Nvidia’s $96B quarter turns the AI boom into industrial-scale infrastructure

The quarter that reset the scale

Nvidia’s latest earnings no longer look like a semiconductor cycle; they look like an infrastructure buildout measured in tens of billions of dollars every three months. For the quarter ended July 26, 2026, the company reported revenue of $96.221 billion, up 18% from the prior quarter and 106% from a year earlier, with GAAP diluted earnings per share of $2.46 and non-GAAP diluted EPS of $2.22 [1]. GAAP operating income rose to $63.734 billion, a 124% year-over-year increase, and net income reached $59.688 billion [1]. The company also told investors to expect third-quarter revenue of $108.0 billion, plus or minus 2%, while explicitly excluding any China data-center compute revenue from that outlook [1].

The cleanest way to read the print is that Nvidia’s core AI engine is still accelerating at a scale few technology suppliers have ever reached. Data Center revenue was $89.0 billion, up 18% sequentially and 117% year over year, meaning the segment represented roughly 92.5% of total quarterly sales [1]. Edge Computing, the smaller segment that now includes areas such as PCs, robotics and other AI-at-the-edge uses, posted $7.2 billion in revenue, up 13% from the prior quarter and 27% from a year earlier [1].

Expectations were high, and Nvidia still cleared them

The surprise is not merely that Nvidia beat forecasts; it is that the beat came when expectations were already stretched. AP reported that revenue of $96.22 billion topped analysts’ average forecast of $92.27 billion, while adjusted EPS of $2.22 exceeded the FactSet consensus of $2.09 [2]. MarketScreener, citing MT Newswires and FactSet, reported that Data Center revenue of $89 billion exceeded the $86.3 billion analyst consensus and that Edge Computing revenue of $7.2 billion topped a $6.55 billion consensus [6].

That matters because Nvidia has become a live proxy for whether artificial intelligence spending is still rising or approaching exhaustion. Reuters described the company as a bellwether for the AI market because its chips power most major data centers and advanced AI models globally [3]. The stock reaction showed how demanding the market has become: Reuters reported that shares first dipped in after-hours trading, then rose 4.2% in heavy volume after executives expanded on the company’s fiscal 2028 growth outlook during the call [3].

The guide is the real headline

The $108 billion third-quarter guide is arguably more important than the $96 billion quarter because it implies that Nvidia is entering its first routine $100 billion-plus quarterly revenue period. Axios reported that the projection would mark the first time Nvidia exceeded $100 billion in a quarter [5]. AP noted that if Nvidia hits the target for the August–October period, revenue would be about 89% higher than a year earlier, suggesting that growth is still very fast even after the company doubled its latest quarter [2].

Management also gave investors an unusual look into the next fiscal year. On the earnings call, CFO Colette Kress said Nvidia expects revenue to grow by about 70% in fiscal 2028, and she framed that as a supply-constrained outlook rather than a demand ceiling [4]. Jensen Huang told analysts that demand was higher than the 70% supply-supported forecast and that the company had “a huge year” ahead as it worked to secure more capacity across the supply chain [4].

Hyperscalers remain central, but the customer map is widening

The numbers still begin with the largest cloud platforms. On the call, Nvidia said hyperscale revenue was $49 billion, up 13% sequentially, driven by Blackwell demand [4]. The company also pointed to cloud backlog above $2 trillion and said capital spending by the top five hyperscalers is expected to reach nearly $800 billion in 2026 and $1.3 trillion in 2027 [4]. AP separately reported Kress’s comment that top-five hyperscaler capital spending is expected to approach those levels [2].

Yet Nvidia is trying to persuade investors that this is not only a Microsoft-Amazon-Google-Meta cycle. Kress said ACIE revenue — a category covering AI clouds, industrial and enterprise customers — was $40 billion, up 25% sequentially and 138% from a year earlier [4]. Huang said the non-hyperscaler side, including sovereign AI, regional AI, neoclouds, startups and enterprises, represents roughly half of Nvidia’s business and is growing about 100% a year [4]. If that mix holds, Nvidia’s growth story is less dependent on a handful of U.S. megacap capital budgets than the headline AI-capex debate sometimes suggests.

Vera Rubin and AWS show the supply chain’s next phase

Product transition risk is the key operational question, and Nvidia is presenting Vera Rubin as both a demand driver and a capacity constraint. The company said in its release that Vera Rubin is ramping into full production with racks running at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius [1]. On the call, Kress said production shipments of Vera Rubin began earlier in August, that purchase orders had been received from every major hyperscaler, AI cloud and system OEM, and that Vera Rubin is expected to account for about 20% of Data Center revenue in Q3 [4].

The most concrete customer expansion came from Amazon Web Services. Nvidia and AWS announced that AWS plans to deploy 2 million additional Nvidia GPUs across its global infrastructure in 2027 and 2028, alongside plans to bring Nvidia Vera CPU-based infrastructure to AWS and deepen work across networking, open models, data processing and robotics [5]. The call framed the deployment as beginning in the current quarter and running through Nvidia’s fiscal Q2 2029, with some Vera CPUs integrated with Rubin and others deployed as standalone processors [4].

Margins, memory and China are the pressure points

The quarter was exceptional, but not frictionless. Nvidia’s GAAP and non-GAAP gross margins were both 75.0% in Q2, but the company guided Q3 gross margin to 74.0%, plus or minus 50 basis points [1]. Kress said on the call that margins are expected to bottom in the 71%–72% range in Q4 before settling around 72%–73% in fiscal 2028, attributing the reset to extreme memory pricing conditions that exceeded prior expectations [4].

China remains another unresolved variable. Nvidia’s official outlook assumes no China data-center compute revenue in Q3 [1]. On the call, Kress said Nvidia shipped less than 1% of total Data Center revenue in Hopper 200 products to China-based customers under U.S. government licenses during Q2, and she said current Hopper shipments were dilutive to corporate gross margins [4]. The exclusion does not mean China demand is irrelevant; it means management is choosing not to count on it while geopolitics remains unstable.

A boom with financing questions

Nvidia is also becoming more than a chip supplier. Kress said the company has invested nearly $50 billion in frontier AI labs and has partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR on financing platforms intended to raise more than $500 billion of third-party capital for AI infrastructure [4]. She acknowledged that some critics would call the approach circular financing, but argued that Nvidia sees the supported labs as once-in-a-generation companies with accelerating usage and customer demand [4].

That is the strategic tension now attached to Nvidia’s results. The quarter shows extraordinary real demand, high utilization and stronger-than-expected guidance. It also shows an AI economy increasingly organized around Nvidia’s ability to secure chips, memory, power, land, financing and customers at the same time. The $96.2 billion quarter is therefore not just a record. It is evidence that the AI supply chain has shifted from a product boom into an industrial capital cycle, with Nvidia at its narrowest and most profitable bottleneck.

Sources from the last 72 hours

1. [1]NVIDIA Announces Financial Results for Second Quarter Fiscal 2027 | NasdaqAug 26, 2026, 8:20 PM UTC
2. [2]Nvidia stock jumps after early dip following results; AI fever is unabatedAug 26, 2026, 9:32 PM UTC
3. [3]NVIDIA (NVDA) Q2 2027 Earnings Call Transcript & AudioAug 26, 2026, 12:00 AM UTC
4. [4]Nvidia projects 70% revenue growth in 2028Aug 26, 2026, 9:51 PM UTC
5. [5]AWS and NVIDIA to Deliver 2 Million Additional GPUs and Next-Generation Infrastructure for Agentic and Physical AI | NVIDIA NewsroomAug 26, 2026, 12:00 AM UTC
6. [6]Nvidia Q2 Data Center Revenue at $89 Billion vs FactSet Analyst Consensus of $86.3 Billion; Edge Computing Revenue at $7.2 Billion vs Consensus of $6.55 Billion | MarketScreenerAug 26, 2026, 9:37 PM UTC
7. [7]Strong AI chip demand powers Nvidia's Q2 results past Wall Street's expectations | AP NewsAug 26, 2026, 8:40 PM UTC

AI-generated article based on recent web research, then preserved as a dated editorial snapshot.