitors and wait for the cycle, or restructure the commercial model in a way that would force investors to reconsider what the company actually was. The first option would have produced a fine business and a decade of oscillation between forty dollars and two hundred. The second was strategic in the sense Michael Porter used the word: a deliberate decision to compete on a different dimension than the industry.

There may also be an irony in Sandisk’s manufacturing structure worth naming, though I would not build the thesis around it. Investors have historically treated the Kioxia joint venture as a constraint because it limits Sandisk’s ability to unilaterally redirect capacity. That same structural limitation may make long-duration customer commitments more credible, because Sandisk cannot easily walk away from a contract even if the math turns against it. This is inference rather than something management disclosed, and it would be one factor among several including financial guarantees, quarterly volume specifications, and strategic engagement. Kioxia’s own commercial interests could still disrupt the arrangement, which is why the argument sits at the level of a plausible contributor rather than a load-bearing claim. But it is worth flagging that the feature usually described as weakness may function differently in a contracted world.

Contracts Change Time

Sandisk’s New Business Models now cover eight customers, including three American hyperscalers. Weighted-average duration exceeds four years, total contract value at floor pricing is $93.9 billion, remaining performance obligations are $91.1 billion, financial guarantees total $16.5 billion, and coverage rises from roughly half of FY2027 bits to about two-thirds in FY2028. Two customers have already expanded their commitments within weeks of signing.

The previous article treated these agreements primarily as protection against falling prices. Investor Day showed their second function: they are an information system. Sandisk previously made decade-long investments while learning every quarter what customers wanted and what they would pay. It now receives product, volume, and timing commitments several years ahead. That visibility does not guarantee forecasts are correct. It makes disciplined supply possible in a way quarterly auctions never did. The contracts therefore do more than raise an earnings floor. They tell Sandisk which bits customers have committed to buy, and which production the company should not create merely because its technology allows it.

Who Captures the Productivity

Management says the BiCS technology roadmap can produce roughly twenty-seven percent annual bit-per-wafer productivity while Sandisk intends to grow actual volume only in the mid-to-high teens. The gap is not an idle technical detail. It determines whether innovation becomes future oversupply or retained economics.

The old memory model converted every cost reduction into more bits, lower prices, and value transferred to the customer. Goeckeler’s declaration at Investor Day that Sandisk’s cost reductions are “ours” was more consequential than it sounded. Management wants market pricing to determine revenue while technology reduces the wafers and capital required to serve committed demand. Sandisk cannot impose that discipline on Samsung, Micron, SK Hynix, or Kioxia. Industry supply can still break the economics. What Sandisk can control is whether its own productivity becomes additional capacity or free cash flow. Investor Day made clear which outcome management prefers.

What Happens to the Cash

Sandisk intends to return one hundred percent of excess cash after funding the business and maintaining a debt-free balance sheet, with repurchases as the current mechanism. This is usually presented as capital allocation. In a memory company, it is also supply discipline expressed through the balance sheet.

The traditional cycle contains its own destruction: high prices produce high cash flow, high cash flow funds more capacity, more capacity destroys prices. Sandisk is proposing to let technology provide most of the bit growth and send the residual cash to shareholders instead of recycling it into wafers. The buyback is not merely financial engineering. It is the mechanism through which the current regime becomes permanent even if the regime itself does not last.

It also makes the share count endogenous. If the market remains sceptical and the stock trades at a low multiple, each dollar buys back more of the company. If investors accept the duration argument and rerate the stock, repurchases become less accretive but shareholders receive the rerating immediately. The genuinely bad outcome is not that the multiple stays low. It is that the earnings disappear before the denominator changes.

The Category Question Follows

If this system works through even one full downturn, the market will eventually have to decide whether Sandisk still belongs in the peer group it currently sits in. A memory specialist asks where NAND pricing will be next quarter. A generalist can begin asking how much four years of contracted cash flow will reduce the share count. Those investors do not pay the same multiple because they are not underwriting the same object. But category change is the consequence of the mechanism working, not the thesis. It is what happens if the argument is right, not the argument itself.

Three Futures

The scenarios below use FY2030 earnings as the forward basis for a stock value roughly three years from now. They are estimates, not management guidance.

The bear case assumes customers renegotiate when spot prices fall, competitor supply returns, and buybacks lack the time and scale to alter the share count materially. The base case assumes management is too optimistic on margins but broadly right on duration: NAND remains cyclical, HBF contributes nothing material, and three to four years of high cash generation permanently improve per-share economics. The bull case requires NBM coverage to keep rising, industry supply discipline to persist, and management’s framework to prove broadly achievable. I would weight these 20 / 55 / 25, implying a rough central value near $2,650. The base case matters most because it does not require Sandisk to abolish the NAND cycle. It requires the current regime to last long enough.

What to Watch

NBM coverage must continue past two-thirds rather than stall. Customer expansions must continue rather than reverse. Contract economics must remain intact when spot pricing weakens. Sandisk must hold actual bit growth below technological productivity, keep capital intensity in the mid-single digits, and reduce the diluted share count rather than merely announce repurchase authorisations. Competitor capacity remains the external variable the company cannot contract away. HBF should be tracked separately as upside rather than requirement.

Long Enough

The previous article asked how much of Sandisk’s earnings would stay. Investor Day did not answer that question. It suggested we may have been asking it too literally. The earnings do not need to stay forever. They need to stay long enough for technology to reduce the capital required to produce them and for the cash they generate to reduce permanently the number of claims against them. The NAND cycle may return. The shares Sandisk bought back cannot.

Subscribe now

General Disclaimer: The information presented in this communication reflects the views of the author and does not necessarily represent the views of any other individual or organization. It is provided for informational purposes only and should not be construed as investment advice, a recommendation, an offer to sell, or a solicitation to buy any securities or financial products.

While the information is believed to be obtained from reliable sources, its accuracy, completeness, or timeliness cannot be guaranteed. No representation or warranty, express or implied, is made regarding the fairness or reliability of the information presented. Any opinions or estimates are subject to change without notice.

Past performance is not a reliable indicator of future performance. All investments carry risk, including the potential loss of principal. This communication does not consider the specific investment objectives, financial situation, or particular needs of any individual.

The author and any associated parties disclaim any liability for any direct or consequential loss arising from the use of this material and undertake no obligation to update or revise it.

• The thesis has broadened: Applied is benefiting not only because AI chips require more process complexity, but because customers are increasingly constrained by clean-room space, yield, uptime and time-to-ramp.

• Scarcity is showing up in pricing: Semiconductor Systems gross margin reached 55.4%, helped by value-based pricing for tools that save space, improve yield and accelerate production.

• The key question is durability: the market already believes this cycle is extraordinary. The next rerating depends on proving that higher content, pricing and service economics survive after today’s scarcity fades.

Applied Materials reported record revenue of $9.12 billion, record adjusted earnings of $3.50 per share and a 50.4% gross margin. It then guided the October quarter to $10.25 billion of revenue and $4.02 of earnings, with Semiconductor Systems alone expected to reach roughly $7.9 billion, up 62% year over year. The stock nevertheless fell after hours.

That contradiction is the right place to begin, because Applied no longer has to persuade investors that artificial intelligence is driving an extraordinary equipment cycle. The market already believes that. The harder question is whether Applied is becoming economically better because semiconductor manufacturing itself is becoming harder.

I have written about this progression three times. In January, I argued that physics was shifting value from drawing smaller features toward constructing increasingly complicated three-dimensional structures. In February, Gate-All-Around, HBM and a 54.5% Semiconductor Systems gross margin made that claim measurable. By May, management had raised its calendar 2026 Semiconductor Systems growth outlook from more than 20% to more than 30%, every test from the prior article had cleared, and the debate had moved from whether the complexity thesis worked to how much of it was already reflected in the price.

Q3 did not overturn that argument. It exposed the half we had not fully described. I had thought the opportunity was primarily that harder chips require more Applied equipment; this quarter makes clear that those chips are also being produced inside increasingly constrained factories. Complexity creates more content, while scarcity changes the value of that content.

Five of Six

The May article set six signposts for Q3. Five cleared.

The accountability matters. Revenue, margins, cash conversion, duration and growth all moved in the direction the thesis required, while working capital did not. The May article’s tests are set out in the prior piece; Q3 subsequently delivered $2.33 billion of free cash flow and record profitability across both Semiconductor Systems and AGS.

Accounts receivable rose from $5.19 billion at fiscal year-end to $7.69 billion in Q3, producing a rough DSO calculation of approximately 77 days. That is worse than the threshold we set, not better. The offset is real: operating cash flow rebounded to $3.04 billion, contract liabilities increased, and the quarter included a $1.2 billion sequential revenue step-up that may have been back-end loaded. Still, the proper conclusion is not that the cash-flow question has disappeared. The cash-flow test cleared; the working-capital test did not. A Q4 DSO below 70 would support the timing explanation, while another quarter above 75 would turn the issue into a genuine earnings-quality debate.

The Three Raises

The simplest way to understand the acceleration is to follow management’s own forecast.

In February, Applied expected calendar 2026 Semiconductor Systems revenue to grow more than 20%. In May, that became more than 30%. This quarter, CFO Brice Hill said the previous greater-than-30% expectation was now too low:

“The greater than 30% that we highlighted last quarter, we’re saying now that it’s greater than that at this point.”

Semiconductor Systems generated $5.97 billion in April, $7.04 billion in July and is guided to approximately $7.9 billion in October. Applied is taking its core equipment business from a roughly $6 billion quarterly run rate to almost $8 billion in six months.
The cause matters more than the magnitude. Customers repeatedly found ways to create clean-room capacity, pulled equipment deliveries forward and supplied Applied with detailed eight-quarter forecasts. Demand did not merely exceed Applied’s estimate; it repeatedly exceeded customers’ physical ability to house the equipment.

Two Kinds of Scarcity

Gary Dickerson described AI as creating two races: one for technology leadership and another for manufacturing capacity. The first requires better transistors, memory, wiring and packaging. The second requires customers to extract more yield and output from existing fabs while bringing new capacity online.

Applied sits at the intersection of both. Technological scarcity is the thesis we already know: leading-edge foundry and logic, DRAM and advanced packaging are expected to account for roughly 80% of WFE growth, while advanced-packaging revenue, previously expected to grow more than 50% this year, is now expected to grow more than 70%.

Physical scarcity is the new part. When fabs are full, the scarce resources are not only wafers and tools but clean-room area, yield, uptime and time-to-ramp. Applied is increasingly selling directly against each constraint. A new DRAM epitaxy system improves performance while using 20% less clean-room space. More than 37,000 chambers are connected to AIx software. AGS is expected to grow more than 20% this year, while process diagnostics and control is expected to grow more than 50%.

This distinction matters because complexity and scarcity play different roles. Complexity determines Applied’s opportunity through the cycle; scarcity determines how valuable that opportunity is at this point in the cycle. A tool that saves floor space is worth more when floor space is the bottleneck. A service that reduces downtime is worth more when every chamber is running. A metrology system that improves yield is worth more when additional fab capacity is years away.

Applied is therefore not simply selling more process steps. It is increasingly selling solutions to increasingly expensive bottlenecks.

What 55% Means

The strategic story matters only if Applied captures the economics.

Semiconductor Systems produced a 55.4% non-GAAP gross margin and a 38.0% operating margin, compared with 53.5% and 33.2% a year ago. AGS reached a 30.1% operating margin.

Management attributes part of the improvement to a value-based pricing system introduced three years ago. Applied now evaluates what a tool contributes to the customer’s economics rather than simply what it costs to build, and says pricing and margins are higher across both new and existing products.

Hill made the mechanism unusually explicit:

“We examine the value of every single tool and put a new price on every single tool.”

The customer is not merely buying a chamber. It is buying earlier time-to-market, more output per square metre, better yield and less downtime. That is why physical scarcity can convert technological differentiation into pricing power.

There is, however, a counter-signal. Q4 corporate gross margin is guided flat at 50.4% despite another large revenue increase. Management points to rapid hiring, manufacturing expansion, service additions and display mix. Those explanations are credible, but they create the next test. If the costs are temporary, corporate gross margin should begin following Semiconductor Systems higher. If company margin remains around 50% while revenue surges, Applied may be creating more value than it captures.

What Management Would Not Say

The most revealing part of the call was the gap between management’s confidence and its specificity.

UBS analyst Timothy Arcuri argued that the trajectory effectively required Semiconductor Systems growth near 40% if Applied was to outgrow the market. Hill replied:

“I think you’ve got the dynamics right.… We’re not giving that number because we’re not guiding that out quarter.”

Management did not endorse 40%, but it also did not reject the arithmetic.

The same pattern appeared around 2027. Dickerson called it another strong growth year; Hill called it another strong, record year. Yet neither would quantify it. When analysts tried to turn the plan to double manufacturing capacity by 2028 into a revenue forecast, Hill pushed back:

“It is capacity. So it is not a revenue forecast for 2028.”

The uncertainty has therefore moved. Management appears confident about demand direction but is less willing to commit on magnitude and timing, because physical clean-room availability still determines when customer intent becomes Applied revenue.

That specificity gap helps explain why an excellent guide did not make the stock work. At roughly thirty times forward earnings, directional confidence is no longer enough. The October investor event now matters because investors need a quantified bridge from today’s shortage to normalized 2027 and 2028 economics.

Three Worlds After Q3

The Q4 midpoint takes fiscal 2026 revenue to approximately $34.3 billion and adjusted EPS to roughly $12.75. The more important question is what remains after the current acceleration fades.

These are scenarios rather than point forecasts.

The Inventory Reckoning assumes that today’s shortage contains meaningful pull-forward. Applied exits the cycle with more content and a higher earnings base than before, but the familiar capital-equipment pattern reasserts itself as utilization falls, urgency fades and customers absorb capacity already ordered.

The Higher Floor is my base case. WFE eventually slows, but packaging, process control and AGS remain larger businesses; Applied captures modestly more content per wafer; and company margins rise after the current ramp costs fade. This thesis does not require the cycle to disappear. It requires the next trough to be materially higher than the last.

The Compounding Era demands more. Applied must consistently outgrow WFE, sustain mid-teens AGS growth, convert EPIC partnerships into earlier design-ins and move corporate margins toward the economics already visible inside Semiconductor Systems. Q3 makes that world possible, but the durability of the scarcity premium remains too untested to make it probable.

What the Stock Now Implies

Before the report, Bloomberg expected FY27 EPS of roughly $17.31. That estimate should rise, but I would not leap immediately to $20 or $21. The January quarter contains 14 weeks, and management deliberately refused to quantify the year. A move toward approximately $18–19-plus is a more disciplined starting point.

The interesting setup is that the earnings denominator is rising while the share-price numerator has fallen.

Below approximately $475, the market would be treating much of the current surge as cyclical despite evidence that the earnings floor has risen. Around $500–550, the stock prices substantial 2027 strength without requiring the full Higher Floor case. Between roughly $650 and $750, that base case becomes increasingly reflected. Above $1,000, investors are effectively accepting that the Compounding Era has arrived.

The first-half story was multiple expansion. From here, earnings revisions must do most of the work.

After the Cycle

The Big Fundamental Question has evolved:

Can Applied turn the growing physical and technological difficulty of producing AI chips into structurally more content, pricing and service revenue per unit of semiconductor capacity, such that complexity becomes the compounding engine rather than WFE growth being the thesis?

Q3 moved the answer toward yes. Packaging is growing more than 70%, process diagnostics more than 50%, AGS more than 20%, and Semiconductor Systems gross margin exceeds 55%. Customers are giving Applied longer forecasts, and the company is developing tools whose value comes from extracting more output from limited factory resources.

Three structural tests now matter. Corporate gross margin should move through approximately 51% during FY27. The October event should demonstrate that the 2026 exit rate is not mostly demand pulled forward from later years. Applied should continue to outgrow WFE after the present surge begins to normalize.

Three nearer-term observations will tell us whether those tests are progressing: Q4 Semiconductor Systems revenue near or above $7.9 billion, DSO below 70 days, and a Q1 FY27 outlook that remains strong after adjusting for the extra week.

The stock fell because investors already understand that this equipment cycle is extraordinary. Applied’s next task is harder: it must prove that some of today’s extraordinary economics remain after today’s scarcity fades.

The question is no longer how extraordinary this cycle becomes. It is how much of the extraordinary economics Applied keeps after the cycle is gone.

$AMAT

General Disclaimer: The information presented in this communication reflects the views of the author and does not necessarily represent the views of any other individual or organization. It is provided for informational purposes only and should not be construed as investment advice, a recommendation, an offer to sell, or a solicitation to buy any securities or financial products.

While the information is believed to be obtained from reliable sources, its accuracy, completeness, or timeliness cannot be guaranteed. No representation or warranty, express or implied, is made regarding the fairness or reliability of the information presented. Any opinions or estimates are subject to change without notice.

Past performance is not a reliable indicator of future performance. All investments carry risk, including the potential loss of principal. This communication does not consider the specific investment objectives, financial situation, or particular needs of any individual.

The author and any associated parties disclaim any liability for any direct or consequential loss arising from the use of this material and undertake no obligation to update or revise it.

Subscribe now

• The optical transition is broader than expected: demand is expanding simultaneously across scale-across, 1.6T, optical circuit switching, near-packaged optics, CPO and external laser sources.

• The economics are arriving faster than the architecture: revenue reached $1.01 billion, operating margin hit 36.6%, and management is already guiding to roughly 40% margins at a $1.25 billion quarterly revenue run-rate.

• The real debate is duration: customers are helping fund capacity and signing multi-year commitments, but the thesis depends on Lumentum retaining its yield, reliability and qualification advantage as Chinese and other competitors’ scale.

Fourteen months ago, I wrote that AI had changed the nature of networking. Traditional data centers were built for computers that communicated intermittently: an application requested data, a browser called a server, a database returned a result. AI training was different. Thousands of processors needed to communicate continuously, and a processor waiting on the network was an extraordinarily expensive piece of silicon doing nothing. The network had stopped being plumbing around the computer and had become one of the constraints on the computer itself.

The conclusion of that article was that the optical transition would eventually become one of the most important battlegrounds in AI networking. Copper could be optimized, retimed and packaged more cleverly, but physics would eventually win ever-higher bandwidth would make electrical connections increasingly expensive in power and increasingly difficult in signal integrity.

I think that argument was broadly right. What I underestimated was the shape of the transition.

I imagined something relatively discrete: copper would dominate until it reached its limits, at which point technologies such as co-packaged optics would move light closer to the processor. Lumentum’s fiscal fourth-quarter earnings suggest the transition is happening much more continuously. Optics is spreading through several layers of the AI network at once, and each layer appears to create another demand vector before the previous one has finished growing.

The most striking evidence came from CEO Michael Hurlston:

“For one major hyperscaler, the network capacity connecting just two AI data center sites is double the total global backbone capacity they built over the entirety of the last decade.”

Two sites require twice a decade’s worth of backbone capacity.

That sentence says more about Lumentum’s opportunity than its first billion-dollar quarter does. Revenue reached $1.01 billion, up 109% year-over-year, while non-GAAP gross margin crossed 50% and operating margin reached 36.6%. The company then guided the September quarter to $1.225–1.275 billion of revenue and approximately 40% operating margins.

The numbers are extraordinary. The more important question is whether they describe the top of an unusually violent optical shortage or the early financial expression of something more structural.

The Missing Layer

My previous networking article focused mainly on the companies deciding how AI networks were built: NVIDIA, Broadcom, Arista, Marvell and Credo. NVIDIA had understood earlier than most that a GPU was only as useful as the network connecting it to other GPUs; Mellanox therefore looked less like diversification and more like completing the machine. Meanwhile, the optical transition appeared as a future disruption to today’s hierarchy.

What that framing underweighted was the physical layer beneath those architectural choices: the lasers.

Lumentum’s pump lasers are a good example. Shipments increased more than 80% year-over-year, yet the company remains effectively sold out. Management expects shipment volumes to increase roughly fourfold over the coming quarters and estimates its market share at 70–80%. More unusually, customers are entering mostly three-year arrangements, many effectively take-or-pay, while helping offset the capital expenditure required to create additional capacity.

That is an odd form of customer behavior if this is simply a normal semiconductor shortage waiting to correct itself. Customers usually welcome new capacity because it creates negotiating leverage. Here they are helping finance that capacity because securing enough supply appears more valuable than preserving future optionality.

EMLs tell a similar story. Lumentum expects more than 50% year-over-year unit growth by the December quarter, yet management still expects to remain significantly behind customer demand. Two-hundred-gigabit-per-lane EMLs already represent more than a quarter of EML revenue and should become most volumes around the middle of 2027.

The strongest evidence, though, came from ultra-high-power lasers. Citi asked management whether the supply-demand imbalance had improved from the previous quarter. Hurlston replied:

“We are way behind in our shipments, unfortunately, on high-powered lasers... The demand signal has increased, and we are very much further behind relative to our ability to supply.”

Wupen Yuen immediately clarified the important part: Lumentum’s capacity ramp itself is on schedule. The company is further behind because demand accelerated faster than supply.

This is the central fact in the Lumentum thesis today. Capacity is expanding rapidly, but the demand curve is currently moving away even faster.

The Optical Boundary

The reason that can happen becomes clearer when Lumentum’s products are considered as parts of one architectural progression rather than separate businesses.

The earliest AI networking opportunity was largely scale-across. Vast AI campuses need enormous amounts of high-capacity connectivity between buildings and data centers. That is where pumps and narrow-linewidth lasers matter, and where the “two sites versus a decade of backbone” example is most revealing.

Then comes scale-out. The cluster itself becomes larger, pushing 800G toward 1.6T and eventually 3.2T. Lumentum began 1.6T production in Q4, but the majority of transceiver shipments were still 800G. Management expects 1.6T uptake to intensify beginning this quarter and sustain through calendar 2027. In other words, one of the most obvious growth drivers in the story has barely started.

Then there is switching. Optical circuit switches allow increasingly large clusters to change connectivity dynamically without converting optical signals back into electrical ones. Lumentum doubled OCS shipments sequentially in Q4 and says the current quarter will be its first with OCS revenue meaningfully above $100 million. Management remains on track for roughly $400 million of OCS revenue in the second half of calendar 2026.

The more interesting OCS development is competitive rather than financial. Hurlston described Lumentum as effectively the only scaled merchant supplier currently shipping at meaningful volume. He expects the company to become the largest supplier to one major customer in early 2027, while new higher-port-count, lower-port-count and in-tray designs expand an opportunity that was not fully contemplated in the previous market framework.

Finally comes scale-up, where the optical boundary moves inside the rack and closer to the accelerator itself.

This is where my 2025 framework was too binary. I treated co-packaged optics as the transition point: copper today, optics tomorrow. The industry appears to be developing an intermediate step in near-packaged optics, placing optical engines close to the accelerator before eventually moving them onto the package.

Lumentum benefits from both.

Hurlston addressed recent concern about CPO timing directly:

“Our lead CPO customer’s production plans remain very much on track, and their demand signal has increased since our last update.”

That would already have been an important answer. The next sentence made it more tangible:

“We were recently given our first external light source, or ELS, module purchase for delivery by the second half of calendar 2027.”

The word purchase matters. CPO has spent years existing in roadmaps, demonstrations and future market estimates. A purchase order does not prove the eventual size of the market, but it moves part of the opportunity from possibility to commercial evidence.

ELS also raises the amount of value Lumentum can capture. Instead of selling only the laser, Lumentum supplies a higher-ASP module to customers that do not want to integrate individual laser chips themselves. Management says margins will be lower than on the standalone laser but still above the corporate average.

The optical transition therefore looks less like one architectural event and more like a moving boundary: between data centers, between racks, inside racks, beside accelerators and eventually on the package itself.

Lumentum has exposure at nearly every step.

The Economics Arrived Early

That breadth would matter less if moving from components into systems destroyed the economics.

So far, the opposite has happened.

Revenue has progressed from roughly $666 million to $808 million to $1.01 billion and now to a $1.25 billion guidance midpoint. Over the same period, operating margin has moved from 25.2% to 32.2% to 36.6% and now approximately 40%.

The old Lumentum financial model associated $2 billion of quarterly revenue with operating margins of 38–42%. The company is already around the midpoint of that range with only $1.25 billion of quarterly revenue.

When Mizuho asked CFO Wajid Ali what this meant for the longer-term model, the answer was unusually consequential:

“Think of the 42% outlier more as a midpoint, with the range probably moving up 100-200 basis points versus what we showed at OFC.”

That changes the earnings equation.

The usual high-growth semiconductor story asks investors to choose between revenue growth and margins. Capacity must be built before it is filled, new systems dilute component economics, R&D rises, and new customers require investment.

Lumentum currently has revenue estimates moving higher while the margins attached to those revenues also move higher.

My working FY27 assumptions are now roughly $6.2–6.5 billion of revenue and $23–25 of adjusted EPS. FY28 could reach $8.5–9.5 billion and $35–41 of EPS if the 1.6T, OCS, pump and CW ramps overlap as expected. Those numbers are estimates rather than company guidance, but the direction is more important than the exact decimal point: the company’s old financial model is no longer a useful ceiling.

The Other Fab

There is, of course, a familiar way this story can end.

Scarcity produces extraordinary margins. Extraordinary margins attract capital. New fabs come online, customers qualify second sources, pricing weakens, utilization falls and what once looked structural turns out to have been cyclical.

The most important bearish question on the call therefore came from Raymond James, which asked about new Chinese indium phosphide capacity.

Hurlston’s answer was that Lumentum has not yet seen meaningful impact. He argued that EMLs and high-power lasers remain differentiated and that even CW lasers can command substantial premiums because consistency and narrow specifications improve customers finished-transceiver yields. A cheaper laser is not necessarily cheaper if it causes more modules to fail.

That is a credible advantage. It is not an immutable one.

The street discussion after earnings highlighted Source Photonics and other Chinese suppliers progressing through 100G and 200G EMLs and potentially toward high-power lasers. I would not treat every channel claim as established fact, but the strategic question is exactly right.

The bear case is not that China needs to invent a laser.

It is that Chinese suppliers eventually achieve comparable yield, reliability, specification consistency and hyperscaler qualification at meaningfully lower prices.

If they do, Lumentum’s current scarcity economics can unwind even if AI networking demand continues growing.

This is the counterweight to the entire thesis. It is also why I do not assume today’s valuation multiple or scarcity margins indefinitely.

The Price of Duration

The market already knows Lumentum is having an extraordinary year. The TMT discussion around the print suggested that many investors were already using a September-quarter revenue bogey of roughly $1.20–1.25 billion rather than published consensus. By the time earnings arrived, some investors apparently wanted $1.30 billion for the stock to work.

That explains the initially muted reaction to objectively spectacular guidance.

Lumentum is a consensus long. Existing optical and semiconductor specialists have enormous gains. Some will raise their earnings estimates and sell shares at the same time because the position has become too large.

The next leg therefore requires a different buyer: large-cap growth and generalist technology investors willing to believe this is a multi-year earnings-duration story rather than the last year of an optical shortage.

That brings the investment question back to duration.

Using roughly $821 as the pre-earnings reference price, my probability-weighted three-year value is around $1,700.

The bear case assumes today’s scarcity is temporary. Capacity catches demand, Chinese and other suppliers close the technology gap, pricing normalizes and operating margins retreat toward the mid-30s.

The base case assumes several optical waves overlap. Pumps and scale-across continue growing; 1.6T and OCS drive FY27; NPO, CPO and ELS begin contributing before the previous growth vectors mature. The multiple can compress materially while earnings continue growing fast enough to carry the stock.

The bull case is more ambitious. It assumes optical connectivity meaningfully displaces copper closer to the compute and that Lumentum retains technological scarcity across enough of those connections to capture a disproportionate share of the value.

That produces my rough price path of $975 in three months, $1,100 in six months, $1,300 in twelve months and approximately $1,700 in three years.

I have the least confidence in the first number. Positioning can overwhelm fundamentals over three months. I have more confidence in the twelve-month question because by then investors should be valuing Lumentum against FY28 earnings rather than arguing about whether the September-quarter guide was $20 million above someone’s whisper.

What Would Change My Mind

The next quarter does not need to prove that AI demand is strong. That debate is largely finished.

It needs to show that $1.25 billion is another step rather than a peak. OCS should continue beyond its first $100 million-plus quarter. 1.6T should accelerate as promised. Gross margin should remain near or above 50%, and operating margin should remain above 40% rather than immediately mean-reverting.

More important for the three-year thesis, EML and pump demand should remain ahead of expanding supply; high-power laser capacity should convert into revenue without eliminating scarcity; additional ELS or NPO orders should turn architectural discussion into commercial evidence; and the lead CPO customer should remain on schedule for the H2 2027 laser ramp.

The bear signals are equally clear. Pricing weakness before supply is fully available would matter. Gross margin falling below roughly 49% while revenue continues growing would suggest today’s economics are more cyclical than I think. A closing supply-demand gap caused by softer demand rather than better output would change the interpretation quickly. Most importantly, credible Chinese products qualifying at hyperscalers with comparable yields would challenge the scarce-layer thesis at its foundation.

The useful thing about Lumentum today is that the argument has become testable.

When the Bottleneck Became Optical

My 2025 article ended with the idea that the next phase of AI networking would be won by companies building for requirements that seemed extraordinary today but would later appear inevitable.

A year later, two AI data centers apparently require twice the networking capacity one hyperscaler built across its entire global backbone over the previous decade.

That is an extraordinary requirement.

It is also increasingly real.

The network became a bottleneck because AI turned thousands of processors into one distributed computer. What Lumentum is now showing is that the bottleneck is moving again from protocols and switches toward the physics of moving an ever-growing amount of information.

That is why the company’s first billion-dollar quarter is almost beside the point. Lumentum is already guiding to $1.25 billion. OCS is crossing $100 million quarterly revenue. 1.6T has barely started. Pump output is heading toward four times today’s level. High-power laser demand is outrunning that expansion. NPO is appearing before CPO. ELS has moved from a concept to an order. And the operating-margin structure once associated with $2 billion of quarterly revenue has arrived hundreds of millions of dollars early.

Perhaps this is simply what the top of an optical cycle looks like when demand is unusually strong and supply unusually scarce. Capacity will arrive, competitors will qualify, customers will dual-source and margins will normalize.

But the alternative is more interesting.

AI may not simply require more networking. Each generation of AI may require optics in places where the previous generation could still use copper.

If that is true, Lumentum’s opportunity is not one shortage or one product cycle. It is the movement of the optical boundary itself.

Subscribe now

General Disclaimer: The information presented in this communication reflects the views of the author and does not necessarily represent the views of any other individual or organization. It is provided for informational purposes only and should not be construed as investment advice, a recommendation, an offer to sell, or a solicitation to buy any securities or financial products.

While the information is believed to be obtained from reliable sources, its accuracy, completeness, or timeliness cannot be guaranteed. No representation or warranty, express or implied, is made regarding the fairness or reliability of the information presented. Any opinions or estimates are subject to change without notice.

Past performance is not a reliable indicator of future performance. All investments carry risk, including the potential loss of principal. This communication does not consider the specific investment objectives, financial situation, or needs of any individual.

The author and any associated parties disclaim any liability for any direct or consequential loss arising from the use of this material and undertake no obligation to update or revise it.

• AI infrastructure is entering a new financing phase: Nvidia and major capital providers are building repeatable structures to mobilise hundreds of billions of dollars beyond hyperscaler balance sheets.

• The key metric is no longer just capex, but support: if each new dollar of AI capacity requires less vendor backing, the system is moving from Sponsorship toward genuine Self-funding.

• Nvidia may be building a financial moat: if Nvidia-based projects consistently secure better financing terms than competing architectures, CUDA’s advantage extends beyond software into the cost of capital itself.

Six weeks ago, in The AI Handoff, we argued that the AI buildout was moving through four stages: Scarcity → Sponsorship → Conversion → Self-funding. Scarcity had produced extraordinary supplier economics. Sponsorship was emerging as vendors, strategic investors and capital providers helped customers build infrastructure before the applications using it generated enough cash to pay for it independently. Conversion would arrive when those applications produced durable revenue and savings. Self-funding would begin when the returns on one generation of infrastructure financed the next.

The central question was whether falling costs and improving capability could create enough broad, independently financed end-user demand before the physical capacity and financial obligations built in anticipation of that demand outran the resulting cash flows.

This week Nvidia put a number on the next stage of that experiment.

The company announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR intended to mobilise more than $500 billion of third-party capital for AI infrastructure. Nvidia says it may provide residual-value support of up to 25% on individual opportunities. Two months earlier, Apollo and Blackstone had structured a $35 billion Broadcom AI financing with materially more extensive residual protection for senior capital.

The structures are not identical enough to declare that the difference is simply “the price of CUDA”. Borrower quality, tranche seniority, tenor, customer commitments and collateral assumptions matter. But the comparison gives us something we did not have six weeks ago: the beginnings of an observable market price for how much financial value investors attach to different compute architectures.

That matters because Nvidia and its financing partners are not necessarily trying to create the same end state.

Nvidia benefits if its architecture remains differentiated enough that Nvidia-based infrastructure receives lower financing costs, requires less vendor support and retains higher residual values. Capital markets benefit from something broader: a liquid, standardised asset class in which Nvidia, Broadcom, Google TPUs and eventually other architectures can all be financed and compared.

The strategic question is therefore no longer simply whether compute can be financed. It is whether Nvidia can financialise compute without helping commoditise it — while the wider AI system reduces its dependence on sponsorship and moves toward genuine self-funding.

Seven Men on Television

The unusual thing about Nvidia’s CNBC appearance was not the size of the number. It was the people sitting around Jensen Huang.

Goldman Sachs CEO David Solomon, BlackRock CEO Larry Fink, Blackstone President Jon Gray, Apollo President Jim Zelter, Brookfield CEO Bruce Flatt and KKR’s global head of digital infrastructure joined Huang to explain the new financing platforms. These institutions compete for capital and transactions, yet Nvidia had assembled them to describe different parts of what increasingly looks like a new financial system.

Huang started with the technology.

“The computer is now part of the infrastructure like electricity, like the internet.”

His argument was that Nvidia had crossed from selling components into supplying productive infrastructure. AI factories generate revenue, can run many workloads, serve multiple potential users and retain economic value beyond the credit quality of the original buyer.

His partners talked increasingly about the asset.

Gray compared compute finance with houses and aircraft: lenders underwrite the borrower, but also the value of the underlying asset. Brookfield said investors had lacked the proper “format” through which to invest and described Nvidia’s architecture and support as part of what makes the infrastructure transferable to another operator. Fink compared the moment with the early mortgage-backed-securities market. KKR described the value chain as running “from molecule to token”.
There is no contradiction yet. Nvidia may genuinely represent the highest-quality collateral in an emerging market. The divergence is in where each participant wants the market to end up.

Nvidia wants Nvidia compute to become a financial standard.

Wall Street wants compute to become a financial asset class.

The first depends on differentiation. The second eventually benefits from standardisation.

That tension is the most interesting part of the announcement.

Sponsorship, Institutionalised

The mechanics are relatively simple. Financing vehicles can raise third-party capital, purchase AI infrastructure and lease the resulting compute to customers. Goldman can structure and distribute the debt. Infrastructure and private-capital firms can provide equity or credit. Nvidia supplies the systems and may provide limited support against residual-value risk.

What has changed is the investor base.

The first phase of AI infrastructure was largely financed by capital close to the technology: hyperscaler retained earnings, venture capital, strategic equity and corporate investment-grade debt. Nvidia’s coalition is designed to reach capital that does not need to believe a particular AI stock will rise. Pension funds, insurers and long-duration credit investors want contracted cash flows, recoverable collateral and yield.

That turns Sponsorship from a series of bespoke arrangements into something closer to a repeatable capital market.

This does not make the underlying demand fake. Young infrastructure systems often require capital before customers can fully fund themselves. A model company needs compute before it can create the product and revenue that make the compute affordable. A data-centre developer needs a committed tenant before lenders will finance construction.

But financing changes what demand proves. As we wrote in July:

“A vendor-supported purchase says the supplier believes strongly enough in the customer’s future to help create the purchasing power. The second may be an excellent investment. It is not independent validation.”

The $500 billion announcement does not mark the end of Sponsorship.

It marks its institutionalisation.

25% Versus 100%

This is where the financing terms become more interesting than the headline amount.

Nvidia has described potential residual-value support of up to 25% on qualifying projects. In the earlier Broadcom financing, Broadcom provided much broader protection against residual shortfalls for senior capital.

Two different concepts need to remain separate.

The sponsorship ratio measures how much incremental infrastructure relies on vendor guarantees, strategic equity or related-party financing rather than independent customer cash and credit.

The residual-support ratio measures how much collateral downside the vendor agrees to absorb if the borrower fails and the equipment has to be recovered or re-leased.

The second may tell us something important about the first, but they are not interchangeable.

What the comparison may eventually reveal is an architecture financing premium.

If comparable Nvidia projects consistently require less residual support, lower coupons, less equity, smaller collateral haircuts or longer maturities than competing architectures, then Nvidia’s ecosystem has become more than a software moat. It has become a financial moat.

This is a much stronger proposition than saying CUDA makes GPUs valuable collateral. It means Nvidia’s installed base, software compatibility and secondary-user ecosystem reduce the cost of capital for anyone deploying Nvidia systems.

That advantage could reinforce itself. Lower financing costs create more Nvidia installations. More installations deepen the potential re-lease market. A deeper secondary market improves residual values. Better residual values lower financing costs again.

The opposite outcome is equally important. If economically comparable Nvidia, Broadcom and TPU projects eventually require similar enhancement and clear at similar spreads, compute is becoming financeable while Nvidia’s architecture premium is disappearing.

The executed deals, not the MOUs, will tell us which world is emerging.

Three Clocks

The financing initiative also extends the framework from The AI Handoff.

The construction clock runs on engineering. Chips have to be manufactured, power secured, substations built, data centres completed, networks installed and systems commissioned.

The conversion clock runs on enterprises. Companies must connect models to proprietary data, establish permissions, redesign workflows, test reliability, change budgets and persuade humans to work differently.

The capital clock runs on debt service.

Interest has a due date. Leases must be paid. Covenants have test dates. Refinancing windows open and close. Equipment depreciates. Debt matures.

The third clock has the hardest deadline.

A technology can be right over twenty years while the security financing it is wrong over five.

This is where Liaquat Ahamed’s 1873, the book Satya Nadella recently called “the book to be read”, becomes useful. Ahamed notes that roughly $500 million a year flowed into US railway bonds during the early 1870s, equivalent to around $600 billion relative to the size of the modern economy.

The railway itself was not the mistake. Northern Pacific eventually completed its line after multiple bankruptcies and ultimately became part of the system that formed BNSF. The infrastructure survived. Some of the securities used to finance it did not.

The parallel is not the outcome. It is the financing problem.

Capital must arrive before the full economic activity that justifies the infrastructure.

Nvidia’s coalition is designed to give Conversion more time before the capital clock runs out.

There is also a feedback loop. More financing creates more infrastructure. More infrastructure lowers the price of intelligence. Lower prices make new workloads economical. Those workloads accelerate Conversion.

Finance can therefore do more than wait for demand.

It can help create it.

The Great Elasticity Bet

That feedback loop contains the deepest assumption in the entire structure.

KKR noted that token prices have fallen dramatically even as usage has risen.

This makes AI an unusual infrastructure asset. It has the fixed-capital intensity of a utility but the output deflation of software. Investors are being asked to provide long-duration capital against a product whose unit price may decline by orders of magnitude while the equipment producing it is itself improving rapidly.

The financing works only if demand for intelligence is extraordinarily elastic.

If useful intelligence becomes ten times cheaper, society must consume substantially more than ten times as much for the aggregate revenue pool to expand strongly enough to support a growing infrastructure base.

There are good reasons to believe that will happen. Human cognition has historically been expensive, so companies rationed it. They reviewed some transactions rather than all transactions, ran a limited number of experiments, produced periodic analysis rather than continuous analysis and assigned finite numbers of people to tasks where dozens of digital agents might eventually operate in parallel.

Cheaper cognition can create workloads that did not exist at all at previous prices.

But rising token volume is not sufficient. What matters to infrastructure investors is revenue density: the economic output generated per unit of deployed capacity after utilisation, power, operating expense and depreciation.

If prices fall faster than paid usage expands, AI can become extraordinarily successful for users while infrastructure owners earn poor returns.

The technology and the securities can reach different verdicts.

Follow the Payer

This is why the same earnings season matters.

Palantir produced rapid revenue growth and strong free-cash-flow conversion. Atlassian showed expanding contracted demand and improving economics around AI-connected products. At the same time, several hardware companies reported extraordinary growth and still sold off.

This was not simply a rotation from hardware to software.

The market was separating capacity from economic conversion.

Ultimately, the financing vehicle does not get paid in tokens. It gets paid in dollars.

Those dollars must originate with an enterprise, consumer or government receiving enough economic value from AI to support the chain beneath it. That value may come from new revenue, substitution for existing expenditure or productivity savings. Whatever the mechanism, customer economics must ultimately fund model revenue, which funds infrastructure payments, which service the capital.

Huang’s claim that AI tokens are “incredibly profitable” therefore needs one further distinction.

A token may generate an attractive gross margin after inference cost. A lab may generate operating profit after salaries and product expenses. Yet the capital structure ultimately needs free cash flow after the company funds the next model, secures future compute and maintains its competitive capability.

Token profitability validates the product.

Free cash flow validates the infrastructure.

That makes a future OpenAI or Anthropic IPO important for a different reason than the valuation headline.

An IPO is a financial handoff. It can broaden the investor base and reduce the cost of capital.

Self-funding is an economic handoff. It occurs when customer-generated cash can support the next infrastructure cycle without ever-larger injections of external capital.

A company can achieve the first without achieving the second.

Public filings will finally allow investors to distinguish them.

Nvidia, Wall Street and Google

The financing coalition also exposes a longer-term strategic tension around who defines the asset class.

Apollo has already participated in the Broadcom structure. Blackstone is involved with Nvidia, Broadcom and a large Google TPU joint venture. Its own underwriting analogy focused on the economic value of compute rather than on Nvidia specifically.

These firms do not need Nvidia to dominate AI. They need AI infrastructure to become financeable.

Google makes that distinction particularly important.

Google can monetise AI through Gemini, Google Cloud, advertising, security, storage and its own TPU infrastructure. It can also earn money supplying compute to laboratories that compete with Gemini. It therefore has reasons to make TPUs broadly usable even if the leading model is produced elsewhere.

As non-Nvidia systems acquire independent customers, rental histories and financing transactions, capital markets gain something they previously lacked: comparables.

At that point Nvidia’s financing advantage becomes measurable.

Does an Nvidia project receive a lower coupon than a TPU project with the same borrower? Does the lender require less first-loss protection? Is the advance rate higher? Is the equipment assigned a better residual curve? Can the debt mature later?

Nvidia may win all of those comparisons.

But the existence of the comparison itself changes the market.

Nvidia is helping create an enormous financing category while simultaneously betting that its own systems remain the premium asset inside it.

Wall Street wants the category.

Nvidia wants the premium.

The Scorecard

We now have enough evidence to track the handoff rather than merely describe it.

The direction matters more than the current level.

Five Calls

First, watch hyperscaler capex. The largest buyers still fund substantial infrastructure from their own economics. A meaningful capex-guide reduction from a major hyperscaler would be the clearest signal that the construction clock is encountering an economic constraint rather than a physical one.

Second, watch long-term memory commitments. A material cancellation or roll-down in contracted HBM or memory demand would turn the oversupply thesis from narrative into operating evidence.

Third, watch lab disclosure rather than IPO valuation. A spectacular IPO accompanied by persistent dependence on outside capital is financial graduation, not economic graduation. The real test is gross margin, free cash flow and compute obligations.

Fourth, watch Nvidia’s financing premium. If comparable Nvidia projects clear with materially less support and better terms than competing architectures, Nvidia has created a financial moat. If the terms converge, the compute asset class is standardising faster than Nvidia would prefer.

Fifth, and most importantly, watch the Sponsorship Ratio. If the first $100 billion of financed capacity requires 25% support and the next $100 billion requires 35%, the system is becoming more dependent on external support as it grows. If the requirement falls toward 15%, 10% and eventually zero because customers increasingly finance projects using their own cash and credit, the handoff is working.

The View

The Nvidia announcement is bullish for the pace of the AI buildout. It opens pools of capital to customers that do not possess hyperscaler balance sheets, accelerates infrastructure construction and should push the price of intelligence lower.

It is also the clearest evidence yet that AI has entered a new financial phase.

The question is no longer whether Wall Street can finance AI. It clearly can.

The first question is whether the amount of sponsorship required declines as the system scales. That tells us whether Conversion is producing enough independent cash flow for AI to move toward Self-funding.

The second is whether Nvidia retains structurally better financing terms than competing architectures. That tells us whether CUDA and Nvidia’s installed ecosystem can remain differentiated as compute itself becomes increasingly standardised and financeable.

The best outcome for Nvidia is straightforward: the Sponsorship Ratio falls while the architecture financing premium remains high. AI becomes increasingly self-funding, and Nvidia remains the preferred asset inside the capital market built around it.

The weaker outcome is the reverse: sponsorship rises while financing terms converge across architectures. AI may still transform the economy, but infrastructure remains dependent on external capital while compute commoditises underneath Nvidia.

That is why the executed transactions matter more than the $500 billion headline.

Six weeks ago, the Sponsorship Ratio was a framework.

This week, we began to see what might price it.

The next deals will tell us whether the number is going down.

Subscribe now

• Shopee’s underlying economics are improving: advertising grew more than 70%, core marketplace revenue rose 65.6% on 28.4% GMV growth, and logistics unit costs continued falling.

• The missing margin is being reinvested: Sea beat revenue expectations by nearly $700 million but delivered EBITDA roughly in line, as better economics funded customer acquisition, fulfillment, VIP and credit expansion.

• The integration moat has limits: Monee’s economics are strongest inside Shopee. Off-platform lending and Brazil introduce higher provisions and will test how well Sea’s advantage travels beyond its core ecosystem.

“Our improving operational efficiency and growing scale have strengthened our unit economics. We can now profitably serve a wider range of users, enabling us to lean further into user acquisition.”

— Forrest Li, Sea Q2 2026 earnings call

Sea reported $7.8 billion of second-quarter revenue, nearly $700 million above the Bloomberg consensus we had going into the print. Adjusted EBITDA was $917 million. Consensus was $919 million.

That juxtaposition is more interesting than either number by itself.

Sea generated far more revenue than investors expected and almost none of that upside appeared as incremental reported EBITDA. The obvious conclusion is that the company is still spending aggressively. The more important question is whether that spending reflects poor underlying economics or increasingly attractive opportunities to reinvest economics that are improving.

I have been writing about Sea for five quarters, and there is a useful correction to make before answering that question. I have called essentially the same thesis a flywheel, a fortress, the rice-field advantage, and finally integration density. The terminology moved around more than the investment thesis did. I am going to stop renaming it.

The thesis is simply that Sea builds dense consumer activity in difficult emerging markets and can reuse that density across commerce, logistics, advertising, and financial services. More activity should lower service costs, make access to consumers more valuable to sellers, and generate better information for underwriting. If those improvements exceed the cost of acquiring the next customer, the economics can finance still more activity.

Six months ago, in The Choice Not to Harvest, I admitted that I had been early on when those economics would appear in Shopee’s reported margin. Q2 does not erase that mistake. Shopee adjusted EBITDA is still only around 0.7% of GMV. What Q2 does provide is the strongest evidence yet that the machinery underneath that margin is beginning to work.

What We Said, What Happened

The best way to avoid changing the story every quarter is to score the old one.

In Q3 2025 I laid out several explicit signposts for Q2 2026. Two worked considerably better than expected; the one that mattered most for reported profitability did not.

The margin miss matters. I had assumed that investments made through 2025 would begin maturing and visibly release profit by now. Some are maturing, but Sea keeps finding new things worth funding: fulfillment, instant delivery, VIP, content distribution, Brazil, and new Monee customers.

That suggests my old mental model was incomplete. I implicitly assumed:

invest → mature → harvest.

Sea increasingly looks more like:

invest → economics improve → more customers or use cases become economically attractive → reinvest.

That distinction is crucial because successful investment can delay the reported margin precisely because it is successful.

Forrest Li’s statement that Shopee can now “profitably serve a wider range of users” is the clearest expression of this. Better economics have not simply produced more EBITDA. They have expanded the pool of customers Sea believes it can acquire profitably.

The Investment Cycle That Keeps Renewing

The evidence begins with customers. Average monthly new active buyers grew more than 35%, active buyers increased 18%, and purchase frequency rose 8%. Shopee was simultaneously adding customers and extracting greater engagement from existing ones.

Then there is monetization. Advertising revenue grew more than 70%, ad take rate increased over 90 basis points, the number of ad-paying sellers rose around 45%, and average ad spend per seller increased more than 15%. Core marketplace revenue — transaction fees plus advertising — grew 65.6%, more than twice Shopee’s 28.4% GMV growth.

This is a healthier form of monetization than simply increasing commissions. Management acknowledged that the pace of fixed commission increases will probably slow, but sees substantial remaining opportunity in advertising, where AI is improving targeting, campaign diagnosis, personalized vouchers, and seller returns.

The physical economics are even more revealing. Instant-delivery volumes in Indonesia grew roughly 80% while cost per order fell around 20%. Fulfillment volumes increased more than 20% sequentially, and listings shifted into fulfillment generated more than 20% higher orders on average in Southeast Asia.

That is the old rice-field idea finally appearing in economics rather than anecdotes:

Fulfillment remains early. Sea says its penetration is still well below several peers, particularly in Brazil, but its economics are improving as seller participation increases, warehouses integrate more tightly with SPX, utilization rises, and automation lowers cost. Importantly, Sea rents fulfillment facilities rather than owning the underlying land and warehouses, limiting the capital required to pursue the opportunity.

This is why the $917 million of group EBITDA matters. The extraordinary revenue beat did not produce an EBITDA beat because Sea is still reinvesting aggressively. That does not mean Sea could simply have reported whatever EBITDA it wanted; different revenue streams carry different incremental costs. What the quarter does tell us is that reported profitability continues to understate how rapidly some of the underlying economics are improving.

The comparison with Amazon’s long reinvestment period is tempting, but it should not be taken too far. The relevant lesson is not that Sea is another Amazon. It is that reported margins can be a poor measure of economic progress when successful investments repeatedly create adjacent opportunities worth funding. Sea still has to prove that those opportunities earn attractive returns.

So far, Shopee increasingly is.

Management reaffirmed that 2–3% EBITDA/GMV remains within reach and described a straightforward bridge: investment programs mature, logistics and fulfillment costs decline, and take rate improves through advertising and other monetization. It also confirmed that absolute Shopee EBITDA in the second half should exceed the first half.

That does not tell us when 2–3% arrives. It does tell us why 0.7% need not be the ceiling.

Where the Advantage Gets Thinner

Monee is where the integration thesis is both most impressive and most vulnerable.

The loan book grew 62% to $11.1 billion while 90-day NPLs remained at 1.0%. Sea’s newer underwriting models, trained on sequences of behavioral and transactional information, increased approval rates by around 10% without increasing observed risk. The company added 5.3 million first-time borrowers, active credit users rose 34% to more than 40 million, and average balances per user increased 20%.

This is the mechanism we have been arguing for several quarters: commerce produces information that allows Monee to make better credit decisions.

But Q2 also revealed the boundary.

Off-Shopee SPayLater now exceeds 20% of the portfolio, reaching 35% in some countries, while Sea is preparing a standalone financial-services application in Brazil. Management explicitly said higher credit provisions reflect greater exposure to off-Shopee lending and Brazil, which it described as a higher-interest, higher-risk market.

Monee’s numbers show the trade-off:

Source: Sea Q2 2026 results.

The old question was whether Sea’s proprietary data advantage existed. Q2 makes that increasingly difficult to dismiss.

The better question is now how quickly that advantage decays as Monee moves farther from Shopee.

Management says every product, segment, and country must maintain stable risk and positive returns on assets. That may make higher provisions perfectly rational if higher-risk markets also produce appropriately higher yields.

But this is exactly where the thesis should be tested rather than celebrated.

The Brazil Test

Brazil increasingly matters because it tests whether Sea’s operating model travels.

Shopee remains Sea’s fastest-growing market there. Management says it is growing faster than both the market and close competitors despite competitive changes to shipping thresholds and pricing and expects to continue outgrowing the market profitably.

At the same time, fulfillment remains underpenetrated, VIP has already exceeded one million Brazilian members, Monee lending is expanding, and Sea plans to launch a broader financial-services application with capabilities comparable to existing local competitors.

Sea does not need to defeat MercadoLibre for this thesis to work. It needs to demonstrate that the sequence developed in Southeast Asia, commerce density, lower service costs, higher monetization, and eventually broader financial services can be reproduced somewhere else.

If Brazil does that, integration is not simply a regional accident. It is an operating capability.

What We Know Now

The variant perception is not that investors have failed to notice the $1 billion Shopee EBITDA guide. They have.

The more interesting possibility is that the market still treats Shopee’s reported margin as a reasonable proxy for its underlying economics. Q2 suggests that assumption is becoming less safe. Logistics costs are declining with density, advertising is monetizing that density at much higher rates, and Sea is simultaneously adding customers without resetting Shopee EBITDA.

That does not mean management can manufacture any margin it wants. It means the eventual margin may be considerably higher than the current P&L implies.

My three-year framework therefore depends less on predicting perpetual 25–30% GMV growth and more on how much of these improving economics eventually reaches earnings.

At roughly $127–128 following the print, the broad probability-weighted value lands around $250–260, roughly a mid-to-high-20s annualized return over three years. The base case does not require Sea to reach 3% margins. It requires transaction density to continue improving economics gradually while Monee scales without a material deterioration in risk.

The tracking ledger from here should therefore remain fixed: Shopee EBITDA/GMV sustainably above 0.8–0.9%; advertising continuing to materially outgrow GMV; logistics unit costs continuing to fall; Monee NPL staying around 1.0–1.2%; credit provisions stabilizing as off-Shopee lending grows; and Brazil maintaining above-market growth without requiring escalating subsidies.

Six months ago, I wrote that the fields were not ready to harvest. I still would not declare them ready today.

What Q2 changes is our understanding of why.

The investments we expected to mature are maturing. The problem, if it can be called one, is that their success keeps revealing another economically attractive place to invest. That makes Sea’s eventual margin harder to time.

It also makes the underlying earnings power easier to believe.

Subscribe now

• The pricing mechanism changed: Ibiden had historically assumed routine quarterly price reductions. In Q1, like-for-like prices held flat while new products reset higher, turning manufacturing difficulty into direct pricing power rather than merely richer product mix.

• Customers are financing the scarcity: advance payments surged from ¥81.0 billion to ¥154.5 billion as customers continued asking Ibiden to build as much capacity as possible despite competing expansion elsewhere.

• The remaining question is durability: Q1 shows that customers will pay for Ibiden’s capability today. The real test is whether pricing, margins and ROIC remain elevated after Cell 6, Cell 8 and competing capacity arrive. That determines whether Ibiden is a cyclical beneficiary, a compounder, or genuine infrastructure.

In April, I wrote that the core question about Ibiden was not whether AI substrates were getting harder, they obviously were, but whether that difficulty would compound into Ibiden’s economics or merely force the company to spend ever more capital to stay relevant. I framed three classifications: Cyclical, Compounder, Infrastructure. I identified customer advance payments as the leading indicator. And I said that if advances kept rising, the market had underestimated what Ibiden was becoming.

In May, the results offered a partial answer. The profit bridge showed a fourteen-to-one ratio of value to volume in the Electronics OP revision, ¥20.5 billion from ASP and product mix versus ¥1.5 billion from volume. Difficulty was entering the P&L. But the advance payment balance declined, from ¥92.1 billion to ¥81.0 billion, and I confronted that honestly: “I set the signpost. The direction was wrong. That deserves weight.” My classification: Compounder, with optionality toward Infrastructure. Check back in two quarters.

One quarter later, the evidence moved further than I expected, and differently from how I expected. I was right about the mechanism. I was wrong about the magnitude, and about the channel through which it would express itself. The gap between what I anticipated and what happened is where this article lives.

When the Price-Down Stopped

For most component manufacturers, productivity comes with an implicit obligation. The supplier learns how to make the product more efficiently, and the customer claims part of that improvement through regular price reductions. Ibiden’s own planning model reflected this convention. Every quarter, management assumed another price-down. In May, I wrote that the resolution of Ibiden’s pricing tension would probably come through mix rather than rate, each generation’s substrate larger, more complex and higher-priced, even if the pricing on any given product stayed disciplined. I cited CEO Kawashima’s statement that the company did not want to increase prices “beyond the standard zone” and framed this as the TSMC playbook: restrained pricing for long-term incumbency.

Q1 broke that framework.

Goldman Sachs’ Daiki Takayama asked management to decompose the ¥26.5 billion ASP and product mix contribution in the new Electronics OP revision. Shinji Miyazaki, the Director and Senior Executive Officer, was direct:

“According to our business plan, every quarter, we incorporate a price reduction. Because of our past engagement with the customers, we reflect that price reduction. However, in the negotiations since 11th of May, this price decline curve became less steep, or in some cases, we didn’t have to reduce the price throughout our negotiation.”

Takayama pressed to confirm his understanding: on an apple-to-apple basis within the same product, there was no price reduction, and for new products, the pricing was reset higher?

“Yes. Your understanding is correct.”

The quarterly price decline curve, the economic gravity of the ABF substrate industry, the assumption encoded in every sell-side model, in management’s own May 11 guidance, in my own May framework, stopped functioning. On existing products, prices held flat. On new products, the baseline was reset higher. JP Morgan confirmed that most of the ¥26.5 billion was pure ASP, not product mix. The combined effect was operating profit that was in nobody’s plan three months ago, including Ibiden’s.

A larger, more complex substrate naturally carries a higher absolute price because it contains more material and processing. That is mix. A customer abandoning the expectation that the same product gets cheaper every quarter is something else. That is bargaining power.

I had the right mechanism. Difficulty was compounding into economics, exactly as the yield-learning loop predicted. But I assumed it would express itself indirectly, through mix. What actually happened was more direct: the difficulty entered the price line itself, because at the highest complexity tiers, customers concluded they had no practical alternative at comparable yield and quality, and they stopped negotiating as though they did.

This distinction matters for classification. Mix-driven margin expansion is a Compounder characteristic, the products get harder and more valuable, but any given product remains competitively priced. Price-driven margin expansion suggests something closer to Infrastructure, the capacity itself becomes the scarce asset. Q1 showed both operating simultaneously.

Capacity Is Not Capability

The conventional response to high prices is more supply. Ibiden is investing. Unimicron, AT&S, Samsung Electro-Mechanics and others are investing. Eventually, the argument goes, capacity will arrive and margins will normalize.

That argument treats all capacity as equivalent.

It is not. Installed capacity, qualified capacity and economically productive capacity are three different things, separated by years of engineering work and yield learning that capital alone cannot compress. An advanced AI substrate requires twenty-plus routing layers built across more than sixty discrete process steps at sub-ten-micron tolerances. Cumulative yield for the most demanding specifications is substantially lower than for prior generations, and the exact numbers are closely held. A competitor can buy the same equipment. They cannot buy the yield.

The more important development in Q1 was that this capability appears to be generalising. Ibiden is no longer responding only to GPU demand. UBS’s Shingo Hirata asked whether ARM-based CPU substrates would carry comparable complexity and margin to GPU products. Miyazaki’s answer:

“The difficulty level will be similar with the AI GPU that we are currently working on.”

High-end x86 CPUs, ARM inference CPUs and switching ICs are now consuming the same scarce manufacturing capability as GPU substrates. EMIB-T adds silicon-bridge embedding, higher power delivery and more complicated core structures. When Mizuho’s Fumihide Gotoh asked whether EMIB-T faced any barriers that might prove insurmountable at mass manufacturing scale, Miyazaki said there was “no showstopper” and that yields had “improved quite a lot” from the start of development.

Six months ago, I framed Ibiden’s advantage as potentially GPU-specific, which carried customer concentration risk. What Q1 suggests is that the underlying asset may be more general: expertise in manufacturing the hardest substrates, regardless of the application on the other side. That broadens the addressable market without requiring new capability development.

Our thesis is not that competition does not exist. It is that the complexity frontier may be moving faster than competitors can close the qualification gap. Each harder generation narrows the qualified field, concentrates volume among the survivors, and accelerates the learning that wins the next generation. That mechanism, which I described in April, received its most direct financial confirmation in Q1.

What the Customers Are Saying

The balance sheet told a clearer story than the earnings release.

Customer advance payments rose from ¥81.0 billion to ¥154.5 billion during the quarter, driven by ¥73.6 billion of new prepayments. In April, I set ¥130 billion as the confirmation threshold and ¥70 billion as the reassessment floor. In May, I tightened to ¥90 billion after the decline. The balance is now ¥154.5 billion, through both thresholds, and decisively.

Nomura’s Manabu Akizuki asked whether a Taiwanese competitor’s aggressive expansion had made customers more reluctant to prepay. Miyazaki:

“Customers’ stance hasn’t changed from the Cell 6 situation. Taiwanese company is working very hard to grow their businesses. We understand that as well. At the customer side, it’s not that customers have become tougher and has become hesitant to pay the advance payments. We still see strong demand coming from them, we are asked to build the capacity as much as possible. That tone or nuance has remained the same.”

Customers see the competition being built. They are still prepaying Ibiden. That is not how you treat a supplier you expect to have leverage over in eighteen months.

When Citi’s Takayuki Naito asked whether Ibiden would offset the fixed costs of new capacity through customer prepayments, Miyazaki confirmed: “As much as possible, we’d like to receive the advance payments to compensate for the increase in the fixed costs.” Customers absorb part of the investment risk while Ibiden retains the manufacturing economics. That is an unusually favourable capital structure for a manufacturing supplier.

Pricing tells us what current qualified capacity is worth. Advance payments tell us what customers believe future qualified capacity will be worth. Both are moving in the same direction.

The Miss Was in the Model

Q1 operating profit beat consensus by 29%. The full-year guide came in 23% above consensus. Ibiden itself raised its forecast by 41% in three months.

The failure was not primarily one of information. Sophisticated investors knew AI demand was strong. They knew capacity was tight. They knew Ono was improving and that customers were asking about future supply. The physical facts were hardly secret.

A channel check saying that every factory is full can lead to two conclusions. The first is that volume upside is capped. The second is that allocation to the existing volume has become more valuable. The first interpretation fits the old substrate model. The second describes what happened in Q1. The data was available. The framework for interpreting it was not.

The models were not wrong because the analysts are bad. They were wrong because the quarterly ASP decline assumption, correct for two decades, had been hardcoded as a fixed input rather than modelled as a variable that could change when the market crossed from capacity-constrained to capability-constrained. Even management’s own planning system contained the old rule. The customers invalidated it, and neither the company nor the Street knew until the negotiations took place.

The more interesting implication is about the remaining variant. Several sell-side houses already model consolidated operating margins above 30% by the end of the decade. The debate is not simply whether consensus is too low on the numbers. It is whether these earnings deserve to be treated as peak-cycle earnings or as evidence of a more durable economic structure. The variant perception, in other words, is about duration, not magnitude.

Scarcity or Moat?

There is an uncomfortable problem with all of the evidence assembled above: scarcity economics and moat economics look almost identical while everybody is sold out.

Ibiden raising prices today does not prove pricing survives Cell 6, Cell 8 and competing capacity. Customers making advance payments today does not prove they keep doing so once credible second sources exist. Engineering scarcity today does not prove Ibiden can scale its organisation without turning its own talent constraint into a growth constraint. EMIB-T development progressing does not prove mass-production yields will generate attractive returns.

Our April question was whether difficulty compounds faster than capital consumes it. Q1 answered the first half: customers are paying for the difficulty. They are paying more for current capacity and financing the next generation of it. What they cannot answer is whether the economics survive the factories they are helping to build.

That is the experiment now underway. Ibiden is about to add substantial qualified capacity, and its competitors are doing the same. If like-for-like pricing holds, margins remain elevated and return on capital rises as Cell 6 and Cell 8 mature, then complexity rather than temporary scarcity is setting the economics. If price-downs resume and margins compress as supply arrives, Q1 will mark the most profitable shortage in the company’s history rather than a structural change in how the industry works.

In May, I said ROIC inflecting above 12% by FY2028 was “the ultimate verdict” separating a well-positioned cyclical from a structural compounder. That test has not occurred. It cannot occur yet. Everything between now and then is a leading indicator. The leading indicators are strong. But the verdict belongs to ROIC, not to a single quarter’s pricing data.

Three Futures

The scenarios below begin from Ibiden’s FY3/27 revenue guide of ¥550 billion and value the company on FY3/30 earnings. Each column answers a different version of the same question: what kind of company is Ibiden becoming? Prices are on the current pre-October 2026 split basis.

Cyclical: Complexity is real, but qualified capacity catches up. Price-downs return. The current scarcity rent is competed away as Cell 6, Cell 8 and competing capacity come online. Depreciation from the ¥210 billion annual capex program absorbs much of the margin expansion. The work keeps getting harder; the economics do not scale.

Compounder: Qualified supply remains constrained enough that like-for-like price erosion stays well below historical norms. New capacity fills. CPU and switching demand broaden the revenue base. Margins settle in the high-20s to low-30s. Customer co-funding continues around each new build. This is where the current evidence points, and it is my base case.

Infrastructure: The complexity frontier keeps moving faster than effective supply. EMIB-T succeeds at commercially viable yields. Ibiden generalises its manufacturing lead across GPU, CPU, switching and advanced bridge architectures. Margins expand even while capacity scales. ROIC rises despite the huge capital program. The market reclassifies.

What Resolves It

In April, I asked whether difficulty would compound faster than capital consumed it. The customers have now answered the first half of that question: they are paying more for Ibiden’s current capacity and financing the next one. The pricing mechanism that governed this industry for two decades has, for the first time, stopped functioning at the leading edge.

What the customers cannot answer is whether those economics survive the factories they are helping to build.

That is the experiment now underway. If pricing and returns hold as qualified supply expands, Ibiden is infrastructure. If they do not, this was an extraordinarily profitable shortage. The distinction is worth considerably more than one quarter’s earnings beat, and ROIC over the next two years will settle it.

The river in Gifu has not stopped flowing. The advance payments have spoken. The old bargain has broken. Whether a new one has taken its place is the question this company must now answer at scale.

Subscribe now

• The core thesis is being validated: Constellation signed roughly 920 MW of long-term nuclear agreements averaging 18.5 years, raised 2026 EPS guidance, and progressed Crane and the Calpine integration. The “boring” contracting thesis is working.

• Regulation looks more like friction than a veto: the new PPAs were signed before PJM and FERC rules were fully resolved, suggesting scarce, already-connected generation can be monetized even without perfect regulatory clarity.

• The rerating case is increasingly about earnings quality, not estimate upside: the opportunity is for more of Constellation’s already-expected earnings to become long-duration, contracted and defensible not necessarily for 2029 EPS to dramatically beat consensus.

“I recognize that the last time we spoke, I indicated that we expected to be done with an important transaction by this call, but we’re not ready to announce anything today.”

That was Joe Dominguez in March. Constellation had just laid out 20% base-EPS growth through 2029, authorized $5 billion of repurchases, and presented a financial framework designed to make the company look less like a merchant generator and more like a business with increasingly visible earnings. The stock fell 9.3% because none of that was what investors had shown up to hear. They wanted a hyperscaler deal.

Four months later, Constellation delivered one. Or rather, several. The company signed approximately 920 MW of long-term nuclear agreements averaging 18.5 years, raised 2026 adjusted EPS guidance to $11.50–$12.50, cleared two important regulatory hurdles for Crane, and reached agreement on the final Calpine disposal required by the DOJ. Dominguez’s summary of the fears from March was direct: “Those concerns have not materialized.”

And yet I don’t think Q2 changed the thesis. What it did was more useful: it reduced the probability that the thesis we already had would fail.

The Big Fundamental Question remains the same:

Can Constellation convert the growing scarcity of reliable, already-connected power into durable contracted earnings faster than new supply and regulation erode that scarcity?

Q2 moved the answer toward yes.

What We Said, What Happened

Our view of Constellation has evolved less than the language around it suggests. In August 2025, we argued that Constellation sold pressure rather than power. Nuclear was the reservoir: dependable, carbon-free generation available whenever customers needed it. The PTC was the levee protecting economics when power prices weakened; capacity markets rewarded readiness; long-term contracts were the pipes carrying that dependable output to customers.

We then corrected ourselves. We were right about the reservoir and early on the pipes. Co-location proved slower and more regulated than expected, while Calpine widened the ways Constellation could reach customers through gas, storage, retail relationships and powered sites. After Q1, we sharpened the customer problem again: increasingly, the buyer was not simply purchasing clean electricity but a date, power delivered when a data centre or industrial project needed it.

Q2 does not invalidate any of this. It does make one distinction clearer. The core thesis, that scarce operating generation can be converted into increasingly durable earnings, is working. The more ambitious extension, that Calpine lets Constellation repeatedly sell Freestone-like packages combining site, connection, generation and delivery schedule, remains unproved.

That matters because the 920 MW signed this quarter were not spectacular transactions. They were long-term nuclear PPAs. The boring version of the thesis worked.

The Pipe Moved Before the Permit

This is the most important fact from Q2.

In March, management blamed regulatory uncertainty for slowing contract execution. Customers did not know what PJM would require of large loads or how future rules would affect procurement. The obvious bear chain was straightforward: no clear rules meant no customer certainty; no certainty meant no contracts; no contracts meant scarcity remained stranded.

Q2 broke that chain.

When Jefferies’ Julien Dumoulin-Smith pressed management on whether the signed deals depended on the Reliability Backstop Procurement or future FERC decisions, management said they did not. These are executed contracts, not placeholders waiting for regulation to catch up.

PJM still matters. Management expects the RBP process to produce results by year-end and broader co-location clarity around the first or second quarter of 2027. But regulation increasingly looks like a constraint on pace and structure rather than a veto on monetizing the fleet.

The deeper reason is the time asymmetry underneath the entire power story. Large customers can commit capital and add load much faster than new dependable generation, and transmission can be permitted, interconnected and built. Existing, grid-connected generation is therefore the scarce asset during that gap, and Constellation owns 55 GW of it; management estimates that rebuilding the fleet today would cost more than three times the company’s current enterprise value.

Dominguez explained the same point in grid language:

“We have plenty of unused capacity in generation and in the wires grid over 99% of the hours of the year. We have a peak capacity concern, not an energy concern.”

His argument is that batteries, demand response, backup generation and curtailment can handle the relatively small number of difficult peak hours while existing generation serves incremental load during the rest.

That mechanism does not require a landmark co-location agreement. Capacity markets already pay more for scarcity. Bilateral PPAs already monetize long-term nuclear value. Commercial margins capture volatility. Buybacks turn cash into per-share earnings. The routing layer could make the flywheel turn faster, but Q2 showed that the flywheel does not depend on it.

Optionality Gets Attention; Base Gets a Multiple

Our Q1 article contained the line that still best explains the financial mechanism: “Optionality gets attention; base earnings get a multiple.”

An uncontracted nuclear MWh is an option on tighter power markets, rising capacity values and customer urgency. Sign an 18-year agreement and Constellation gives away some of that open-ended upside. In return, it gets duration.

The new contracts average 18.5 years and take roughly 30% of expected clean baseload generation under long-term agreement by 2032. Yet Shane Smith made clear that they start late enough that they do not materially affect 2029 earnings.

That sounds disappointing until you look at what is already in the model. Bloomberg expects adjusted EPS of roughly $11.75 in 2026, $13.06 in 2027, $16.61 in 2028 and $19.10 in 2029, despite very little revenue growth after Calpine enters the base. The Street is already assuming earnings conversion.

What the new contracts do is improve the durability of earnings beyond the period that dominates today’s forecast.

The same idea appears in the Illinois transition. Constellation models 53 million MWh of Illinois CMC nuclear output in 2026 at $34.09/MWh, falling to 23 million MWh in 2027, while nuclear volumes supported in the PTC bucket rise from 101 million to 127 million MWh and the assumed floor increases from $44.75 to $45.75/MWh. A casual reading sees a subsidy ending. The economic reading sees a large block of output migrating from one support mechanism into another.

That is what management’s base-earnings architecture is trying to show.

The obvious bull case is no longer differentiated. Nuclear scarcity is well understood. AI electricity demand is well understood. Twenty-one analysts rate CEG a Buy and none rate it a Sell.

I do not think the interesting bet is that $19.10 of 2029 EPS is wildly too low. The interesting bet is that the character of those earnings changes faster than the multiple anticipates. If more of the $19 becomes long-term contracted, inflation-linked and backed by assets that cannot be quickly replicated, then CEG does not need a dramatic estimate beat to work. It may simply deserve a better multiple on the earnings already expected.

The risk is the mirror image. Constellation could deliver the EPS and still fail to re-rate if enhanced earnings remain large, PPA pricing sits near the low end of management’s $20–$50/MWh premium range, and the broader customer-solutions business never becomes repeatable.

Steve Fleishman pushed Shane Smith directly on the pricing range. Smith refused to narrow it. That leaves an important gap in the evidence. Volume is disclosed. Duration is disclosed. Counterparty quality is disclosed. The economic premium is not.

What Management Is Really Selling

The prepared remarks and Q&A suggest management is trying to make three arguments at once.

The first is that the existing fleet, not speculative new build, is the asset. When KeyBanc’s Sophie Karp asked about new nuclear, Dominguez said conversations in New York were interesting but “not that imminent.” Management wants investors focused on the value of megawatts that already exist.

The second is that regulatory clarity will increase deal velocity rather than create demand from scratch. When JPMorgan’s Jeremy Tonet asked how customer conversations had changed, Dominguez said ambiguity had been the enemy of deal execution and predicted that contracting could “kick off with a bit of a bang” once rules settle. That is a strong claim, and it now has a timetable.

The third is that Calpine should be judged as more than $2 of EPS accretion. Calpine contributed $1.028 billion of Q2 RNF, $2.147 billion of operating revenue less $1.119 billion of purchased power and fuel, while Brazos Valley is being sold for about $1,420/kW. Required disposals should generate roughly $5.9 billion at close to $1,200/kW against approximately $960/kW implied by the acquisition.

That supports the asset-value argument. It does not yet prove the full time-to-power thesis. We said after Q1 that if Freestone remained a one-off, Constellation could still own an excellent clean-firm contracting business while the broader opportunity turned out to be less valuable than we imagined.

That test remains open.

A Good Quarter With Messy Accounting

Adjusted EPS of $2.55 beat Bloomberg’s $2.32 estimate, and management raised full-year guidance by $0.50 to $11.50–$12.50. But adjusted EPS was also $1.13 above GAAP EPS of $1.42.

The bridge is substantial. Unrealized fair-value adjustments added $0.94 per share to the reconciliation, acquired commodity-contract amortization added $0.41, Calpine merger and integration costs added $0.23, pension and OPEB added $0.06 and legal/environmental items added $0.10, partly offset by $0.61 of decommissioning-related activity.

I think most of those adjustments are understandable for a power business absorbing a large acquisition. Still, an improving earnings-quality thesis eventually needs to appear in cash. H1 operating cash flow was $1.55 billion versus $1.58 billion last year, while capex rose to $2.52 billion as Calpine, Crane and other projects absorbed capital.

Bloomberg expects free cash flow to reach about $6.1 billion by 2029 and net debt to fall below $14 billion. If that does not happen, the multiple argument weakens considerably.

Three Years From Now

For an August 2029 value, I care more about earnings composition than revenue growth.

The base case is not heroic. Bloomberg already has $19.10 of EPS in 2029. It requires continued contracting, Crane broadly on schedule, nuclear reliability near historical levels, higher free cash flow and a growing base share of earnings.

The bull case needs more: several gigawatts of premium contracting, stronger gas utilization, at least one repeatable Freestone-like model and continued accretive capital allocation. The bear is not “AI disappears.” It is that supply and regulation catch up before Constellation converts enough scarcity into contracts, while enhanced earnings normalize and the market returns the stock toward conventional power-company valuation.

The Scoreboard

The next year should tell us whether Q2 was the beginning of a cadence or merely a strong quarter. Cumulative new long-term nuclear contracting above 2 GW by the Q4 2026 earnings call would signal acceleration; below 1.5 GW would suggest that Q2 did not change velocity. Contract pricing remains harder to observe, but evidence that deals are clearing comfortably inside the $20–$50/MWh premium range would strengthen the rerating case; repeated deals near the low end would weaken it.

PJM needs to deliver the expected RBP outcome by year-end and broader co-location clarity by 1H27. Meaningful slippage beyond that would restore some of the regulatory discount. Nuclear capacity factor should remain around 93–94% or better outside planned-outage noise; sustained readings below 92% would raise a different and more serious question about the asset itself.

Finally, the accounting has to catch up with the story. The absolute GAAP-to-adjusted gap should shrink as Calpine purchase-accounting and integration effects fade, while free cash flow should begin moving toward the $6 billion level Bloomberg expects late in the decade. And if Constellation signs another material Freestone-like agreement by the end of 2027, the broader time-to-power thesis would finally begin to move from possibility to repetition.

The Same Bet, Better Evidence

For a year, we kept asking when the routing layer would activate. Q2 suggests that was too narrow a question.

The structural advantage is the time gap between demand and new supply. Capacity markets monetize it one way. Long-term PPAs monetize it another. Commercial execution captures volatility. The PTC raises the floor. Buybacks turn cash into per-share growth. None of those mechanisms requires the spectacular transaction investors have spent eighteen months waiting for.

The routing layer could make the machine better. It could give Constellation more ways to solve a customer’s entire power problem rather than simply sell electricity. But Q2 showed that the core engine can already compound without it.

The thesis did not change.

The evidence did.

Subscribe now

• The quarter was exceptional, but the thesis has moved: Sandisk reported an 84.6% gross margin and $39.25 of adjusted EPS, yet the more important development was $93.9 billion of minimum contracted revenue at floor prices.

• Sandisk is trading upside for duration: more than half of FY2027 bits and roughly two-thirds of FY2028 bits are already committed, creating a contracted core with multi-year agreements, price protection and financial guarantees.

• The market may be using the wrong trough: current margins are not sustainable, but the old boom-and-bust framework may no longer apply. The investment case now depends less on further estimate upgrades and more on how much normalized earnings persist through the next downturn.

“We want to increase the visibility and the durability of the franchise. We want to get this kind of boom and bust out of it.”

David Goeckeler gave that answer when Cantor’s CJ Muse asked why Sandisk’s next-quarter gross-margin guide was not higher even though NAND pricing was still rising. The wording matters. A chief executive whose company had just reported an 84.6% gross margin described the mid-80s as a fair return, then immediately changed the subject from price to duration. Management is no longer trying only to extract the highest possible margin from the shortage. It is trying to make an extraordinary margin last.

The quarter was certainly extraordinary. Revenue reached $8.97 billion, gross margin was 84.6%, adjusted EPS was $39.25, and adjusted free cash flow was $5.04 billion. Datacenter revenue doubled sequentially, while Datacenter’s share of bits rose from 12% a year ago to 38%. Sandisk guided the next quarter to $10.3–10.8 billion of revenue, an 83–85% gross margin, and $44–46 of adjusted EPS. Yet the most important disclosure was not in the income statement. It was that Sandisk now has $93.9 billion of minimum contracted revenue at floor prices.

The question is no longer how high earnings can go. It is how much of them stay.

The Thesis Has Moved

I have written three pieces on Sandisk. The first argued that AI created a missing storage layer between HBM and hard drives. The second connected that layer to context memory and KV cache inside new AI systems. The third accepted that the demand thesis had been proved and shifted the debate to contracts: could Sandisk turn temporary scarcity into a higher earnings floor?

I ended that article by asking whether the gross-margin floor was 55%, 45%, or 30%. At the time, New Business Models covered just over one-third of FY2027 bits. I said coverage needed to cross 50%, Datacenter needed to become a larger part of the business, and the contracts needed enough financial protection to matter when pricing eventually weakened.

FY4Q crossed those thresholds. More than half of FY2027 bits are committed, roughly two-thirds of FY2028 bits are spoken for, QLC Stargate is shipping for revenue, and customers that signed only one quarter ago are already asking for more supply over the next three to five years. The contract mechanism is no longer a thesis. It is becoming the business.

That changes my view. I no longer think Sandisk should be analysed mainly as a memory producer enjoying the strongest cycle in its history. It is building a contracted core inside that cyclical business, and the size of that core is now large enough to change normalised earnings.

A Fair Return

Sandisk has agreements with eight Datacenter and Edge customers. Their weighted-average duration is more than four years. Pro forma remaining obligations are $91.1 billion, backed by $16.5 billion of financial guarantees. Pricing combines fixed and variable components, with floors and ceilings, while supply and demand are specified by year and quarter.

This is a different commercial relationship from the one memory companies have traditionally had. In the old model, Sandisk invested against an industry forecast and renegotiated with procurement teams every quarter. The company carried the capacity risk; the customer retained the right to wait for lower prices. In the new model, customers secure future supply by committing volume, product mix, and financial protection years ahead.

Management says NBM gross margins should remain around 80%, with some upside when market prices rise. That means Sandisk is giving up part of the final price spike in exchange for a floor. The choice is deliberate. Management is choosing duration over maximising every last dollar of the boom.

The customer relationship may matter as much as the price floor. Sandisk says the agreements include detailed forecasts of product mix and deployment timing. That pulls the company closer to the architecture decisions that determine which capacities, controllers, and performance characteristics future AI systems require. A supplier that learns about demand after the product is designed competes on price. A supplier involved while the product is being designed competes on execution and trust.

The buyback reinforces the same choice. Sandisk repurchased $4.5 billion of shares during FY4Q and has $15.5 billion of authorisation remaining. A contracted earnings base makes future cash flow easier to plan; a low memory multiple lets the company turn that cash into a larger claim on future earnings for each remaining share.

From Acceleration to Duration

Two things moved in opposite directions this quarter. Contractual duration became much stronger, while cyclical acceleration weakened.

Revenue grew 51% sequentially, but one-third of the increase came from bits and two-thirds from pricing. The next guide assumes both bit growth and only modest further price increases. Gross-margin guidance is roughly flat. The business remains at an exceptional level, but the slope is changing.

That is not just a Sandisk issue. Memory spent five quarters as the cleanest expression of AI scarcity: every report brought larger price increases, larger beats, and another estimate reset. The TMT discussion described the current phase well: the business has reached cruising speed. Across the sector, beat magnitudes are shrinking, some Asian suppliers are missing, and leadership is rotating toward other parts of the AI buildout. Part of the selling is fundamental, not merely a crowded-position unwind.

CJ Muse framed the new debate directly:

“The challenge for semi investors is moving less of a focus on margin and EPS revisions, but rather buying into the durability of this cycle.”

The rest of the call followed that question. Bernstein’s Mark Newman pressed on the lighter revenue guide. Analysts asked why margins were not rising further, whether customers could walk away, and when supply would catch demand. They were not challenging whether AI demand exists or whether the contracts are real. They were challenging management’s larger claim that those contracts have changed the cycle.

I think both sides have part of the answer. The rate of estimate improvement is peaking. The old source of stock upside—another ASP surprise followed by another guide raise—is fading. But slower acceleration is not deteriorating durability. Sandisk is building more duration just as memory supplies less momentum.

Two Books, One Income Statement

Sandisk now contains two economic books. The contracted book has multi-year purchase commitments, price floors, financial guarantees, and detailed product visibility. The uncontracted book still floats with the NAND market.

The valuation debate comes down to how those books blend when spot pricing falls. If 60% of output earns an 80% gross margin and the residual 40% falls to 25%, the blended margin is still 58%. At two-thirds coverage, it is about 62%. Revenue per bit will differ across products, so this is not an earnings forecast. It is the reason the old trough framework no longer fits.

The market is right that an 84.6% company-wide gross margin is not a sensible terminal assumption. It is wrong to assume that Sandisk must therefore return to old-cycle margins. That would require contract coverage to reverse, contract economics to fail, or both.

I would not build the thesis around management’s forecast that the NAND market approaches $500 billion in 2027. Sandisk did not show how much of that forecast comes from bits, price, or mix, and the number requires extraordinary pricing to persist. The contracts are better evidence. They show what named customers have agreed to buy, the minimum economics Sandisk expects to earn, and how much future production is already committed.

This is the updated variant perception: the next phase of the Sandisk thesis is a transfer from estimate upside to earnings duration. The company can earn less than the most aggressive FY2027 forecast and still be worth more if the remaining earnings persist for longer and the share count keeps falling.

The Test That Comes in a Downturn

There is an awkward feature to this thesis: the floor can only be proved when the cycle turns. At today’s margins, it is invisible. A downturn will initially look like the thesis is failing because spot pricing and the uncontracted book will weaken first. Only then will we see whether the contracted economics hold.

The $16.5 billion of guarantees matter, but I would not treat them as the whole enforcement mechanism. The stronger commercial deterrent may be allocation. A customer that breaks a commitment during a shortage risks losing preferred access to Sandisk’s future supply when qualified capacity is hardest to replace. That is an inference, not a disclosed contract term, but it explains why customers are willing to sign and then expand multi-year commitments. Management expects bits to remain on allocation beyond calendar 2027, while the guarantees rise in coverage relative to the remaining obligations as the agreements mature.

I do not need to wait for a downturn to decide whether the earnings floor has moved. The contract coverage, margin framework, purchase orders, and customer expansions are enough to conclude that it has. The downturn will determine the exact height of the floor, not whether one exists.

Three Futures

A three-year stock price should be based on the earnings available to investors at that point, so I use FY2030 EPS as the forward number for a 2029 valuation. The scenarios below are my estimates, not company guidance.

The bear case assumes the commercial model fails the test: customers renegotiate, variable pricing resets lower, coverage retreats, competitor capacity returns, and the uncontracted book collapses. This is a real bear case, not a softened base case.

The base case does not require Sandisk to escape the cycle. Revenue reaches $44 billion, margins normalise well below current levels, and EPS falls far below peak-cycle expectations. The return comes from applying 12 times earnings to a more dependable $151 rather than 7–8 times an apparently temporary $200.

The bull case requires contract coverage above 75%, continued customer expansions, disciplined industry supply, and aggressive buybacks. Under those conditions, Sandisk would have created a new earnings regime rather than merely a better cycle.

At roughly $1,350, the probability-weighted value is about $1,790, or close to a 10% annual return over three years. More important, the base case works without requiring the bull case to carry the argument. I would own Sandisk here.

What I Track Now

The first three items matter most. Coverage tells us how much of the business has been moved out of the quarterly market. Gross margin tells us whether the promised floor is real. The mix of bits and price tells us whether Sandisk can grow after the pricing cycle stops doing the work.

The Earnings That Stay

Management is no longer choosing between a slightly higher or lower quarterly margin. It is choosing what kind of company Sandisk becomes when the shortage ends.

The old Sandisk maximised the upcycle and accepted the bust. The new Sandisk is contracting supply, sharing risk with customers, planning products several years ahead, and using the resulting cash to shrink the share count. The peak rate of earnings improvement may be behind it. That is not the end of the thesis.

Sandisk spent five quarters proving how high earnings could go. FY4Q showed why far more of them may stay.

Subscribe now

• The margin arrived from an unexpected direction: adjusted gross margin rose 200 basis points to an all-time high even though net revenue margin was unchanged. DoorDash did not charge more; the fulfillment system became more efficient.

• The Digital Shelf remains under-monetized: DashPass is accelerating grocery adoption, advertising is already contributing to earnings, and merchant relationships contain commercial value that DoorDash has yet to quantify or fully exercise.

• Per-share conversion remains the constraint: the widening GAAP-to-adjusted gap, stock compensation, recurring legal costs and continued reinvestment mean stronger network economics may not flow cleanly to shareholders.

Two numbers from DoorDash’s second quarter define everything that follows.

Adjusted Gross Margin: 54.2%. All-time high. Up two hundred basis points year-over-year. Net Revenue Margin: 13.5%. Identical to Q2 2025. To the decimal. DoorDash expanded margins by two hundred basis points without moving its take rate a single basis point.

Adjusted EBITDA of $914 million beat consensus by 8.4% and cleared the top of management’s own guidance range. GAAP operating income was $156 million, down from $163 million a year ago on 36% revenue growth, missing consensus by 19%. The stock moved four-tenths of a percent.

In The Two Businesses Inside after Q1, I wrote that margin expansion had to come from the intent layer, because: “Marginal costs are real and irreducible. Labor scales linearly with volume.” In The Meituan Referendum after Q4, I identified the Digital Shelf, grocery-enabled CPG advertising, as the single variable determining whether DoorDash was a $150 stock or a $300 stock. Both pieces assumed the fulfillment layer was the substrate that needed to reach zero. The intent layer was where the margin lived.

That framework needs correcting. The fulfillment layer didn’t get to zero. It got cheaper. On its own. While the intent layer had only been partly exercised.

Management knows it. They broke their shareholder letter format, said so explicitly, and spent five pages arguing DoorDash is a compounding membership flywheel, organized around strategy rather than financial results. They published two cohort charts they have never shown before: Adjusted Gross Profit per MAU and DashPass penetration compounding together over five years. Ravi Inukonda volunteered that the EBITDA beat came “later in the quarter” and they didn’t have time to reinvest, pre-framing Q3 before anyone asked. Twelve analysts asked questions. Not one pushed back.

This is a management team constructing the communication infrastructure for reclassification. The cohort charts are not evidence offered in passing. They are strategy. And the sell side is buying the narrative without interrogation.

The Cost Curve

Here is the mechanism behind the 200 basis points, and the caveat that matters.

Cost of revenue, as a percentage of Marketplace GOV, fell from 6.7% in Q2 2025 to 6.4% in Q2 2026. Contribution margin went from 34.9% to 36.8%. When Doug Anmuth of JP Morgan asked about the take rate improvement, Ravi gave an answer more useful for what it ruled out:

We’re not operating the business towards take rate or net revenue margin percentage. Our goal has been always to optimize for overall profit dollars… I would think about it as flattish from Q2 to Q3, and then Q4, which is normally a quarter with higher Dasher costs, it’ll be slightly lower.

Management is guiding you not to model take rate expansion. The margin improvement came from below the revenue line.

I should be precise about what this does and does not prove. The year-over-year comparison strips out seasonality, Q2 2025 had the same seasonal Dasher cost profile as Q2 2026, and adjusted gross margin was 52.2% then versus 54.2% now. That is not a seasonal pattern. It is also not fully explained by Deliveroo consolidation, since Deliveroo was already contributing in Q4 2025 and Q1 2026 without producing a comparable gross margin expansion.

The evidence is consistent with a density-driven cost curve: better batching, routing, Dasher utilization per zone, and international unit economics converging toward the US standard. But the materials do not provide a full bridge showing how much came from structural delivery efficiency versus Deliveroo mix, advertising contribution, insurance timing, or basket composition. Management cited ads and subtotal as specific drivers and attributed take rate movement primarily to seasonal Dasher costs.

The honest read: something structural improved in the economics of fulfillment. The precise decomposition is not observable from outside. The direction of the change matters more than the magnitude of any single quarter, because if fulfillment itself has an improving cost curve at DoorDash’s density, it changes the architecture of the entire thesis.

The prior articles framed it as: fulfillment gets to zero, then advertising takes over. The actual structure may be different. If fulfillment gets cheaper over time, and advertising layers on top, the two businesses don’t merely stack. They compound. That shifts the terminal margin ceiling upward, not by the 200 basis points per year I asserted in an earlier draft (one quarter does not establish an annual rate, and cost curves flatten as efficiency improves), but by a cumulative 100-200 basis points through 2029 in a base case, with more available in a bull case where the global technology platform and autonomous delivery contribute additional gains.

Directionally, DoorDash’s delivery economics appear to be converging toward Uber’s levels on a steeper improvement trajectory, though segment definitions differ enough that a precise numerical comparison is unreliable. The conceptual point stands: DoorDash has order-volume share leadership and the relationship between density and delivery cost is non-linear. A logistics business whose unit costs decline non-linearly with scale is exhibiting economics that sit somewhere between pure logistics and platform.

DashPass Before the Digital Shelf

The most important update to the earlier thesis is one of sequencing.

The prior articles assumed the path ran: grocery reaches breakeven → CPG advertising unlocks → margins expand. Q2 suggests the mechanism starts one step earlier.

DoorDash disclosed that it added more paid U.S. DashPass members in the twelve months through Q2 than in the previous twenty-four months combined. As cohorts age, DashPass penetration rises, order rates increase, and adjusted gross profit per MAU increases. Most tellingly: DashPass members placed approximately 75% of U.S. grocery and retail orders during Q2.

DoorDash is not acquiring a restaurant consumer and then separately acquiring a grocery consumer. It is making restaurant delivery reliable enough to become habitual, converting the habitual consumer into a member, and using membership to lower the cost of trying grocery and retail. More categories make DashPass more valuable. More DashPass members make new categories easier to scale. The Digital Shelf may be the margin engine. DashPass is what keeps filling it.

There is a cost to this loop. DashPass orders carry lower gross-margin percentages because members pay lower consumer fees. Management accepts this trade when frequency, retention, and lifetime profit compensate, and Q2’s cohort charts argue it does. But the company still does not disclose the cohort-level cash economics needed to settle the question quantitatively. The direction is clear. The magnitude is taken on trust.

The Lever Nobody Asked About

If the cost curve is doing margin work on its own, the monetization layer becomes additive rather than compensatory. That changes the math.

Michael Morton of MoffettNathanson opened Q&A with the question that matters most. Some grocery partners reportedly pay near-zero take rates. What’s the repricing opportunity? Tony Xu declined to comment on any specific partner, then said this:

When you are the fastest grower in the market for them, and you are their source of growth, we might be 100% of the growth that they see, you certainly have opportunities to grow your business with them.

I don’t want to overstate what that means. “Opportunities to grow your business with them” could mean higher take rates, but it could also mean advertising, fulfillment services, data products, or expanded assortment. It is not a confirmation of unused pricing power. It is a statement that commercial bargaining position exists and management sees ways to monetize it. But it is still a CEO describing latent economic value in a relationship the market cannot observe.

The grocery thesis is advancing alongside it. DashMart warehouses run near 24/7 with what Tony described as “10x better error rates” because DoorDash controls the inventory. The blind storefront problem I identified in The Ice King, the information layer promising something the physical layer cannot reliably deliver, is being solved through vertical integration, exactly as the Tudor framework predicted. Management reaffirmed new verticals reaching gross profit positive in H2 2026. The Digital Shelf thesis isn’t wrong. It is understated, because it sits on top of a cost structure that is itself improving.

Then the silence. When Youssef Squali of Truist asked what drove the EBITDA outperformance, Ravi answered:

The unit economic improvement came in ahead of our expectations, specifically in a couple of areas. Ads was one of them, subtotal was the other one.

Twelve analysts asked questions. Not one, across MoffettNathanson, Evercore, Bernstein, JP Morgan, Morgan Stanley, Bank of America, Citi, Barclays, asked how large the advertising business is.

The advertising lever is not untouched. It is already contributing to results, Ravi said so. But it is under-disclosed. Instacart breaks out advertising separately and is valued on it. Uber discloses an advertising run-rate. DoorDash does neither, which means sell-side models assign marketplace economics to a revenue stream that likely carries much higher incremental margins. Companies segment-report when they want a business valued separately. SevenRooms venues growing 100%+ YoY, digital ordering at 150,000+ merchants with revenue up 40%+, reservations bookings up 150% sequentially, all described in prose, none tabulated.

The variant perception, stated carefully: the market appears to capitalize DoorDash’s current margin trajectory while assigning limited incremental value to a separately disclosed advertising or merchant-services segment. That segment is growing and already affecting EBITDA, but cannot be modelled from outside. The day DoorDash tabulates it, the sum-of-the-parts changes. That day hasn’t come.

The Ice Houses, Rebuilt

Two threads from the prior articles that Q2 resolved or materially advanced.

The Ice King assigned meaningful probability to a $1.5-2 billion Deliveroo impairment by 2027-2028. European food delivery had collectively destroyed $20 billion in shareholder value. Q2 showed the opposite. Deliveroo accelerated across MAUs, Total Orders, and subscription growth, all at two-year highs, while exceeding internal profit expectations and turning contribution profit positive. The global technology platform completes H1 2027; the efficiency benefits haven’t arrived yet. Management backed the thesis with capital: $1.05 billion of buybacks at roughly $154 average, the same buyback whose twelve-month absence I flagged in Q4 as a bear signal. That signal reversed.

Deliveroo has moved from probable problem to promising proof point. It has not yet been proven as a successful acquisition. DoorDash still does not disclose Deliveroo’s standalone EBITDA, cash contribution, integration spending, or return on purchase price. Contribution-profit positivity is evidence. It is not a return calculation.

The Ice King also posed the refrigeration question: would DoorDash’s infrastructure compound or be obsoleted? Justin Post of Bank of America asked about agentic AI traffic. Tony:

I think there’d be two big wars. One is the battle for attention, and you see that playing out with chat assistants. The other is the battle for atoms… Our focus is squarely on making sure that we master the physical world so we can be the most useful to all these digital assistants.

DoorDash isn’t waiting for the refrigerator. It’s building it, the Autonomous Delivery Platform orchestrating human dashers, Dot robots, DoorDash Air drones, and third-party AVs in a single routing system. Dot targeting a high single-digit percentage of Phoenix orders by year-end is the first falsifiable commitment. And Meituan’s 2025 experience, where JD.com and Alibaba destroyed sixty billion yuan of profit at full maturity, is a reminder that the cost curve, not pricing power, is the moat that survives competitive attack. DoorDash appears to be building the right one.

The Counterweight

The bears are not wrong about everything.

GAAP operating income declined 4.3% year-over-year on 36% revenue growth. The bridge from $199 million of net income to $914 million of EBITDA is $715 million, up 93% year-over-year, now 78% of the adjusted number.

Three items that need stating plainly. SBC was $349 million, 7.8% of revenue. Management “lowered” the FY guide to $1.2-1.3 billion, but H1 was $580 million, implying H2 of $620-720 million. Q2 is the run-rate, not the anomaly. The legal and regulatory add-back ran $29M, $48M, $29M, $45M, then $98M across five quarters. A charge excluded from EBITDA every quarter for years is an operating cost, and at $250-300M annualized it represents 7-8% of adjusted EBITDA. R&D grew 52% to $535 million, funding the technology platform and autonomous delivery.

Nobody on the call raised any of it.

The honest way to see this company’s earnings is through three lenses, not one. GAAP operating income was $156 million, the floor, burdened by acquisition amortization, SBC, and everything else. Adjusted operating income excluding the roughly $112 million per quarter of finite Deliveroo intangible amortization, which is non-cash and runs off on a defined schedule, but whose associated acquisition consideration was very real, was approximately $268 million, up roughly 64%. Owner earnings, after adding back the amortization but deducting normalized SBC and recurring legal costs, sits somewhere between these two. None of these views is more correct than the others. Each answers a different question.

On cash: management disclosed that FY26 reported free cash flow will be reduced $700-800 million by year-end merchant-payment timing. Clean FY26 FCF is therefore approximately $3.7-3.8 billion. The stock trades at roughly 24 times clean free cash flow, not the 30 times the headline implies. That is a useful anchor.

The GAAP-to-adjusted gap is real, and it widened this quarter. The cost curve argues it is transitional, the underlying business economics are improving even as the P&L carries acquisition and investment noise. But the gap is widening faster than the business is growing, and if the cost curve flattens while the add-backs persist, the adjusted story was always the generous one.

Three Worlds, 2029

The Q4 article’s base case was $175. The Q1 article’s was $210. FY2026 EBITDA is tracking approximately $3.73 billion, roughly matching the Q4 article’s 2027 base case. We are ahead of schedule on profitability. But the right response to being ahead of schedule is to revise the estimates, not to assume the rate continues.

A margin bridge makes the assumptions auditable:

The reinvestment deduction is where intellectual honesty lives. DoorDash’s reward for proving one investment works is permission to begin three more. Management keeps finding attractive areas to fund, grocery infrastructure, autonomous delivery, merchant software, international expansion. The company may grow for longer than expected while converting less of that growth into per-share free cash flow than expected. That tension must be in the model, not outside it.

Share counts assume continued buyback at roughly the current pace (~$1.5B annually) partially offsetting SBC dilution of 8-10 million shares per year. Net cash of $8-12 billion by 2029 (varying by scenario) is added to the EV-to-equity bridge. These are approximations, not a full equity model, but the arithmetic is visible.

At approximately $208, the stock prices the lower half of the base case against a probability-weighted value of roughly $260. That gap is real but not extreme, roughly 25%, and the catalysts are defined: Q3 EBITDA in November, advertising disclosure at any point, and Deliveroo’s continued trajectory.

The updated BFQ. The prior articles asked whether DoorDash was a logistics company or a media company. That question has partially resolved. The new question: How long does the fulfillment cost curve run before it flattens, how much under-disclosed monetization remains above it, and how much of both reaches each share rather than funding the next layer of expansion?

The first two parts are bullish inputs. The third is the constraint that keeps this from being a simple call.

The variant stated carefully. Consensus sees a delivery company in an investment cycle that will eventually produce margin expansion. The variant: the margin expansion has already begun in the cost curve, and the monetization layer, advertising, grocery commercial terms, merchant services, is under-disclosed and only partly exercised. The market is pricing the margin DoorDash has earned. The less visible question is how much additional margin exists that DoorDash has not yet chosen to show.

What Resolves This

Six signposts, each with a threshold.

Q3 Adjusted EBITDA (November). Above $1,060M strengthens the case that Q2’s improvement was structural. Below $990M suggests it was timing and reinvestment deferral.

Q3 Adjusted Gross Margin. At or above 54% means the year-over-year expansion holds on a different seasonal base. Below 52.5% means the curve may have flattened or Q2 was mix-driven.

Advertising disclosure. Any quarter, any format. The single cleanest re-rating catalyst, the moment the second business becomes defensible in numbers rather than prose.

Organic GOV growth. Above 24% reverses the Q1-to-Q2 deceleration. Below 21% compresses the multiple regardless of margin.

Dot in Phoenix by year-end. Hit means the autonomy R&D is a funded option with observable progress. Miss or silence means discount the autonomous delivery narrative.

Legal/regulatory add-back. Below $55M in Q3 means the $98M was discrete. Above $80M again means treating it as a permanent operating cost and haircutting EBITDA accordingly.

The ice houses are still useful. Q2 showed us something the prior articles didn’t expect: the ice is getting cheaper to cut even before Tudor raises the price. The DashPass flywheel is filling the shelves. The advertising layer is contributing before being disclosed. Deliveroo is improving before the platform rebuild is complete.

The compounding mechanism appears real. The remaining question, the one that keeps this from being a simple buy, is how much of that compounding reaches each share, and how much funds the next ambition management decides to pursue. DoorDash has begun proving the network generates more value as it grows. It has not yet proved how much of that value it intends to keep

Subscribe now

• The flywheel is working: Brazil items sold rose 56%, ecosystemic users grew 37%, credit quality improved, and operating expenses gained 260 basis points of sequential leverage.

• Shareholders still cannot see the payoff: management reinvested all the operating leverage into lower prices and incentives, while declining to disclose the ecosystemic-user penetration and economics needed to value the strategy.

• The debate is now about time: MercadoLibre appears to be building a more valuable integrated system, but the stock needs evidence that customer value will eventually become reported margins and per-share earnings.

From MercadoLibre’s Q2 2026 shareholder letter:

Contribution profit per ecosystemic user is multiples of the sum of the profit of a marketplace-only user and a fintech-only user. Profitability doesn’t simply add when a user becomes ecosystemic, it multiplies.

A year ago, we called the free shipping threshold cut a strategic gambit, a victory lap disguised as a skirmish. Since then: four articles, one consistent thesis, and a stock that went from roughly $2,700 to $1,888. Down 30%.

The commerce call was right. Every signpost cleared. Brazil items sold grew 56% year-on-year. Unit shipping costs kept falling. The managed network hit 96% penetration. Ecosystemic user growth in Brazil accelerated from 35% before the threshold change to almost 50%. The logistics density thesis played out exactly as we described in the first article, then the second, then the third.

The credit call was wrong in Q4 and Q1. We underestimated how fast a $12.5 billion lending operation would change how the market categorized the entire business. We wrote about the logistics war while the credit transformation happened in peripheral vision. Q2 materially weakened the immediate bear case, NIMAL improved 290 basis points, NPLs dropped to historic lows, but the damage to the stock was already done.

The margin call was half right and half wrong. We said operating leverage would emerge. It did, opex improved 260 basis points sequentially this quarter. We assumed management would show some of it. They showed none. They reinvested every dollar.

The hard lesson, five quarters in: being right about the flywheel was necessary but not sufficient. The market doesn’t pay for flywheels it can’t see in the P&L. What we underestimated, across every article, was that management would reinvest 100% of the operating leverage, every quarter, with no signal of when it ends. The gap between the business and the stock is not analytical error. It is a time horizon mismatch between a management team playing a decade-long game and a market that prices in quarters.

Q2 doesn’t close that gap. But it gives us, for the first time, the economic unit that might explain what’s on the other side of it.

The Surplus Question Gets a Candidate Answer

Our Q1 article asked the question that matters: who captures the surplus MercadoLibre is creating? The company was producing enormous value, cheaper shipping, faster delivery, broader selection, easier credit, better financial tools, but the stock case required showing that value accrued to shareholders after subsidies, provisions, capex, and funding costs.

Q2 gives that question a candidate mechanism. Not yet a complete answer. A mechanism.

The concept is the ecosystemic user, someone who engages with both the marketplace and Mercado Pago. Management says these users generate 70% more GMV than marketplace-only users. They buy 55% more items. Their payment volume is 90% higher. Their assets under management are more than double. And their contribution profit is not additive but multiplicative, multiples of the combined profit of a marketplace-only and fintech-only user.

These users are growing at 37% year-on-year, the fastest-growing segment since Q4 2023, and their share of the total user base has expanded by eight percentage points over that period. Credit card holders are two to three times more likely to become ecosystemic than users without the card. MELI+ subscribers grew 72%.

Here is what connects every investment we’ve written about across four articles. The free shipping threshold converts marketplace-only buyers into more frequent buyers who adopt Mercado Pago. The credit card converts Pago users into marketplace shoppers. PIX discounts drive payment adoption. 1P and cross-border trade expand selection that pulls users deeper into the ecosystem. MELI+ bundles it explicitly. These are not separate initiatives with separate ROI calculations. They are different doors into the same room.

The argument is credible. The disclosure is incomplete. MercadoLibre did not tell us what percentage of users are ecosystemic. It did not disclose absolute contribution profit, conversion cost, or the time required to recover that cost. Management gave us the uplift because it supports the story. It withheld the denominator that would let us value it. When BTIG’s Marvin Fong asked directly for the penetration number, Martin de los Santos declined:

We don’t disclose the actual share of users. We just want to make sure that those are important users to us, growing very fast.

Management wants the concept in the narrative without the number that constrains it. That tells you the ecosystemic user may be MercadoLibre’s real economic unit, but investors still cannot build an ecosystemic-user model.

Three things need to be separated. Q2 provides strong evidence that the system creates more user value. It provides meaningful evidence that the system creates more company-level economic value. It provides incomplete evidence that this value will accrue to shareholders within a reasonable period. The article, and the investment case, rests on preserving that hierarchy.

Management Is Selling the Ecosystem. The Market Is Pricing the Balance Sheet.

Management spent the entire shareholder letter and earnings call selling this reframing. CEO Ariel Szarfsztejn closed the call with the thesis stated as plainly as a management team will ever state it:

We are building something quite unique globally, an ecosystem of commerce and financial services that compounds on itself. These are not two businesses running in parallel. This is one flywheel, each side making the other one more valuable.

No timeline. No margin target. No guidance the market can model against.

The questions on the call revealed where the disagreement sits. Goldman’s Irma Sgarz led with “feel obliged to ask about margin” and pushed on gross margin compression sources. Cantor asked whether current margin levels were sustainable through the second half. Citi pressed on credit card NIMAL inflection. UBS challenged the 90+ day NPL metric. BTIG fished for the ecosystemic penetration number.

The questions split into three camps: analysts exploring the flywheel and its upside, analysts testing the timing of margin realisation, and analysts still focused on credit risk. The second group dominated the call. Nobody disputed that engagement improved. Everyone wanted the bridge from engagement to earnings.

Management’s answer was consistent: credit profitability recovered, opex scaled, and management then reinvested those benefits into commerce. That is exactly the pattern we’ve tracked for five quarters. The analysts accept the evidence that the investments work. They are not willing to pay today for profits management has neither quantified nor scheduled.

“Elected”

Here is the section that matters for the stock.

Product development expense scaled from 8.4% of revenue to 7.2% year-on-year. AI productivity drove it: code submissions up 110%, engineering headcount flat for the first time in years, more than half of Help Portal requests resolved by the AI assistant without human intervention. Provisions flat quarter-on-quarter despite credit portfolio growing 75%. Total opex improved 260 basis points sequentially. The operating leverage we spent four articles waiting for showed up in the numbers.

Then management launched new PIX buyer discounts and seller take rate cuts in Brazil. In April. During the quarter. Gross margin compressed 280 basis points sequentially, consuming every dollar of opex improvement. Martin de los Santos on the call:

We elected to reinvest that margin into other areas of the business.

Elected. He used the word three times.

The business gets stronger. The margin doesn’t move. The stock sits.

Our Q1 article asked whether the dial has a minimum setting that competition enforces. Q2 doesn’t fully answer that. Whether the new Brazil initiatives were proactive or reactive doesn’t determine whether they’re discretionary in an economic sense. Management may invest proactively because it sees competitive dynamics, Shopee in low-ASP, Amazon in logistics, Nubank in financial services, that make it necessary even without an immediate threat. The earlier framework remains the honest one: management controls the intensity, but competition and customer expectations influence the minimum.

What Q2 does clarify is that the leverage mechanism works. The opex lines proved it. If management ever chose to let two or three quarters of opex improvement flow through without launching a new initiative, operating margin would move above 8% rapidly. They have not chosen to. There is no external catalyst that forces them to. It comes when they decide, and they have given no signal of when that will be.

Behind the reinvestment decision, though, something structural is building that deserves attention. The AI-driven productivity gains, Product Development at 7.2% and falling, customer service running on 30% fewer humans while the business tripled, 2026 the first year engineering isn’t growing, represent 100-200 basis points of annual operating leverage that compounds regardless of management’s commerce and fintech investment decisions. This happens in the background. Over three years, it could contribute meaningfully to margin recovery even if management continues to invest aggressively elsewhere. The market is not modelling this because it is distributed across multiple line items. It may be the most underappreciated structural feature of the entire P&L.

Credit Recovered. The Clock Did Not.

The Q4 article’s $12.5 billion identity crisis, the market re-categorizing MELI from tech platform to Latin American lender, has a materially better data set now. Not a resolution. A better data set.

NIMAL recovered from 17.8% to 20.7%. The 15-90 day NPL dropped from 8.0% to 7.0%, near historic lows. Credit card NPL at 4.6%. Provisions flat quarter-on-quarter. Credit revenue net of provisions as a share of revenue higher than a year ago. The portfolio is scaling profitably at the current level of growth.

Our Q4 signpost said NIMAL must hold above 22%. It didn’t. The miss is from credit card mix, now 47% of the portfolio at -2.5% NIMAL because issuance accelerated to 2.6 million cards per quarter. Each cohort breaks even in 12-18 months, but every wave of new issuance resets the average age of the book. The portfolio is majority young. Recalibrate: watch NIMAL trending toward 22% as issuance growth moderates and cohorts season.

The immediate bear case, credit growth outrunning underwriting, weakened materially this quarter. The structural credit debate, whether the portfolio can survive a full Brazilian downturn at this scale, did not resolve. UBS correctly flagged the 90+ day NPL uptick. Management attributed it to technical factors and product mix. Q3 will confirm whether it was noise or signal.

Our View, Updated

We were right that logistics density could be used offensively. We were right that integrated commerce, payments, credit, and data create a system competitors cannot easily replicate. We were right that Engine A, advertising growing 62% FX-neutral and now above 10% of Latin American digital ad share, AI productivity compounding in every operating line, would eventually become visible.

We underestimated how quickly credit would change the company’s financial identity. We underestimated how long the investment cycle could persist. Most of all, we treated value creation as if it would automatically become shareholder value. The Q1 article corrected that by asking who captures the surplus. Q2 gives the surplus question a named mechanism, the ecosystemic user, without yet proving how much of the value reaches the bottom line.

Our variant view is narrower than it was a year ago: MercadoLibre is becoming a more valuable integrated system even while reported margins remain low, and Q2 suggests the cost of building that system may be temporary rather than permanent. The missing variable is the meaning of “temporary.”

The bear case does not require MercadoLibre to lose. It requires the surplus to keep flowing to customers, sellers, and credit losses while margins stay near current levels. The base case assumes the ecosystem keeps compounding, credit seasons normally, and the cost of defending the system is higher than we once assumed but manageable. The bull case is management’s story: ecosystemic users become a large share of the base, credit cards turn profitable, advertising and AI change the margin structure, and the company gets re-categorized as integrated digital infrastructure.

Signposts graded and updated: Brazil FXN GMV above 35%, cleared at 39%, maintain. NIMAL trending toward 22%, below at 20.7%, recalibrate to “watch as issuance moderates.” 15-90 NPL below 8%, cleared at 7.0%. Operating margin above 7.5% as inflection signal, below 6.5% as stress. Advertising FXN growth above 50%, cleared at 62%, watch for quarterly revenue approaching $500 million. New: ecosystemic user penetration disclosure, when management starts reporting the number, it means they’re confident in the curve. New: 90+ day NPL below 18% confirms noise, above 20% reengages concern.

Five quarters. Five articles. The commerce thesis was right. The credit thesis scared us and recovered. The surplus question now has a named mechanism, if not yet a proven one. The moat is wider than it has ever been.

And the stock is down 30% since we started writing.

MercadoLibre is building something that does not exist anywhere else, an integrated commerce and financial services flywheel in a market where both digital shifts are happening simultaneously to 650 million people. The flywheel works. Engagement compounds. Q2 suggests the economics may follow. Management still has to let them appear.

The price of the flywheel is time. And time, for now, is what management is buying, whether the market likes it or not.

Subscribe now

• The operating thesis strengthened: bookings grew 22% in constant currency, profit grew faster, AV utilization is improving, and Uber’s market position remains resilient in mature autonomous-vehicle cities.

• The valuation thesis weakened: Uber still withheld AV take rates and comparable Uber One metrics, while committing more than $10 billion to help build the autonomous supplier ecosystem.

• The key question has shifted: it is no longer whether Uber matters in an autonomous world, but how much capital it must contribute and how much cash ultimately remains for shareholders.

In May I wrote that Uber had moved from being priced as an autonomous-vehicle casualty in the high $60s to being priced as a business with a plausible coordination thesis at $80. I thought the remaining distance to my $120–$130 base case depended on Uber One becoming more deeply embedded, Delivery margins reaching 4%, and insurance savings accelerating U.S. Mobility.

Three months later, Uber delivered the strongest evidence for that thesis yet, and the stock returned to roughly $68. The repricing did not merely stall; it reversed.

That does not mean the original thesis was wrong. It means my explanation for what would move the stock was incomplete.

I had assumed Uber would rerate as management proved that autonomous fleets need its demand. Q2 suggests that proof is no longer sufficient. Uber is now providing increasingly precise evidence of what preserving that position will cost, while still withholding the numbers that show what it will earn.

The Scorecard

The honest place to begin is with the tests I set before the quarter.

After Q1, I wrote:

“If Uber reports Q2 without disclosing AV take rates, I will interpret that as a signal the economics are not yet where management wants them.”

Uber did not disclose the take rate. When Bank of America’s Justin Post asked how the more-than-$10-billion autonomous commitment would affect earnings and margins, Balaji Krishnamurthy said Uber would provide more visibility as deployment approached scale.

I do not think that answer means the economics are necessarily poor. The commercial arrangements are early, likely vary by partner, and may change as vehicle volumes grow. But I committed in advance to what continued silence would mean. I do not get to rewrite the test after seeing encouraging evidence elsewhere.

Uber One created a second disclosure problem. Q1’s defining number was fifty million members generating half of total Gross Bookings across Mobility and Delivery. In Q2, management said membership reached another record and that members generated more than 70% of Delivery Gross Bookings, but it did not update the member count or the share of total bookings.

The new statistic may be useful. It is not an update to the old one. It changes both the metric and the denominator. One quarter may be harmless; a second omission would look deliberate.

The operational evidence was almost the mirror image. Gross Bookings reached $58.0 billion and grew 22% in constant currency. Adjusted EBITDA increased 33%, non-GAAP operating income rose 40%, Delivery bookings grew 25%, and Delivery operating income grew 38%.

Insurance savings passed through lower prices produced faster trip growth in Los Angeles and San Francisco. Autonomous partners were completing trips in the high twenties to low thirties per vehicle per day. Uber’s category position in Los Angeles, San Francisco, and Phoenix; the mature U.S. AV markets, was higher than a year earlier.

The pattern is hard to dismiss: the operating tests strengthened, while the two disclosures needed to value that strength failed.

The Proof

The original Uber thesis began with utilization.

Human drivers provide their own vehicles and decide when to work. When demand falls, they turn off the app. An autonomous vehicle is a fixed asset that depreciates whether it is carrying a passenger or sitting idle. Its owner must cover charging, cleaning, maintenance, insurance, remote assistance, and financing before earning a return.

That makes dense demand more valuable after autonomy arrives, not less.

Q2 supplied the best evidence yet that Uber can provide it. AVs are now live on Uber in seven cities, with as many as fifteen planned by year-end. Partners have committed approximately 120,000 vehicles over the coming years, and management says Uber’s position has strengthened rather than weakened in the cities where autonomous competition is most advanced.

AV Labs adds another layer. Uber is operating sensor-equipped vehicles that collect rideshare-specific driving data for its partners. Combined with demand, dispatch, fleet operations, insurance, financing, charging, and local regulatory work, Uber is becoming an input into autonomous commercialization rather than merely another app through which a completed product can be booked.

The strategic argument is much stronger than it was eighteen months ago. Uber will almost certainly matter in an autonomous world.

And yet relevance is not ownership.

The Invoice

In 2005 Apple prepaid $1.25 billion to five memory suppliers to secure NAND flash for the iPod. Apple was not trying to manufacture semiconductors; it was using the strength of a customer relationship it already controlled to make sure scarce supply expanded around it.

Uber is attempting something similar, with one dangerous difference. Apple’s memory suppliers did not aspire to own the iPod customer. Uber’s autonomous partners may eventually want the rider, the app, and the entire trip economics.

Dara Khosrowshahi’s desired market structure is clear: several AV developers compete across technologies, vehicles, and regions; each benefits from Uber’s demand and local operating capabilities; and no single provider gains enough power to dictate commercial terms.

I had treated that fragmentation as the likely outcome of the industry. Q2 suggests Uber may have to help finance it into existence.

Uber expects to commit more than $10 billion across equity investments, infrastructure, and vehicle offtake arrangements. It is helping software partners raise capital, anchoring vehicle demand, collecting shared data, and building the capabilities needed to launch fleets city by city. Management intends to bring in outside financiers as deployments scale.

The intended loop is compelling. Uber’s recurring demand raises vehicle utilization; better utilization attracts suppliers and outside capital; more suppliers reduce Uber’s dependence on Waymo or any other developer; better coverage and price produce more demand.

But the loop begins with Uber’s balance sheet.

That creates two risks. The first is that Uber supports weaker suppliers that never become competitive with the leading autonomous systems. Lucid can have PIF backing, Nuro software, and a large Uber order without proving that it can manufacture the right vehicle at the right cost.

The second is that successful suppliers use Uber’s capital, data, and demand to reach scale before seeking better economics or distributing directly. Uber may help create the competition it needs and still fail to own the value that competition produces.

The non-obvious insight from Q2 is that Uber is using cash from its capital-light business to create the supplier market its capital-light future requires.

Where the Cost Appears

Uber does not disclose a clean bridge from AV investment to current operating expenses, so it would be wrong to treat every increase in corporate costs as autonomous spending. The consolidated results still show where the broader ambition first meets the income statement.

Source: Uber Q2 2026 earnings release.

Mobility and Delivery generated $775 million of incremental operating income, while corporate costs absorbed $168 million. That line contains shared finance, legal, technology, mapping, payments, and other functions, so it cannot be assigned solely to AV Labs, partnerships, or M&A. It is still the visible line through which some of the expanding corporate agenda reaches shareholders.

The Q3 guide made the tension more immediate. Uber guided to non-GAAP EPS of $0.84–$0.88, versus an analyst expectation of approximately $0.89, while its bookings midpoint was broadly in line with the Street. The quarter proved the strategic case but did not create near-term estimate upside.

Capital allocation told the same story. Uber generated $10.1 billion of trailing free cash flow, but deployed approximately $4 billion to establish a 37% economic interest in Delivery Hero and repurchased only $510 million of its own stock.

Reported free cash flow measures what the operating business generates. It does not measure what remains available to owners after acquisitions, strategic equity investments, and AV commitments.

Management is winning the strategic argument and losing the multiple because the costs can be placed into a model today, while the returns remain a distant assumption.

What the Next Year Can and Cannot Tell Us

The next few quarters cannot determine whether Uber’s AV strategy creates value. The vehicles, financing structures, supplier relationships, and commercial agreements will take years to mature. Uber itself has said meaningful commercialization will take much longer than the underlying technical progress.

Corporate costs moderating in Q3 would not prove that the $10 billion earns an attractive return. A restored buyback would not prove that outside financiers will eventually absorb the vehicle risk. Even a disclosed take rate would tell us little without knowing what Uber spent or guaranteed to obtain it.

What the next year can show is whether Uber can carry the build without weakening the business that funds it.

Through early 2027, the test is whether Mobility and Delivery continue producing bookings growth near 20%, profit grows faster than bookings, central costs become more controlled, leverage remains bounded, and buybacks recover. That would show that management can incubate a long-duration project without sacrificing current per-share compounding.

During 2027, the test shifts toward risk transfer. Uber needs to demonstrate scaled third-party vehicle financing, clarify which of the 120,000 commitments are firm or conditional, and show how much capital or guarantee exposure remains with Uber.

Only from 2028 onward can the final economic question begin to resolve: AV gross profit, take rate after all support, return on Uber capital, utilization across several suppliers, and whether those suppliers remain on Uber after reaching scale.

The next few quarters do not resolve the long-term thesis. They determine whether Uber earns the right to keep pursuing it.

Three Futures

These scenarios use my assumptions from the FY2026 earnings base:

The bear case is not Uber disappearing. The company remains relevant, but central costs stay elevated, Delivery Hero consumes capital through its integration, suppliers gain bargaining power, and buybacks become residual. Earnings rise, but the multiple remains compressed because strategic importance never becomes distributable cash.

The base case assumes that Mobility and Delivery continue compounding, Uber contains the incubation cost, and third-party financing gradually takes over much of the vehicle funding. AVs remain too small to drive the consolidated model, but they no longer justify a large displacement discount.

The bull case requires several competitive AV suppliers, visible economics above a 20% take rate, limited retained fleet risk, and a core business strong enough to fund investment while reducing the share count.

The probability-weighted value is approximately $112. The wide range does not mainly reflect uncertainty about how many robotaxi trips Uber completes. It reflects uncertainty over how much capital Uber must leave behind and how much cash per share remains for owners.

I still believe the coordination thesis. Q2 made it stronger. What changed is my understanding of why the stock has not responded.

I expected Uber to rerate when it proved that autonomous vehicles would need its demand. Instead, each new piece of proof arrived with a more visible invoice and no disclosed toll.

Uber is building a long-duration asset whose value will not be known for at least another year or two. The next few quarters will not tell us whether that asset works. They will tell us whether Uber can afford to wait.

Uber has proved that it can make autonomous vehicles busy. It has not yet proved that it can stop paying once they are.

Subscribe now

• The core continues to compound: Toast added a record 9,500 net locations, recurring gross profit reached $595 million, and monetisation rose to 98 basis points without relying on higher pricing.

• AI now has revenue attached: Toast IQ Grow is on track to become the company’s fastest product to $10 million in ARR, using transaction-level data to identify opportunities, launch campaigns, and measure the orders they generate.

• The economics remain unproved: Early demand and customer outcomes are encouraging, but Toast has not disclosed paid accounts, retention, service intensity, or gross margins, the evidence needed to show that Grow can scale like software rather than a managed service.

One Idea, Several Names

For eighteen months our argument has been the same. Financial technology accounts for most of reported revenue, so the market files Toast alongside Square and Clover and applies a payment multiple. That is the wrong lens, because no restaurant selects Toast for its processing rates. They select it because Toast runs the business, and payments flow through because of owning the transaction layer. Toast is a vertically integrated restaurant operating system with payments as the monetization rail.

We gave that one idea three names across four articles, the integration layer, the profit sanctuary, the invisible manager, which was a writing habit rather than a change of mind. Two things in the record deserve correcting.

The first is timing. For six quarters we pointed at Toast IQ adoption figures and treated usage as evidence the AI layer was arriving. Usage is not revenue, and a market that declines to underwrite what it cannot measure was behaving sensibly. We were describing a mechanism and asking to be paid for optionality.

The second matters more. We consistently defended Toast’s decision to reinvest rather than harvest, distinguishing chosen lower margins from structurally lower margins. That distinction holds. But we treated the act of reinvestment as though it were evidence of value creation, and it is not. Investment is not the conclusion. Return on investment is. Toast still discloses no acquisition cost, payback period, or contribution margin for any business it is funding.

Q2 does not require a new identity for Toast. It gives better evidence about the company we have been studying since the start.

The System Still Compounds

Toast added a record 9,500 net locations, taking the base to roughly 180,000, with the majority from core SMB and mid-market rather than a few large enterprise deployments. GPV reached $60.7 billion and recurring gross profit $595 million.

Location count measures distribution, ARR measures monetization, and recurring gross profit measures how much Toast actually captures from that distribution. All three rose, and gross profit rose fastest. GPV per location was flat after falling 1% in the fourth quarter, SaaS gross margin expanded roughly 240 basis points, and total take rate, recurring gross profit over GPV, reached 98 basis points, up five year over year. Elena Gomez attributed the take rate gain to product adoption and cost optimization rather than pricing, which makes it durable. Toast did not sustain growth by adding visibly weaker customers.

One detail sits awkwardly. Net adds beat expectations by more than a third while ARR landed at consensus, which raises the question of why such a large location surprise produced no monetization surprise. New locations ramp over time, many may have gone live late in the quarter, and enterprise or international deployments may carry lower initial ARPU, but Toast disclosed no cohort data to settle it. Worth tracking rather than a thesis break.

The Closed Loop, and What It Hasn’t Proven

Toast IQ Grow is a marketing product built on restaurant data. Management said it is on track to be the fastest product in company history to $10 million in ARR, that it already runs at positive gross margins, and that one restaurant cut agency spending 70% while generating over $100,000 in marketing-attributed sales in under two months.

What matters is not that Toast built an AI marketing tool. What matters is that it can close the loop. A standalone vendor can produce a competent email. Toast can observe that Tuesday nights are underused, identify guests who stopped returning, select an offer with attractive margin, deliver the campaign, process the resulting orders, and verify whether incremental sales occurred. Narang put it directly:

“Because we power the point-of-sale, we can close the loop with each campaign tied directly to the orders it drives.”

The model is not the scarce asset, since foundation models are available to everyone. The harder combination is context, distribution, workflow access, and transaction-level measurement inside one system. This is why AI most likely increases the value of Toast’s existing position rather than creating a different company.

The prize is adjacent spending. Restaurants outsource marketing, payroll, bookkeeping, tax, and scheduling, and as Narang noted, they are often “spending a multiple of what they spend on software” on those functions. Toast’s total ARR works out to roughly $13,400 per location. It does not need to capture all of the surrounding budget for the economics to change; a modest share would materially raise revenue per location. How much is technologically addressable and economically capturable, this quarter did not answer.

Which is the discipline required here. Three standards of proof apply. Demand proof asks whether restaurants will buy, and Grow’s trajectory suggests some will. Outcome proof asks whether the product creates measurable value, and the early figures suggest it can. Economic proof asks whether Toast can deliver at scalable margins, and that remains entirely open.

Bernstein’s Harshita Rawat pressed on exactly this, given the marketing success managers who review AI-generated work. Narang answered:

“As we’ve gotten some early scale, we’ve already seen the gross margins improve. Frankly, I have no concerns long term about what the gross margins of that business could be.”

Confident, possibly right, not a disclosed unit-economic model. Toast provided no paid customer count, no conversion rate, no ARPU, no retention, no current gross margin, and no ratio of accounts to marketing managers. A human-in-the-loop service can show improving margins at small scale and still settle far below software economics, and that gap is most of the upside case.

One further qualification to something we have argued before: more locations do not automatically make the agents smarter. They expand distribution and may enrich the benchmark set, but only if Toast can standardize patterns across restaurants that differ enormously in cuisine, format, and check size. A new taqueria improves the dataset for taquerias. Whether it improves the agent serving a steakhouse is an engineering question, not an arithmetic one.

The Same Disagreement, Third Quarter Running

Toast beat on revenue, recurring gross profit, EBITDA, GPV, and location adds, and the stock fell 1.3%. Adjusted EBITDA of $221 million included roughly $10 million from a tariff refund, leaving underlying EBITDA near $211 million against consensus of $195.7 million. Full-year EBITDA guidance rose to $805–825 million, a $15 million midpoint raise against a $25 million beat. Gomez explained the gap without hedging:

“We strategically chose to reinvest the tariff refund into key growth initiatives and to seed long-term bets. As a result, we’re increasing our full-year adjusted EBITDA guidance by less than the Q2 beat.”

Close to close, Q4 2025 took the stock down about 7%, Q1 2026 down 14.7%, and this quarter down 1.3%. Three consecutive quarters of strong results, guidance withheld, negative reaction. This is not a communications failure that better investor relations can repair. Management is optimizing for the number of durable businesses it can build; the marginal shareholder wants evidence core profits will reach consolidated earnings. The disagreement ends only when the newer investments produce enough gross profit to make their returns visible.

Q2 moved that forward without settling it. New market ARR is near $200 million and doubling, retail SaaS ARPU is approaching core levels within two years of launch, and Grow has revenue attached rather than usage statistics. Against that, free cash flow fell to $130 million from $208 million as inventories rose to $217 million from $114 million at year-end, with Toast securing components ahead of memory-market pressure. That may be the right trade, but the cash has left the building and the explanation now carries a deadline.

A note on something we over-weighted. Last quarter Toast Local had its own section as the answer to DoorDash’s point-of-sale ambitions. This quarter it appeared once, briefly, at the end of the final answer. The demand-side thesis is not dead, but the evidence is much weaker than on the merchant side and we should weight it accordingly. DoorDash remains the real counterpoint: Toast knows the restaurant better, aggregators own more consumer intent. Toast needs enough connection to the guest to avoid becoming a back-office utility while someone else controls demand.

Three Outcomes for the Same Company

None of these requires Toast to become a different business. In each it remains a restaurant operating system, and the only question is how deeply it monetizes that position. From roughly $32, with about 580 million diluted shares and $1.7 billion of net cash, looking to FY2028.

The margin trap (20%). Restaurant activity weakens, GPV per location turns negative, Toast Capital losses rise, and Grow finds adoption but settles into managed-service economics. The core stays valuable while reinvestment fails to earn its cost. Revenue compounds at 12–14%, margins near 11–12%, and at 10–12 times EBITDA that supports roughly $22–28.

The quiet compounder (55%). Toast keeps taking share, attachment rises, enterprise and international scale steadily, retail becomes a credible adjacency, and AI improves support costs and retention without becoming a separate profit pool. Revenue compounds at 17–19%, margins reach 15–16%, and at 16–18 times EBITDA that supports roughly $45–57.

The expanding wallet (25%). Grow proves repeatable, a second agent succeeds, human review per account falls, and restaurants shift part of their outsourced spending onto the platform. Total monetization moves from 98 basis points toward 120. Revenue compounds at 21–23%, margins reach 18–20%, and at 20–22 times EBITDA that supports roughly $72–88.

The base case has sat in the $45–57 range across three articles while the stock traded $37, then $26, then $24, and now $32. The destination has not moved. What changed is that the upside now has a mechanism attached rather than an aspiration.

What Would Change Our Mind

Grow above roughly $50 million in ARR by mid-2027, without service headcount rising at the same rate. A second agent shipping inside four quarters. Disclosure of paid accounts, attach rate, or gross margin, usage anecdotes have stopped being sufficient.

Recurring gross profit growth staying several points above GPV growth, with the spread coming from subscription adoption and paid services rather than pricing or credit.

GPV per location positive by the fourth quarter, once the World Cup benefit washes out. Cash conversion back toward 70% of EBITDA in the second half. Toast Capital credit losses below roughly $30 million a quarter.

And the core holding above 20% growth at 40% margins, because that is what funds everything else.

Our view has not changed. Toast is a vertically integrated restaurant operating system with payments as the monetization rail, and workflow ownership is what drives distribution, attachment, retention, and data context. Enterprise, international, retail, Toast Local, and agentic products are extensions of that architecture, not quarterly reinventions of the company. Q2 strengthened the core thesis and produced the first measurable evidence that the same position can be used to perform work restaurants currently outsource, which is not a new definition of Toast, but a larger answer to the question we have been asking from the start: how much can be built on top of a workflow Toast already owns?

The register was the wedge, and the operating system is the moat. The opportunity now is making that operating system worth more to every restaurant that joins it.

Subscribe now

• The transaction is getting richer: Revenue grew 8% against room-night growth of 5.3%, extending an eight-quarter pattern in which monetisation has outpaced volume as merchant mix has risen to 73%.

• The acquisition treadmill remains: Marketing expense grew faster than revenue, marketing intensity increased to 4.7% of gross bookings, and direct mix remained stuck in the mid-60%s despite greater Genius and app engagement.

• The decisive test is the second booking: Booking has shown it can extract more value from the trip it executes today. The moat will be proven only when that trip lowers the cost of acquiring the customer tomorrow.

From Booking Holdings’ Q2 earnings call:

Room nights, gross bookings, revenue, and adjusted EBITDA all exceeded the high end of our guidance for the second quarter... Overall, if I also may point you to our full-year guidance, actually, if you look at our full-year guidance, we are at a high single-digit level for gross bookings and revenues and at mid-teens level for EPS at the high end. That is still in line with our long-term algorithm and is also in line at a constant currency basis with our original guidance for the year, even though in our assumptions we have seven out of twelve-month impact of the Middle East.

That last sentence from CFO Ewout Steenbergen is the quarter. Booking absorbed seven months of a regional war, a war that closed the Strait of Hormuz, grounded flights across the Gulf, and pushed airfares up across every long-haul corridor, and held its original full-year guidance at constant currency. Not a trimmed guide with an asterisk. The original algorithm.

The numbers underneath: room nights grew 5.3% to 325 million, gross bookings rose 9% to $51 billion, revenue grew 8% to $7.35 billion, adjusted EBITDA grew 9% with margins expanding 40 basis points to 36%. Every metric above the high end of guidance.

And yet the composition is what caught my attention. Room nights missed the street, consensus was around 329 million, while revenue beat it by $167 million. Fewer nights, more money. That gap, roughly three percentage points between revenue growth and volume growth, is where this quarter’s real argument lives.

In my Q1 piece I set a threshold I refused to explain away: if room nights missed even the lowered 2-4% guide, the thesis needed revision rather than another external excuse. They came in at 5.3%. The threshold held. I’ll also note, because the series deserves the honesty, that I’ve overreached before, the “fintech disguised as travel” framing, the Amazon analogy, the suggestion that Booking’s payment rails could escape travel entirely. I’m retiring that. What survived the pruning is narrower and better: Booking is moving from discovery, which AI is commoditising, toward execution, payments, settlement, refunds, service, coordination, which it is not. The question was always whether the execution layer captures its own economics. Q2 gave evidence on both sides.

The Spread

Start with the side that worked.

The three-point spread between revenue growth and room night growth needs careful handling, because I’ve learned the cost of assigning too much causality to one quarter. Walk the bridge: constant-currency ADRs rose about 2%, driven by genuine pricing in Europe and the US, management was explicit this wasn’t a mix artifact. FX added roughly a point. Flights and other verticals contributed. Revenue timing from March’s elevated cancellations added noise. Those factors get you close to the full spread without invoking merchant economics at all.

So, the single quarter proves less than I’d like. The pattern is what matters:

Eight consecutive quarters of positive spread, widening as merchant mix climbs from 66% to 73%. Merchant revenue grew 15% this quarter while agency revenue declined 7%. Management confirmed again that incremental payment revenues exceeded incremental payment costs over the trailing twelve months, a statement that has appeared before, but one that survived another quarter at higher mix while sales and other expenses still showed cost efficiency despite the heavier merchant load. The payment economics are confirming, not arriving.

If the pattern holds, room night growth systematically understates Booking’s earnings trajectory, and a market that values this stock on volume is using the wrong gauge. That’s the bull case, and the table is the best evidence for it that exists. But a pattern is not yet proof of permanence. Two quarters of the spread collapsing toward zero and this argument weakens considerably.

The Treadmill

Now the side that didn’t work.

If executing more of the trip, payments, service, loyalty, bundling, makes each transaction richer, the next test is whether it makes the next customer cheaper to acquire. That’s the difference between a better margin and a genuine moat. And Q2 did not show it.

Marketing expense grew 11% against revenue of 8%. Marketing as a percentage of gross bookings rose to 4.7% from 4.6%. B2C direct mix held in the mid-60%s, stable, not improving. SEO declined, and when Mizuho’s Lloyd Walmsley pushed on whether that reflected AI search or something else, Fogel answered more directly than I expected:

I do believe that some of the changes that were made in the display at Google definitely put some pressure on SEO... Our direct, mid-60s%. Hasn’t gone down at all. In fact, total number, absolute number, obviously, had to go up to maintain the same percentage share in the direct, because obviously we grew.

The deflection is fair as far as it goes. Both direct and paid traffic grew in absolute terms; the ratio stayed flat because both sides rose. But a ratio can stay flat on a treadmill. Booking is spending more on paid channels to hold direct mix steady while organic discovery erodes underneath it. That is not what a flywheel looks like.

I want to be careful here too, because one quarter of marketing deleverage doesn’t invalidate customer loyalty. Booking is still investing in the US and Asia. Connected Trip is a low-double-digit share of transactions, large enough to matter directionally, too small to move the consolidated marketing line. Paid spend can rationally grow during geographic expansion. But the strategic metrics have reached enough scale that some evidence should be showing up in acquisition economics. Genius Level 2 and 3 members generate a high-50% share of room nights, up from mid-50%s. Mobile app mix is in the high-50%s, up from mid-50%s. These loyalty layers are supposed to convert engagement into cheaper demand. Q2 showed the engagement. It did not show the conversion.

The Second Booking

Steenbergen gave the behavioural disclosure the whole thesis rests on:

Travelers who book with us across multiple travel verticals return more frequently.

That’s the mechanism, one trip creates context, trust and service history, and the next booking should come back more directly and more cheaply. The mechanism is visible. Its economics are not. In my Q4 piece I called the real AI risk toll extraction rather than replacement: Booking stays essential while paying rising rent to whichever interface controls customer intent. Google collects that rent today through $8 billion-plus of annual marketing spend. Q2’s evidence on the interface itself was actually encouraging, LLM referral traffic remains below 1% of room nights and flat, Google launched agentic booking with Booking as a first partner, OpenAI went CPC, which Fogel called “right in our wheelhouse.” The AI layer is fragmenting, not concentrating. But the rent did not fall. Both things were true in the same quarter, and the tension between them is now the entire investment debate.

AI’s clearest contribution so far is on the other side of the P&L anyway. Customer service cost per booking is falling at a double-digit rate. Voice AI handles most eligible inbound calls. Steenbergen disclosed internal metrics, AI cost per merge request, declining, that tell you this management team treats AI spending the way it treats marketing spending: measured, ROI-gated, unsentimental. AI is making the execution layer cheaper to run before it makes demand cheaper to acquire. That’s worth something. It isn’t the moat.

So, the question I’ve been circling for four articles has narrowed to one sentence: can Booking turn the trip it executes today into the customer it owns tomorrow? The merchant model makes each trip richer; the table shows it. Genius, the app and Connected Trip give the traveller reasons to return, Steenbergen’s disclosure confirms it. But until the return shows up as lower acquisition cost, the moat is behavioural rather than financial, and the market won’t pay a premium multiple for behaviour.

Three Years Out

From roughly $195:

The bull case requires no new category label; I’ve spent enough of this series asking the market to rename the company. It requires direct mix rising above 70%, Connected Trip reaching 20-25% penetration, and marketing falling toward 4.0-4.3% of gross bookings. Operational outcomes, not a reclassification event. The bear case, the toll road, marketing intensity rising, margins flat, is roughly what’s already priced in.

Management is making its own bet on which scenario plays out. Booking returned a record $4.1 billion to shareholders in Q2 and repurchased $7.4 billion of stock in the first half at an average price of $173, a pace that retires roughly 6% of the float per year. When a company with a $10 billion free cash flow engine buys back its own equity that aggressively at these levels, it is telling you where it thinks intrinsic value sits.

I own it here, and I’d add meaningfully toward $175, near the company’s own buyback average, absent new evidence of thesis deterioration. Sized for patience, not certainty: two of three scenarios pay you, and the one that doesn’t cost almost nothing from here.

What Settles the Argument

The Q1 threshold was met, so here is the new one, and I will not attribute a breach to conflict, channel mix, or geographic investment: if marketing exceeds 5.0% of gross bookings in any of the next four quarters, the switching-cost thesis at the core of my bull case needs revision. Below 4.5% in any quarter, and the conflict-distortion explanation is confirmed and the flywheel is turning. Between those bounds, watch B2C direct mix, above the mid-60%s means repeat behaviour is converting into cheaper demand, and watch for the number nobody has disclosed: what it costs Booking to acquire a customer through an AI agent versus Google Search. Every quarter of silence on that number extends the uncertainty discount.

Booking has shown it can earn more from the trip in front of it. The moat will be proven only when that trip lowers the cost of winning the next one.

That proof will not come from AI demos or partnership announcements. It will come from the income statement.

It always does.

Subscribe now

• The proof: Premium subscribers reached 300 million, ARPU rose 7.4% in constant currency, and gross margin hit a record 33.4%, challenging the view that royalty costs permanently cap Spotify’s economics.

• The bet: Management is deliberately trading some low-quality user growth for higher conversion and monetisation, introducing more friction into the Free tier and prioritising revenue over raw MAU expansion.

• The next layer: Audiobooks+ has surpassed $100 million in annual recurring revenue, offering an early glimpse of a metered model in which Spotify charges its most engaged users for additional hours, credits, and AI-powered experiences.

Three months ago, I argued that the market wanted financial proof the Taste Graph was an engine, not a thesis. Q2 supplied some of that proof: 300 million Premium subscribers, constant-currency ARPU growth of 7.4%, a record 33.4% consolidated gross margin, and €797 million in free cash flow. Yet the most important moment in the quarter was not a number. It was a decision, one that changes how I think about the next phase of Spotify’s business.

The Old Ceiling and How It Broke

For most of its existence, Spotify had a problem that scale could not fix. Music royalties are generally calculated as the greater of a percentage of revenue and a per-user amount. As Spotify grew, content costs grew in rough proportion. Gross margin sat around 25-27% for years. The company could add subscribers and revenue, but it could not add margin, because the content cost structure moved in step with the top line. That was the bear case from the IPO through 2023: a business that could get bigger but never meaningfully richer.

The margin trajectory tells you when something changed: 25.3% in 2022, 26.7% in 2023, 31.1% in 2024, and now 33.4%. Q2’s Premium segment was even sharper, gross margin of 34.9%, up 174 basis points, as Premium revenue grew 16% on a constant-currency basis while associated content costs grew only 12%. The filing attributes the improvement to revenue growth outpacing music royalty costs net of marketplace programmes, audiobook licensing costs, and Spotify Partner Program costs. That is all Spotify disclosed about the mechanics, and it is enough: the old claim that scale could never improve Spotify’s economics is broken in the data.

The Taste Graph matters here because it is the asset that makes price increases absorbable. Spotify raised prices across dozens of markets and subscriber growth remained healthy, 7 million net additions against guidance of 6 million, with management reporting no visible damage to intake. If Spotify were a commodity distributor, a 7% ARPU increase should have produced meaningful churn. It did not. The Large Taste Model, AI DJ at roughly 100 million subscribers, Prompted Playlist, years of accumulated listening history, these create a switching cost that is invisible in any single quarter and powerful across many. The subscription proved it can escape the old economic ceiling. The question is what Spotify builds on top of it.

The Decision

Alex Norström, on the Q2 earnings call:

“We’ve been making changes like tweaking the sign-up to get a higher quality MAU throughput. We’ve deprecated a lower-end Android device support which builds the business and makes it more efficient for us. We’ve carefully introduced some friction in both ad load and some limitations in our Free tier. Gustav and I have this saying where we say sometimes we pull the growth lever, and sometimes we pull the monetization lever. Here we’re starting to pull the monetization lever.”

Spotify guided Q3 MAUs to 788 million, up 11 million from 777 million, a meaningful step down from the 16 million added in Q2 and the 15-20 million quarterly pace of recent years. The company is not shrinking. It is choosing fewer marginal users in selected emerging markets in the belief that the remaining funnel will generate more value. Norström pointed to Latin America as the precedent, where Spotify followed the same progression from high MAU growth and low conversion to a more monetisable base over several years.

I think this framing is credible. At 777 million users, raw scale is less scarce than monetisable intent, and the conversion gap in Rest of World, which represents a large share of MAUs but a much smaller share of subscribers, is precisely the kind of opportunity that rewards tighter funnels rather than wider ones.

But I also think it is important to name what this is. It is a wager, not a proof. The proof comes later: does conversion improve, does churn hold, does advertising revenue per Free user rise, does the emerging-market base generate more revenue despite slower intake? Management has changed the scoreboard before showing the new score. That does not make the decision wrong. It means the evidence that validates the decision has not yet arrived.

The Meter

The smaller number in Q2 that I think matters more than most analysts will credit: Audiobooks+ passed $100 million in annual recurring revenue.

Norström described Spotify’s usage curve in a way that connected this directly to the Taste Graph:

“At the head of that curve are many millions of people who simply want more, and they’re willing to pay for it.”

Premium includes an audiobook allowance. Heavier listeners pay for more hours. Personal Podcasts follows the same logic, with monthly credits included in Premium and the option to buy more. Gustav Söderström made the connection to AI inference explicitly: generated experiences carry a real marginal cost, so unlimited use at a flat price would separate cost from revenue at the exact moment Spotify needs them connected.

This is the business model that may sit on top of the Taste Graph. The Taste Graph identifies what each user values. The usage-based layer, the meter, lets Spotify charge its most intensive users more, rather than raising the flat subscription price for everyone equally. That creates a different kind of revenue growth than conventional price increases: it is driven by engagement intensity rather than across-the-board repricing, and it scales with the depth of the personalisation engine rather than with management’s appetite for pricing risk.

The loop this creates is stronger than better recommendations alone. More users generate more data. Better data improves discovery and generated products. Better products reveal which users sit at the head of the usage curve. Those users buy credits, hours, or add-ons, and the revenue funds the next product. At $100 million of ARR against nearly €4.8 billion of quarterly revenue, Audiobooks+ proves the architecture works. It does not yet prove materiality. The meter needs to become a business rather than an elegant explanation.

What Remains Unproved

Advertising has stabilised but not inflected. Ad-supported revenue grew 3% in constant currency, reversing the decline from Q1. Automated channels reached nearly 40% of ad revenue, up from roughly 30% last quarter, and active advertisers grew 60% to 33,000. CFO Christian Luiga, whose guidance credibility on gross margin has been near-perfect for two years, committed to double-digit ad growth in the second half. The system is built. The excuse period should end. If double-digit growth arrives in Q3 or Q4, the market has to reprice a segment it had written off. If it does not, Luiga’s credibility on this specific promise takes a hit.

One risk I have not previously discussed sits on page 25 of the 6-K. The Mechanical Licensing Collective has filed an amended complaint and has sought permission for an interlocutory appeal of the January 2025 court ruling that classified Spotify’s Premium Service as a qualifying bundle for royalty purposes. If the MLC were to prevail in full, Spotify estimates a potential liability of approximately €473 million for March 2024 through June 2026, plus possible penalties. More important than the one-time charge, an adverse outcome could raise the forward royalty rate on the music component of Premium, which would directly threaten the improved content economics that Q2 demonstrated. Management did not discuss this on the call. I am not treating it as a base case, but it is the structural tail risk to the margin story and it deserves attention.

The €200 million of incremental marketing and AI spending in 2026 has not yet been tied to quantified returns. Management calls it temporary, expects Q4 moderation, and has kept headcount flat at 7,302. Söderström’s claim that “our margin is a managed outcome” is believable given the evidence. It is not a permanent pass. Q4 needs to show that “controlled” eventually becomes “productive.”

Three Years From Now

The bear case assumes add-on tiers remain small, advertising stays weak, and AI costs absorb pricing gains, or the MLC ruling resets royalty economics. The base case assumes consensus delivery, several hundred million euros of add-on revenue, and margins approaching management’s long-term targets. The bull case requires usage-based products to become a material revenue category while Spotify retains enough economics from rights holders to exceed a 20% operating margin.

The tracking items: Premium net additions holding near 5 million per quarter despite emerging-market friction. Ad-supported revenue reaching double-digit constant-currency growth by Q4. Gross margin returning above 33% after Q3’s regulatory timing charge. Q4 operating expense growth moderating below 10% excluding currency and social charges. Paid add-on revenue moving well beyond today’s $100 million ARR. And the MLC appeal, any adverse procedural development would change the margin calculus directly.

Q2 proved that the core subscription has escaped its old economic ceiling. The monetisation decision and the meter are management’s wager that the same understanding of taste that improved retention can now support a richer, more layered revenue model. The next phase of this series is no longer about whether the moat exists. It is about whether the moat can bear weight.

Subscribe now

• The mechanism is now visible in the accounts. In twelve quarters, revenue growth rose from 17% to 93%, the sequential addition from $50 million to $303 million, and net dollar retention from 107% to 157%. Revenue is outrunning customer count while RPO, billings, and margins rise with it.

• The Ontology is an institutional compiler. Foundation models supply generic intelligence; the Ontology translates a firm’s objects, permissions, workflows, and judgment into executable instructions. Better models make more workflows possible. The compiler makes them cumulative.

• The race is depth against distribution. Palantir has the deepest deployed model of high-consequence operations; incumbents have far greater distribution. My end-2029 outcomes are approximately $72, $190, and $375.

A compiler is invisible when it works. A programmer writes in human-readable language; the machine requires exact instructions. The compiler resolves definitions, links libraries, enforces rules, and turns intention into execution. The tenth program reuses what the first nine compiled.

Large organisations face the same problem. Their operating knowledge is scattered across software, employees, policies, permissions, exceptions, and accumulated judgment. A foundation model may understand supply chains in general; it does not know which substitution this company permits, who may approve it, what SAP must record, or whether the decision worked.

The model understands the world. It does not understand the firm.

Palantir calls the translation layer between the two the Ontology. It maps an institution’s objects and relationships, connects them to live systems, encodes who may act, and records what follows when an action is taken. I think a better description is an institutional compiler: Palantir turns the messy logic of an organisation into an executable model that employees, applications, and AI agents can use.

The big fundamental question is: can each compiled workflow increase the speed, scope, and value of the next before Microsoft, ServiceNow, AWS, Google, and SAP compile enough of the firm through systems they already distribute?

Q2 matters because the accounts now look like the compiler has begun to compound.

From Coordination to Compounding

Our view of Palantir has changed in stages. In October, I described the Ontology as the modern equivalent of MS-DOS: a common language across incompatible enterprise systems. After Q4, the argument became stronger. Once an institution had mapped its objects, permissions, and operating relationships into Palantir, replacement was no longer a normal software switch; it required reconstructing the institution’s operating model.

Q1 moved the thesis from coordination to absorption. Workflows that once lived inside individual applications could increasingly move into the Ontology, while the applications became interfaces. Our Q1 article also applied the Jevons idea to enterprise AI: as inference prices fall, firms use more intelligence, which creates more agents, tool calls, and governed actions. The force commoditising models increases demand for the layer controlling what those models may see and do.

What I missed was the effect on Palantir’s own growth curve. I understood why cheap models should create more demand. I had not fully understood why every successful workflow might reduce the difficulty of the next one, allowing Palantir to accelerate even as its base became larger.

The Ontology does not merely govern more work. It preserves what was learned while that work was being compiled.

The Acceleration That Should Not Be Possible

Palantir’s year-over-year growth has accelerated for eleven consecutive quarters. The percentage progression is striking; the absolute additions are harder to explain away.

Source: Palantir quarterly earnings materials; calculations.

Palantir added roughly $326 million to its quarterly revenue base in the year ending Q2 2025. It added about $932 million during the following year. Q2’s sequential addition alone was larger than the entire US commercial business one year earlier.

A temporary AI spending surge could produce a few strong quarters. It would be less likely to produce accelerating percentage growth, larger dollar additions, higher retention, expanding contracted demand, and wider margins at the same time. Q2 revenue grew 93% to $1.935 billion; US commercial grew 149% to $764 million; US government grew 90% to $809 million. Adjusted operating margin reached 62%, and adjusted free cash flow was $1.22 billion at a 63% margin.

The most revealing comparison is revenue against customer count. Total customers rose 24% over the year while revenue rose 93%. US commercial customers increased 35%, from 485 to 653, while US commercial revenue rose 149%. Sequentially, customer count grew 6% and revenue grew 28%. The system is going deeper much faster than the customer base is becoming broader.

Net dollar retention confirms that interpretation. It was 107% in Q3 2023 and reached 157% in Q2 2026. Existing cohorts are not merely renewing; they are becoming much larger relationships.

The forward indicators trace the same progression:

Source: Palantir quarterly presentations. Figures may not sum perfectly because of rounding.

Total RPO has increased almost fivefold; long-term RPO more than sixfold. Late 2025 brought a surge in multi-year commitments. Since then, short-term RPO has accelerated toward recognition while the long-term bucket keeps refilling. Billings have nearly quadrupled. Palantir is converting backlog without emptying it.

RPO also understates part of the government business because contracts with initial terms of twelve months or less, and obligations outside termination-for-convenience clauses, are excluded. Management says the reported balance is therefore primarily commercial.

The cost structure completes the argument. From Q2 2025 to Q2 2026, Palantir added about $932 million of revenue and $730 million of adjusted operating income, an incremental adjusted margin near 78%. Revenue grew 93%; adjusted expenses grew 37%. Those are not the economics of a business that repeats the same amount of labour for every new workflow. They imply reuse.

How the Compiler Compounds

Every enterprise AI system contains two models. The first is a foundation model—GPT, Claude, Gemini, Llama, Nemotron—trained on broad information and supplying generic intelligence. The second is a model of the institution: its objects, relationships, permissions, policies, actions, decision history, benchmarks, and outcomes.

The two assets age differently. A foundation model begins losing relative value when a competitor releases a better or cheaper one. The model of the firm becomes richer each time it is used. The first workflow requires Palantir’s Forward Deployed Engineers to connect systems, define objects, encode policy, and establish safe actions. The second workflow inherits much of that work. The tenth inherits the first nine.

This is where the collapse in inference costs matters. Our Q1 article noted that inference prices had fallen roughly a thousand-fold in three years. That did not reduce AI consumption; it expanded the number of tasks that could be run economically. Each new task still needs institutional context, permissioning, auditability, and a way to act.

Three mechanisms reinforce one another. Cheaper models make more workflows viable. Each workflow enriches the customer’s Ontology, making adjacent work faster to compile. Each engagement adds connectors, industry patterns, and field knowledge that Palantir can reuse elsewhere.

Shyam Sankar described the result as an automated model factory inside the customer’s security boundary. Operational telemetry feeds customer-specific benchmarks and post-training, while the customer controls the resulting weights. He also said that a standard open model beat frontier models on five production tasks once Palantir tested it against the customer’s actual work. The important asset was not the model ranking; it was the institution’s own definition of success.

That is why “institutional compiler” is more precise than “control plane.” Palantir is not merely routing agents or governing tool calls. It is translating tacit institutional knowledge into a living system through which agents can reason, act, and learn. Forward Deployed Engineers are performing specification work. The Ontology turns that work into a reusable asset.

The old consulting critique is therefore dead, but a new risk replaces it. If every new institution still requires proportional FDE effort, Palantir has a human-capital ceiling. If templates, tools, partners, and prior deployments allow each engineer to compile more operating surface area, the ceiling rises. Q2’s incremental margins suggest that it is rising; the company does not disclose workflow-level economics, so this remains an inference rather than a settled fact.

The foundation model makes a workflow possible. The institutional compiler makes it cumulative.

The Other Compilers

Palantir is no longer alone in understanding where value is moving. Microsoft, ServiceNow, AWS, and Google now describe products that govern agents, connect enterprise context, and translate intelligence into action. The strategic question is not whether these products exist. It is whose representation becomes authoritative when an agent must decide what an object means, which rule applies, and what action is allowed.

Microsoft starts with the graph of work: identity in Entra, documents and communication in Microsoft 365, code in GitHub, data in Fabric, and processes in Dynamics and Power Platform. Fabric IQ Ontology, still in preview, defines entities, relationships, and rules for agents and workflows. Microsoft has less proven operational depth, but the broadest ingredients and cheapest route into existing customers.

ServiceNow is the closest conceptual rival. It begins with the history of services, incidents, approvals, assets, cases, and business rules. Context Engine maps people, assets, policies, and operational history; ServiceNow says that intelligence “compounds with every workflow.” AI Control Tower governs agents across vendors, while Action Fabric exposes workflows to third-party agents. Its thesis now sounds like Palantir’s: “Intelligence will keep getting cheaper. Trusted execution will keep getting more valuable.”

AWS and Google are strongest one layer below. AgentCore lets customers use any model or framework while handling runtime, tool access, identity, tracing, evaluation, and security. Google’s Agent Platform offers runtime, registry, identity, gateways, policies, and observability for agent interactions. These products complement Palantir by making agents easier to run, but they also let technically capable customers assemble a good-enough compiler themselves.

SAP remains different. It already contains executable transactional logic for finance, procurement, supply chains, and manufacturing. Any cross-system compiler must respect that authority rather than wish it away; Palantir’s own prior examples treated SAP as a source system to connect and, at times, a workflow to simplify.

The contest can be reduced to one line: Palantir’s depth is racing the incumbents’ distribution.

The concept of an ontology is not the moat; everyone now understands it. Palantir’s possible moat is the installed system: a live model already connected to source systems, trusted to execute consequential actions, and accumulating knowledge faster than a rival can reconstruct it. Microsoft and ServiceNow do not need to beat that depth everywhere. They need to make a separate Palantir purchase unnecessary for enough workflows.

Sovereignty Cuts Both Ways

Management’s “AI sovereignty” argument is a claim about ownership of the second model. If a model provider captures a customer’s prompts, reasoning traces, benchmarks, and operating outcomes, the provider’s asset compounds while the customer rents the result. Palantir says the institution should own that learning loop. Its Q2 products extend this idea into model switching, post-training, agent orchestration, and customer-controlled weights.

There is an obvious contradiction: a company may reduce dependence on model laboratories by becoming deeply dependent on Palantir. The more operating logic compiled into the Ontology, the harder it becomes to leave. Palantir’s best answer is architectural rather than rhetorical—make the compiler open above and difficult to reproduce below. Let Microsoft, ServiceNow, AWS, Google, and customer-built agents use the Ontology; keep the advantage in constructing and maintaining the live institutional model.

Sovereignty also cuts geographically. US revenue grew 115% and now represents more than 81% of Palantir’s revenue. International commercial revenue grew 26% and only 2% sequentially. American companies may see a US vendor with defence-grade controls as an advantage; foreign institutions may see dependence on that same vendor as a sovereignty risk.

I would therefore treat international as upside, not as a required base-case input. The US opportunity is large enough to carry the next several years. A company valued as a universal institutional layer will eventually have to prove that the compiler can travel.

What Q2 Changed

Before Q2, I believed model commoditisation increased the value of governed action, the Ontology was absorbing workflows, and delivery capacity was the main constraint. I still believe all three. What changed is that increasing returns inside customers are no longer merely an architectural possibility. Revenue density, retention, contracted demand, billings, and margins now move together in the pattern the thesis predicted.

The new variant view is therefore about duration. Conventional software models assume each sale is another unit added to a larger base, so growth must eventually slow. Palantir’s relevant unit may be different: the amount of institutional operating surface compiled into the Ontology. Each new workflow can increase the value of the workflows already there.

The underappreciated risk also changed. It is no longer that Palantir is consulting in disguise. It is that Microsoft or ServiceNow becomes good enough for ordinary workflows before Palantir can industrialise its depth. The market may be underestimating both the duration of Palantir’s growth and the power of incumbent distribution.

Three Years From Now

I value the outcomes at year-end 2029 using FY2030 earnings, because that is the period public markets would then be pricing. The starting point is Palantir’s new FY2026 revenue guide of roughly $8.15 billion and a current reference price of $125.65.

In the bear case, incumbents capture ordinary workflows while Palantir keeps the hardest government and industrial settings. In the base case, delivery scales and Palantir remains authoritative in high-value cross-system work. In the bull case, third-party agents consume Palantir’s context rather than replace it, making the compiler a neutral layer across clouds and models.

The probability-weighted value is about $207, roughly 65% above the current price, but it is not a precise target. The bull case contributes heavily, and small changes in growth duration or terminal multiple move the answer sharply. The $9.2 billion cash balance supports the downside, but I do not add it separately to a P/E-derived value because part of its benefit already appears through interest income and future capital allocation.

Palantir has proved the margin structure. The harder question is how much of the firm it can compile, and for how long.

What Would Prove Us Wrong

The thesis should remain falsifiable. Revenue must keep growing far faster than customer count; net dollar retention should stay above 145%, with a fall below 135% signalling weaker customer compounding. Short-term RPO must convert while long-term RPO continues to refill, and billings should not trail revenue growth for several quarters. Adjusted operating margin should remain above 55%, with incremental margins above 60%, or the human-capital ceiling deserves another look.

US commercial growth below 60% before late 2027 would weaken the duration case. International commercial growth above 30–40% would expand the runway; sustained growth below 20% would confirm a geographically narrower opportunity. The competitive test is more qualitative but just as clear: are Microsoft and ServiceNow agents using Palantir as the authoritative source of institutional context, or is Palantir becoming one tool inside their systems?

Compilers usually disappear into the systems built on top of them. Users do not think about the definitions, libraries, and dependencies being resolved beneath the surface; they notice that new work becomes easier because the hard translation was done before.

That is what Q2 may be showing. Model intelligence is getting cheaper. More work becomes automatable. Each Palantir workflow adds context that can be used again, while each deployment adds knowledge that can shorten the next one. Revenue accelerates, contracted demand refills, and marginal costs fall.

The model laboratories are teaching machines to understand the world. Palantir’s bet is that the greater prize lies in compiling each institution into something those machines can safely understand and act upon.

Q2 is the first quarter where that compiler looks capable of accelerating itself.

Subscribe now

• Lam cleared nearly every test we set far earlier than expected: gross margin reached 52%, NAND doubled to 23% of systems revenue, and CSBG rose to $2.47 billion.

• The structural thesis is now stronger: more three-dimensional and technically demanding chips require more etch, deposition, and service intensity per wafer—but the spending that expresses that complexity remains cyclical.

• The old bear case was China and margin compression. The new one is that customers are compressing several years of investment into a short window, creating a future capacity digestion cycle that Lam has not yet proved it can compound through.

In April, at $275, I argued that AI was making semiconductor manufacturing more Lam-shaped, increasing the etch and deposition work required per wafer, independent of how many wafers the industry processes. I set four signposts. Gross margin above 50%: proves the operational flywheel. NAND above 15% of systems revenue: proves the conversion pull-forward is real. CSBG above $2 billion: proves the installed-base compounding is structural. CY27 WFE above $155 billion: proves the multi-year cycle extends.

I expected those tests to take time.

Lam cleared almost all of them in one quarter. Gross margin hit 52%, a 20-year high. NAND hit 23% of systems. Doubled. CSBG jumped to $2.47 billion. WFE was raised to “the low $150 billion range” for 2026, with 2027 described as an “extraordinary setup.” Then Lam guided September revenue to $8.1 billion, $970 million above consensus, with a 39.5% operating margin.

Our thesis was right. Our sense of speed was wrong.

One correction. In January I modeled a 48–49% gross margin floor after China declined. The actual number is 52%, and management is targeting mid-50s. I was wrong by 300 basis points. The R&D maturity flywheel I described in April was real, but I underestimated both its magnitude and the pricing contribution that comes from solving the hardest manufacturing problems at the most advanced nodes.

And one prior debate that can be closed. China fell from 34% to 26% of revenue. Lam posted record margins and a massive guide. The old thesis, that declining China revenue would compress margins, has not merely weakened. It has been falsified.

The question in April was whether AI was making semiconductor manufacturing more Lam-shaped. June answered it. The better question now: can rising semiconductor complexity increase Lam’s content per wafer fast enough to outrun the capacity cycle that its own success is helping create?

The same problem, spreading

Lam’s served available market reached roughly 36–36.5% of WFE, ahead of the timeline set at the 2025 Investor Day. Doug Bettinger, when Mizuho’s Vijay Rakesh asked:

“We at that point were talking about our SAM expanding from the low 30% of WFE range into the high 30% range. As we sit here today, we’re probably trending already to that high level. We’re, I don’t know, I’d guess 36%, 36.5% this year.”

Low-30s to 36.5% in 18 months. That doesn’t mean Lam invented a new market. It means more semiconductor spending is flowing toward etch, deposition, and surface treatment where Lam is already strong. The same technical problem, building three-dimensional structures with tighter geometries, is spreading across device categories.

In NAND, the problem is depth. In January I called it the coiled spring. This quarter it fired. NAND went from 12% to 23% of systems revenue, more than doubling sequentially. Tim Archer framed it as structural AI demand:

“Expanding context windows and persistent memory requirements are driving significantly higher demand for flash storage.”

That completes a narrative shift SK Hynix started in January. If NAND spending is genuinely AI-driven, the historical cyclical discount investors apply is wrong.

But I want to note how Archer handled the pushback. When Citi’s Atif Malik raised TrendForce forecasts of NAND oversupply in 2027, Archer didn’t defend NAND directly. He widened the lens, “the Lam story is a lot bigger than just NAND.” A fully confident defense would have met the oversupply question head-on. He redirected instead. The structural reframing is still the right emphasis, but when management deflects rather than defends, I want to register that.

The air pocket risk from April, what fills the revenue stream when the $40 billion conversion completes?, has not been answered, only deferred. Whether conversion hands off to greenfield capacity or enterprise SSD demand sustains the build is what the next articles need to track.

The same physical constraint is appearing in DRAM, where advanced nodes now require technologies first proven in foundry logic. Lam’s Akara platform, adopted for GAA logic at 2nm, is winning conductor etch positions in DRAM, its installed base doubling annually. Advanced packaging is growing more than 70% year-over-year, with panel-level tools shipping across multiple geographies.

Different device. Different format. Same physical problem.

The 52% gross margin confirms the mechanism is financial, not just technical. The toll rate rises, more process steps per wafer, and the collection cost falls: more mature tools, closer-to-customer manufacturing, scale leverage. Pricing contributed, but management didn’t quantify the component, so I can’t cleanly separate it from mix and efficiency gains. The prior articles correctly identified customer concentration as a margin constraint, and TSMC, Samsung, and SK Hynix remain formidable procurement organizations.

Management raised long-term targets to mid-50s gross margins and mid-40s operating margins, a material upgrade from the 2025 Investor Day framework. Several quarters need to pass before declaring 52% the permanent floor rather than a favorable mix quarter.

The installed base earns its keep, and the balance sheet absorbs the strain

CSBG at $2.47 billion, up 43% year-over-year, is the compounding layer from April made visible at scale. Bettinger confirmed it’s “growing faster than the installed base.” That’s ARPU expansion. Equipment Intelligence and Dextro are in “the very early stages of a multi-year rollout” across 100,000+ chambers.

Here’s the tension. The Q4 upside came entirely from CSBG. Systems revenue of $4.25 billion missed consensus by $177 million. If complexity means more tools per wafer, systems should accelerate too.

And the ramp’s financial strain shows. DSO jumped from 64 to 72 days. Receivables grew 29% on 15% revenue growth. Free cash flow was 56% of adjusted net income. Buybacks slowed to $246 million from $796 million. Headcount rose roughly 1,800 in one quarter.

The systems miss is probably timing, the $8.1 billion September guide implies ~$5.6 billion in systems, a 32% sequential jump. Deferred revenue grew $213 million. Japan acceptances rose to $490 million. The tools are in the pipeline.

One quarter of this mismatch is timing. If systems miss again in September and DSO stays above 70, the quality of the growth deserves harder scrutiny.

Success creates the next bear case

The old bear cases have weakened. China-margin compression: falsified. NAND as management talk: $978 million says otherwise. The 50% margin needing China: disproved at 52% with China at 26%.

The new bear case is that a real structural advantage is being compressed into an unusually short spending window. Customers have announced fabs through 2027 and beyond. DRAM, NAND, foundry logic, and packaging are all rising together. That is excellent for orders. It also creates the conditions for excess capacity when those fabs produce simultaneously.

The right distinction, and the one I collapsed too easily in prior versions, is between complexity as a persistent technical direction and spending on complexity as a cyclical financial expression. Complexity may never peak. The rate at which customers spend to address it absolutely can.

What the market prices and what it doesn’t

At what point does our variant perception become consensus?

At $298, with analysts raising targets to $350–$450 and the structural narrative spreading, the qualitative thesis is no longer ours alone.

It’s also worth recognizing that management is actively selling the re-rating. Every disclosure choice, SAM quantified at 36.5%, advanced packaging upgraded to >70%, new profitability targets at peak performance, “compelling” escalated to “extraordinary”, serves one purpose: convince long-only capital that Lam deserves a compounder multiple. That doesn’t make it wrong. During an upcycle, authenticated structural shifts and peak-cycle management confidence produce identical signals. They separate when the cycle turns.

The narrative has shifted. The numbers have not. FY27 consensus at $31.1 billion and $8.20 EPS is difficult to sustain after a $8.1 billion quarterly guide. Estimates need to move materially higher. The revision cycle should provide a mechanical tailwind over the next 60–90 days.

The bear case is not catastrophe. It’s the possibility that the physical thesis proves correct while the financial expression proves cyclical, customers compress spending into FY27, capacity overshoots, and rising content per wafer does not fully offset the decline. Lam exits as a better business. Just not one that has escaped cyclical valuation.

Six signposts:

1. December revenue guidance. Above $8.3B = September was a step. Below $7.8B = timing concern.

2. Gross margin. Below 50% for two consecutive quarters challenges the thesis.

3. CSBG. Above $2.5B confirms the franchise. Below $2.1B during high utilization weakens it.

4. Cash conversion. DSO below 65 and FCF conversion above 80% by December = June was timing.

5. The NAND handoff. Conversion must bridge to greenfield or sustained enterprise SSD demand.

6. WFE after 2027. Spending must stay high enough that content and CSBG offset systems digestion.

The toll booth metaphor has served its purpose across four articles. What this quarter showed is that it compounds: the toll rate rises with complexity, the collection cost falls with scale, and every tool enlarges an installed base that generates rising revenue for years.

Management is selling the conclusion, Lam compounds through cycles, before the test arrives. The test arrives when the fabs being built today begin producing simultaneously and equipment orders decelerate. Until then, the structural narrative and the cyclical data will be indistinguishable.

The mechanism is real. I believe that more than I did in January, more than I did in April, and more than the evidence strictly requires me to. What I do not yet have is evidence that it survives a spending pause. That evidence cannot exist until the pause arrives.

This quarter proved Lam captures more when customers spend. The next cycle will show how much persists when they stop.

Subscribe now

• AI has moved toward Amazon’s architecture. Agents require CPUs, storage, databases, security, and enterprise data—not just accelerators—turning AI into a multiplier on the rest of AWS.

• The Foundry is strengthening the Engine Room. Amazon’s custom silicon and AI services increasingly support retail, advertising, logistics, and internal automation, creating a reciprocal flywheel across the company.

• The remaining question is cash. AWS growth, margins, and backlog validate the investment, but each successful capacity expansion appears to justify an even larger one before shareholders see the free cash flow.

“We’ve long believed AWS could become a few hundred billion dollars revenue business and now believe it will be at least double that and very possibly be $1 trillion annual revenue business for us in time.”

— Andy Jassy

I have spent the past three quarters asking whether Amazon’s Engine Room could finance its Foundry.

In Q3, commerce, Prime, sellers, and advertising were generating enough cash to fund AWS, AI capacity, and custom silicon. In Q4, the Foundry stopped being one destination for Amazon’s surplus and became the organizing principle of capital allocation. Q1 proved that customers wanted what Amazon was building, but the balance sheet entered the story, and I ended with a deliberately hard conclusion: evidence was not yet cash.

Q2 makes the evidence overwhelming and the cash question harder.

AWS growth needed to remain above 27%; it reached 36.7%. Advertising needed to hold above 22%; it accelerated to 26%. Paid units needed to grow at least 12%; they rose 17%. Trainium needed customers beyond Amazon’s investment portfolio; OpenAI, Uber, Pinterest, and a growing group of AI companies appeared.

The two cash tests failed. Trailing free cash flow moved from positive $1.2 billion to negative $7.6 billion. Long-term debt reached $128.9 billion, close to the $150 billion level I had marked as a warning. Amazon then raised expected 2026 cash capital spending from $200 billion to $220 billion.

I was right about the demand. I was too conservative about how quickly it would become revenue and profit, and too optimistic about when those profits would become free cash flow.

That is enough reconciliation.

The deeper question is whether the nature of AI has changed in a way that makes Amazon’s old architecture once poorly suited to frontier-model training unusually well suited to inference and agents. If it has, AWS may have entered a cycle in which AI demand pulls the rest of the cloud behind it, cheaper internal silicon widens the profit pool, and better economics justify still more capacity.

The risk is no longer that the Foundry fails. The risk is that it works so well it never finishes being built.

The Workload Moved Toward Amazon

Three years ago, the case against AWS was persuasive.

The first generative-AI wave revolved around training giant models across thousands of tightly connected Nvidia accelerators. That job rewarded the fastest horizontal networking, full Nvidia systems, and clouds willing to reorganize their architecture around one synchronized cluster. AWS had spent years moving in a different direction: Nitro, its own networking, disaggregated resources, managed services, and custom chips built around the cost structure of general cloud computing.

Those choices made AWS weaker for frontier training. The criticism was right. Then AI changed shape.

The first wave generated answers to prompts. Reasoning models began generating many more tokens before producing an answer. Agents now trigger those reasoning loops repeatedly, retrieve context, call software, query databases, check permissions, save memory, and act again without waiting for another human instruction.

Each shift changes what the underlying system must do well.

Training wants one giant accelerator cluster. Production inference needs reliable serving near the model. Agents need CPUs for orchestration and tool calls, storage for memory, databases for context, identity and security for permissions, and access to the applications and proprietary data on which they are supposed to act.

That sounds much more like AWS.

Nitro was designed to separate and route different resources efficiently. Graviton was built to own the CPU cost curve. AWS’s broader advantage was that much of the enterprise data and software an agent needs already lived there.

AWS did not suddenly repair every weakness identified during the training era.

The workload moved toward the architecture Amazon had already spent a decade building.

Jassy described the commercial consequence:

“Growth in AI drives core because post-training, reinforcement learning, and agent tool use is mostly done on CPUs versus AI accelerators.”

This is a claim about the structure of demand, not a product boast. AI spending does not sit beside conventional cloud spending. Once a workload enters production, it pulls CPU cycles, storage, databases, networking, security, and observability behind it.

AI becomes a multiplier on the rest of AWS.

Microsoft independently described the same mechanism two days earlier. Satya Nadella said CPUs were “just as important as GPUs” for agents, while Microsoft highlighted rising demand for PostgreSQL, Fabric, Foundry, enterprise context, identity, and security. Amy Hood said additional CPU and GPU capacity was monetized almost immediately once engineering teams made it available.

Two competitors are describing the same system. That makes it harder to dismiss Jassy’s explanation as earnings-call framing.

Amazon may be better placed to collect on it because more applications and data already reside in AWS, and because Graviton gives it a credible, widely adopted CPU cost advantage.

The financial evidence is striking:

Q2’s $4.6 billion sequential increase was about 80% larger than AWS’s prior record. Growth accelerated for the fifth straight quarter. Backlog reached $496 billion, more than three times current annualized AWS revenue, although Amazon did not disclose the duration, customer concentration, or conversion schedule. Most 2027 capacity is reserved, and Jassy described demand already committed for 2028 as “striking.”

Then there was margin.

Reported AWS operating margin reached 39.4%, including a $600 million energy-contract gain. Strip that out and clean margin was still roughly 37.9%, slightly above Q1, while depreciation grew more than 30%.

Growth accelerated by nine points. Margin held.

At a $169 billion revenue run rate.

Part of that reflects high utilization, mature core services, software improvements, and cost discipline. Trainium cannot explain every basis point. It does, though, explain why margins like these may persist.

The lower layers of inference are becoming more commodity-like. Customers increasingly consume an outcome through Bedrock or another managed service rather than selecting a chip directly. When models can be swapped and the interface hides the hardware, cost structure matters more. A provider paying full third-party silicon margins sets one cost benchmark; Amazon can run a growing share of its services on Trainium and retain the difference.

That is the old Graviton playbook: sell the higher-value service and use cheaper internal silicon underneath.

Amazon’s AI and chips businesses have each passed $25 billion annual revenue run rates, both growing at triple-digit rates. Graviton is used by 98% of the largest EC2 customers. Trainium has multi-year, multi-gigawatt commitments from Anthropic and OpenAI. Jassy previously said Trainium could save tens of billions of dollars in annual capital spending and add several hundred basis points of margin at scale.

Q2 does not prove the whole claim. It makes it much easier to believe.

The Engine Room Learns from the Foundry

It would be a mistake to write this as an AWS-only quarter.

Online stores grew 15%, up from 9% in Q1. Third-party seller services grew 16%, up from 12%. Paid units increased 17%. North America and International revenue grew 16% and 15%, respectively. Advertising stepped from four quarters at 22% growth to 26%.

Retail’s mechanism is speed.

Amazon delivered more than 40% more items on the same day or overnight in the first half. Monthly active perishables customers rose more than 50% since January. Same-day orders containing perishables carried more than three times as many units. Pharmacy doubled new customers, and same-day prescription deliveries increased nearly fivefold.

A two-day service is useful for a planned purchase. Same-day delivery can become the default for milk, medicine, toiletries, and whatever the household forgot this morning.

Speed increases frequency. Frequency improves inventory and route density. Density lowers unit cost and supports wider selection and faster delivery.

Then it repeats.

There was one crack. Shipping costs grew 19%, two points faster than paid units. For three quarters I had treated units growing faster than shipping costs as evidence that the retail network was gaining operating strength. That spread reversed. Olsavsky blamed fuel inflation and higher linehaul rates, with an FBA surcharge offsetting part of the pressure.

I am willing to accept the explanation for one quarter. I will not ignore the metric.

North America’s reported 7.9% margin also included a $600 million tariff refund. Clean margin was closer to 7.3%, below Q1. Retail demand accelerated, but the cost advantage did not widen.

Advertising converts that demand into cash.

The business reached $19.8 billion in quarterly revenue, nearly an $80 billion annual run rate. Amazon does not disclose its profit, but the economics are attractive: the shopping traffic, purchase data, and customer relationship already exist. An additional sponsored recommendation does not require another warehouse or delivery route.

AI is creating more monetization surfaces. Shoppers who click a sponsored prompt convert 48% more often and spend 21% more. Advertisers using Ads Agent see lower costs per impression and acquisition. Prime Video adds attention outside the store; inventory across the NFL, NBA, WNBA, and NASCAR sold out.

There is a real risk. AI discovery may send more shoppers directly to brand sites and Shopify merchants, leaving Amazon with fulfillment while weakening the discovery moment where its most valuable ads sit. For now, agentic commerce looks more like a referral channel than a distributor that owns checkout. That gives Amazon time to build sponsored prompts into Rufus and Alexa without surrendering the customer relationship.

This is the update to the Engine Room and Foundry framework.

The Engine Room still supplies the cash. Retail creates traffic and transaction data. Advertising monetizes it. Prime deepens the relationship.

The Foundry is beginning to pay some of that investment back. Trainium runs Bedrock; Bedrock supports Rufus, Alexa, advertising tools, and internal automation; better shopping experiences drive conversion and purchase frequency; more commerce creates more advertising revenue; that revenue helps finance the next data center.

The relationship is now reciprocal.

Amazon Supply Chain Services shows the same corporate formula in the physical world. Amazon built logistics for itself, reached extraordinary scale as its first customer, then began selling the capability to P&G, 3M, Lands’ End, and American Eagle.

AWS was born this way. Logistics may be next.

The Market Paid for Proof

Amazon raised expected 2026 cash capital spending from $200 billion to $220 billion, largely because memory costs increased, and the shares rose.

The market does not hate capital spending. It hates spending it cannot underwrite.

Alphabet delivered extraordinary Google Cloud growth, margin, and backlog, yet a larger capex shock and negative quarterly free cash flow overwhelmed that evidence. Microsoft paired accelerating Azure with positive free cash flow and a clearer path to containing consolidated margin pressure. Amazon supplied the most explicit asset-level return math of the three.

Data centers begin consuming cash roughly two years before monetization, but the buildings last more than 30 years and can host several generations of servers. Servers are bought only months before deployment, when Amazon has much better demand visibility. They break even in slightly under three years, last five to six years, and most current AI capacity is contracted for at least five.

The math is attractive. It is not complete.

The sub-three-year break-even applies to servers and networking, not the fully burdened cost of land, buildings, power, financing, and unused capacity during construction. Amazon proved that equipment placed into a busy facility can earn strong returns. It has not shown when the total capital program generates cash for shareholders.

The peer comparison adds another wrinkle. Estimated annualized net-new cloud revenue in the quarter was roughly $19.0 billion for GCP, $18.6 billion for AWS, and $15.9 billion for Azure. The figures are estimates and contract timing differs, so I would not build an entire thesis around them. They do invert the lazy percentage-growth narrative: AWS is adding immense absolute dollars despite its larger base.

Microsoft owns the strongest enterprise distribution through Office, GitHub, identity, security, and Agent 365. Those products create Azure demand, though they also claim scarce compute for Microsoft’s own use. Google has the deepest integration among models, TPUs, Search, YouTube, advertising, and Cloud.

Amazon’s consumer businesses are more physical. AI tends to improve shopping, logistics, and advertising rather than directly replace them. That may give Amazon less defensive urgency when allocating capacity between internal products and external customers.

Less. Not none.

Amazon is building Rufus, Alexa, Quick, Kiro, its own models, and a deep Anthropic partnership. The neutrality advantage remains a hypothesis.

Management’s new disclosures were designed to make the capital easier to accept: $496 billion of backlog, $25 billion-plus AI and chips businesses, capacity reserved through 2028, five-year contracts, and explicit break-even math.

What Amazon left out is where the uncertainty now lives.

It did not disclose how much backlog belongs to Anthropic and OpenAI, how much the AI and chips numbers overlap, when the backlog converts, the split between short- and long-cycle capital, fully burdened returns, or the expected free-cash-flow inflection. Asked about funding, Olsavsky said Amazon had “a lot of options.”

A non-answer.

Anthropic is part asset, part tenant, and part concentration risk. The $53.4 billion pre-tax gain made Q2 net income meaningless for operating analysis. It also confirmed that Amazon owns a valuable financial asset whose value and AWS consumption are tied to the same relationship.

Useful. Complicated.

A Foundry That Never Finishes

The old bear case was that Amazon would build capacity before demand appeared. Q2 weakened it. The new risk is rational perpetual reinvestment.

Every capacity cohort may earn a good return, reveal still more demand, and justify a larger cohort before the cash from the previous one appears in consolidated results. Operating cash flow rose 33% to $161.4 billion. Net property spending rose 64% to $169.0 billion. Free cash flow turned negative. Long-term debt nearly doubled in six months.

The Q3 headline guide is distorted because Prime Day moved from Q3 last year into Q2 this year, with nearly four points of calendar impact and another 80 basis points of currency pressure. AWS has no Prime Day.

I would underwrite a wide $2.5–4.0 billion sequential AWS addition, implying roughly 35–40% growth. The range is wide because Amazon gave strong demand commentary but no quarterly capacity schedule. Below 30% would reopen the question of whether Q2 was a concentrated capacity release. Clean margin below 35% would suggest that depreciation is arriving faster than the benefits from utilization and custom silicon.

For valuation, consolidated operating income is the cleaner primary anchor. A sum-of-the-parts view is useful because AWS, advertising, commerce, and Anthropic have very different economics, but it can count the same traffic, data, and AI investment twice.

At a post-print price around $258, my three-year ranges are:

The numbers are not the thesis. They show where the thesis matters.

AWS must keep adding at least $3 billion of quarterly revenue or explain the capacity timing. Clean AWS margin should remain above 35–36%. Advertising needs to hold above 20%. Shipping-cost growth must fall back below unit growth. Debt should remain near or below $150 billion at year-end. Above all, operating-cash-flow growth needs to begin outrunning capital-spending growth during 2027.

Q4 was the declaration. Q1 proved the demand. Q2 proved the operating economics.

The remaining proof is cash.

The Foundry works. The question is whether Amazon’s success keeps extending the buildout faster than shareholders can ever see it.

Subscribe now

• The advertising thesis is working. Meta grew ad revenue 27% despite daily users rising only 3%, as better recommendation and advertising models increased both inventory and pricing.

• Product return is no longer the same as shareholder return. Capital spending, depreciation, research costs and widening AI ambitions absorbed nearly all of Meta’s operating cash flow during the quarter.

• The key question is whether the lab remains subordinate to the advertiser. Meta does not need to win every frontier-model contest, but it must show that unproven projects face a real capital hurdle and that free cash flow eventually returns to shareholders.

A year ago, in The Infinity Machine Boots Up, I wrote that “the ROI isn’t coming; it’s here.” Meta had shown that better recommendation models could increase engagement, create more ad inventory and improve conversion at the same time. Revenue was accelerating, margins were expanding, and the company could reinvest the resulting profit into still more compute. I called it an infinity machine.

I still believe the core idea.

What I no longer believe is that the return from Meta’s advertising systems and the return from Meta’s entire AI budget can be treated as the same thing.

That distinction did not matter much in 2Q25. It matters now. Meta grew advertising revenue 27% while daily users increased only 3%. Family revenue per person rose about 24%; ad impressions increased 14%, and the average price per ad rose 12%. Yet reported operating margin fell from 43% to 31%, clean margin was closer to 37% after removing legal and severance charges, and $31.1 billion of quarterly capital spending left only $784 million of free cash flow. The product return is real. The shareholder return is no longer automatic.

I now think there are three separate questions. Does AI improve the product? Does that improvement exceed the operating cost of producing it? Does the cash left over accrue per share? In 2Q25, all three answers appeared to be yes. In 2Q26, the first remained emphatic, the second became more conditional, and the third was no—at least for the quarter.

I was right about Meta’s return on intelligence. I had not separated that return from Zuckerberg’s appetite to reinvest it.

The Closed Loop

The fundamental question for Meta is no longer whether AI helps the business. It plainly does.

The question is whether Meta can keep the loop closed: can more intelligence create more relevant content, more attention, better advertising and more profit, which then funds still better intelligence, or will each proof of return simply widen the company’s ambitions until the advertising machine becomes the funding source for an open-ended AI race?

The productive loop is unusually powerful. Better models decide what someone wants to see before that person knows to ask. Better content keeps the person engaged for longer. More engagement creates more commercial moments. Better ad models match those moments with businesses willing to pay, and the resulting conversion data improves the next decision.

Scale strengthens every step. Meta has 3.6 billion daily users, a vast supply of content, millions of businesses and an auction learning from countless interactions. A rival can build a capable model. Recreating all four sides of that system is much harder.

The quarter confirmed the mechanism. Meta’s audience barely grew, but the economic value of that audience jumped. The mix was healthy too: impressions supplied more inventory while higher pricing showed that the new inventory retained value. Worldwide impression growth did slow from 19% in the first quarter to 14%, so I would not extrapolate 27% ad growth forever. Still, this was not a company squeezing a mature user base. It was a system becoming more productive.

The Advertising Company

Last year I wrote that Meta was no longer an advertising company. That was too clever.

Meta is an advertising company. That is the bull case.

What changed is the meaning of advertising. Twenty years ago, a business bought space beside content. Today it gives Meta a budget and an objective; Meta can create the ad, identify the customer, choose the moment, predict the conversion and adjust the campaign. The business is buying an outcome rather than a slot.

Google monetises intent after a user expresses it. Amazon monetises intent near the point of purchase. Meta is strongest one step earlier: it discovers desire. Someone opens Instagram to be entertained and encounters a product they did not know existed. AI improves the content, detects the latent interest, generates the commercial message and finds the seller.

This is why I do not think Meta needs to own the best general-purpose model in every category. Anthropic may lead coding, OpenAI may own more consumer productivity, and Google may have the broadest model-and-distribution bundle. Meta can still earn the highest return from intelligence because it owns a direct route from prediction to commercial demand.

That is my variant view. The AI winner will not necessarily be the lab with the best benchmark. It may be the company with the most profitable use for a model that is good enough.

When Success Widens the Bet

The old thesis assumed a closed financial loop. There is now a second possibility. A self-funding loop can become a self-expanding one.

That does not make the spending foolish. Meta’s aggressive GPU purchases in 2022 looked excessive before they helped repair recommendations after Apple’s privacy changes and TikTok’s rise. The willingness to commit before the payoff is obvious is part of the company’s strength.

But “AI spending” now covers several different claims on capital. Recommendation, ranking and ad automation have earned their spending because the return is visible. Enough model and compute ownership to avoid dependency may be necessary even when the direct return is harder to measure. Frontier research, enterprise agents, external compute, smart glasses and the continuing Reality Labs effort remain less proven.

Meta does not tell us how the budget is divided. That is the gap.

The reported quarter looked worse than the underlying business: $2.4 billion of legal charges and $1.18 billion of severance reduced operating income, and excluding them produces a clean margin near 37%. The legal cost cannot be dismissed—youth-related litigation may recur—but it tells us little about the return on AI. Even on the cleaner number, margin was down roughly six points from a year earlier. Research expense rose 67%, stock compensation rose 58%, depreciation rose 46%, and Reality Labs lost another $4.6 billion. Operating cash flow increased 25%, yet Meta made no share repurchases and issued almost $25 billion of debt as capital spending absorbed nearly all of the cash generated by operations. One quarter of construction payments should not be annualised, but the direction is clear.

The core machine did not fail.

More claimants arrived.

A Price for Ambition

Zuckerberg said Meta has received a “large number of offers” for its compute at a “meaningful premium” to what the company paid. Most people heard the outline of a new cloud business. I heard something more useful: an observable opportunity cost.

An outside rental price tells Meta what an internal workload must beat. If a recommendation model can generate more advertising profit than the same capacity would earn from a customer, Meta should take the compute back. If an internal project cannot clear that bar, it should explain the strategic value of doing so or let someone else use the machines.

Ben Thompson made this point before the quarter, and it is the part of his argument worth keeping: compute rental can be a hurdle rather than a destination. Meta does not need another reason to build. It needs a price for deciding which ambition deserves what it has built.

This is not an argument for replacing founder conviction with a committee. Zuckerberg’s willingness to invest early has created enormous value. The better answer is a rule that lets successful bets win more capital and forces unsuccessful ones to lose access to it.

Meta need not prove that every AI project will work.

It needs to show that failure has a cost.

What I Believe Now

I am more confident that Meta owns one of the best commercial applications of AI in the world. I am also more confident that the company does not need to win every frontier contest to benefit from the technology.

I am less willing to award the wider AI programme the returns already visible in advertising. Compute rental can protect the downside, but returns on rented hardware—however attractive during scarcity—should not receive the same valuation as a self-reinforcing ad system. Agents, glasses and enterprise products may become large businesses; for now, they are claims, not evidence.

This is why the stock’s decline was both understandable and incomplete. The quarter was slightly below the buy-side revenue bar and third-quarter guidance implied slower growth. Yet clean operating income was better than the headline, and the capex midpoint barely moved. Meta managed to spend too much for shareholders focused on free cash flow and too little for semiconductor bulls seeking another upward capex surprise.

The deeper conclusion is simpler:

The quarter raised my estimate of Meta’s underlying earning power and lowered my confidence that shareholders will capture all of it.

Three Years From Now

The valuation turns on whether product return, company return and shareholder return reconnect by 2029.

In the bear case, advertising remains a good business but slows before spending does. New products do not become material, annual capital spending remains near $200 billion, and Meta is valued as a capital-heavy advertiser.

In the base case, advertising compounds in the mid-teens, clean margins recover toward 40%, spending intensity peaks and free cash flow returns. Meta does not need another business of comparable size; it needs the lab to remain subordinate to the economics of the core.

The bull case requires a larger expansion of content and commercial inventory, plus one new profit pool—agents, messaging or glasses—becoming material. Internal uses of compute must earn far more than external rental, and capital spending must stop growing faster than revenue.

At roughly $538 after the print, the base case implies an annualised return of about 18–26% over three years. The bear case is roughly flat to down; the bull case requires evidence I am not yet willing to capitalise.

What Would Change My Mind

I will track five things. Advertising growth above 20% through early 2027 would confirm that the productive loop remains strong; below 15% while spending stays elevated would weaken it. Clean operating margin needs to stabilise above 37–38%, rather than settling below 35%. Quarterly free cash flow should recover above $10 billion before 2028. Meta needs to disclose enough about compute sales or internal returns to show that capacity has a real hurdle. And buybacks should eventually resume at a level that offsets stock compensation.

A year ago, I said the ROI had arrived.

It had.

What I missed was that the arrival of a return does not settle who gets to use it. Meta has proved that it knows how to turn intelligence into revenue. It now has to prove that it knows which ambitions deserve that revenue.

That is the cost of being right.

Subscribe now