Build2 distinct publishers2 min readPublished
Revenue of $96.2bn at a 75.0% gross margin, with 74.0% guided for the next quarter, says the supplier is not trading price for volume. Capacity plans written in 2025 need new pricing inputs.
The Engineer · Build desk

Compiled by The EngineerSomething wrong?How this is made
Start with the margin line, because that is what prices capacity next year. Seventy-five points on $96.2 billion is about $72.2 billion of gross profit in a single quarter [1][2][1]. Hold 74.0% on the guided $108.0 billion and gross profit is roughly $79.9 billion [8][2]: a hundred basis points of compression [3] arriving with about $7.8 billion more gross profit than the quarter just closed [4]. A supplier trading price for volume does not print that. This one is not being asked to.
The plan assumption to delete, then, is the one that had each accelerator generation landing with softer unit pricing. Cost-per-token improvements are still available, but they have to be found in perf-per-watt, utilization and model-side work, because the vendor's take rate has moved by one point across two quarters [2][8].
The shape of growth needs re-basing too. Sequential revenue growth was 18% and the guide implies about 12.3% [1][5], while the sequential dollar add goes from roughly $14.7 billion to $11.8 billion [7][6]. For scale: the quarter cleared the consensus runtimewire cites, $92.1 billion, by $4.1 billion [15][10], which is around a third of the increment already guided for the current quarter [14]. Anyone reading earnings surprises as a demand signal is calibrating on the smaller number.
Concentration is the third input to change. Data Center at $89.0 billion is 92.5% of the company [5][8], and its sequential add of $13.8 billion [6][9] is nearly twice the whole Edge Computing line of $7.2 billion, which grew 27% year over year against Data Center's 117% [13][5]. There is no diversified second engine to smooth a Data Center air pocket, and none to compete for fab and rack capacity either.
Where the binding constraint actually sits shows up in the announcements that have nothing to do with silicon. NVIDIA secured land, power and shell capacity with SB Energy at the PORTS-Pike campus in Ohio [12], and lined up six financial firms to mobilize outside capital for the buildout, subject to definitive agreements [11]. The queue forms around megawatts and balance sheets as much as around dies. Worth noting that in the same quarter, the company sent approximately $26.0 billion back to its own shareholders [16], about 44% of GAAP net income [4][13], while organizing third-party money for its customers' infrastructure.
Jensen Huang's framing is that compute is revenue [14]. From the buy side, that reads as compute being somebody else's revenue at 74 points of margin, for as far ahead as the company is willing to guide [8].
Ranked by verification strength, evidence, and original report placement.
NVIDIA reported revenue for the second quarter ended July 26, 2026 of $96.2 billion, up 18% from the previous quarter and up 106% from a year ago.
GAAP net income for the quarter was $59,688 million.
Second-quarter Data Center revenue was $89.0 billion, up 18% from the previous quarter and up 117% from a year ago.
Data Center revenue in the prior quarter was $75.2 billion, according to runtimewire.
Third-quarter fiscal 2027 revenue is expected to be $108.0 billion, plus or minus 2%, and NVIDIA is not assuming any Data Center compute revenue from China in its outlook.
GAAP and non-GAAP gross margins for the third quarter are expected to be 74.0%, plus or minus 50 basis points.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Strong on the numbers, thin outside them
Every financial figure in the story traces to a dated issuer earnings release with specific revenue, margin, segment and capital-return lines, and the second publisher reproduces them consistently. But the cluster contains one primary document and one restatement of it: the consensus benchmarks are unattributed, no filing tables, unit volumes or pricing data are supplied, and the deployment and benchmark claims are vendor assertions without third-party confirmation.
Large disclosed consumption, vendor-sourced
Adoption evidence is substantial in scale — $89.0bn of Data Center revenue in a single quarter, $7.2bn of Edge Computing revenue, newest-generation racks named as running at five cloud providers, physical land and power secured in Ohio, and a named confidential-inference deployment. It is discounted because every observation originates from the vendor's own release: no customer, hyperscaler or independent benchmark confirmation appears in the cluster, and the $500bn financing capacity is announced intent conditioned on definitive agreements.
Modestly overstated at the framing layer
The quantitative core — $96.2bn, 75.0% realised and 74.0% guided margins, the $108.0bn guide — is disclosed, checkable and understates nothing, and the story's central inference that compute is not getting cheaper follows directly from the two margin percentages. The overshoot sits in the rhetoric and the aggregation: 'compute is revenue', 'no signs of slowing' and an over-$500bn financing headline that is contingent on unsigned agreements, plus a beat quantified against an unnamed consensus, all sit ahead of what the supplied evidence establishes.
Issuer-authored, unchallenged downstream
The originating document is a quarterly earnings press release from the company whose equity, pricing power and buildout narrative are the subject — the strongest possible commercial and market-facing incentive to present demand and margin favourably, including a product highlight reel and a partner-financing headline. The second publisher restates the same release without independent sourcing or a challenging voice, so nothing in the cluster offsets the issuer's incentive.
High on disclosed figures, lower on interpretation
Confidence in the reported and guided numbers is high: they are dated, specific, internally consistent between the two publishers, and consistent under the arithmetic checks (segment sums, sequential increments, implied gross profit). Confidence in the surrounding story is lower because there is effectively a single primary source, no independent corroboration of deployments or benchmarks, an unattributed consensus figure, and no volume or pricing data to test the pricing-power conclusion.
invest
Nvidia's pre-earnings re-rating is a credit story, not a chip story1 distinct publisher
invest
Nvidia's August 26 print: 92% of the quarter rides on one segment1 distinct publisher
leadership
Nvidia's $500B financing club puts the chip supplier inside the buyer's capital stack1 distinct publisher
product
The $500bn compute asset class rests on a depreciation curve Nvidia once denied1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 26, 2026
1 article · August 26, 2026