Nvidia reported fiscal second-quarter revenue of $96.2 billion on Wednesday, more than doubling year-ago figures and handily beating Wall Street estimates, as insatiable demand for AI compute overwhelmed concerns about a potential spending slowdown.
The results showed 106 percent year-over-year revenue growth and 18 percent sequential growth from the prior quarter, sending a clear signal that enterprise and hyperscaler spending on AI infrastructure continues to accelerate despite persistent doubts from some corners of the market. Data center revenue alone hit $89 billion, up 117 percent from a year ago, accounting for over 92 percent of total company sales.
GAAP earnings per diluted share came in at $2.46, up 128 percent from the prior year. Non-GAAP EPS was $2.22, beating analyst consensus of approximately $2.09. Gross margins held steady at 75 percent, while net income under GAAP reached $59.7 billion for the quarter ending July 26. Operating income was $63.7 billion, up 124 percent year-over-year.
Huang: Compute Is Revenue Now
AI has reached its inflection point. It is doing useful work. Its tokens are productive and profitable. Now, compute is revenue. And demand is accelerating. – Jensen Huang, Nvidia CEO
Huang emphasized that the AI market has moved beyond the training phase into massive deployment across enterprises worldwide. The reasoning models driving agentic AI applications consume orders of magnitude more compute than previous generations, creating a self-reinforcing demand cycle that shows no sign of slowing down. He pointed to the rapid adoption of Nvidia platforms across cloud providers, sovereign AI projects, and enterprise customers as evidence that AI spending is structural rather than cyclical.
The CEO also addressed the growing competitive threat from custom silicon being developed by Nvidias own customers. OpenAI recently unveiled its Jalapeno inference chip at the Hot Chips conference, claiming 1.5 to 1.9 times the performance per watt of Nvidias flagship. Anthropic has hired Google chip veteran Amir Salek to build its own semiconductor team. Amazon, Google and Microsoft all operate custom AI chip programs. Huang argued that Nvidias advantage lies in its full-stack platform approach rather than any single chip specification.
$279 Billion Bet on Supply
The most striking number in Nvidias filing was $279 billion in manufacturing, supply and capacity commitments. That figure has nearly tripled from $95.2 billion at the end of January and $119 billion in the prior quarter, representing an unprecedented bet on continued AI infrastructure expansion over the coming years.
On the earnings call, Huang said the company supply chain is gigantic but still insufficient to meet demand. He noted that half the market comes from hyperscalers like Microsoft, Google and Amazon while the other half encompasses enterprise buyers, neoclouds and sovereign AI projects that operate largely outside public view. This second half of the market, he said, does not buy custom chips one at a time but rather purchases Nvidias full platform.
| Metric | Q2 FY27 | Q2 FY26 | Y/Y Change |
|---|---|---|---|
| Revenue | $96.2B | $46.7B | +106% |
| Data Center | $89.0B | $41.0B | +117% |
| Gross Margin | 75.0% | 72.4% | +2.6 pts |
| GAAP EPS | $2.46 | $1.08 | +128% |
| Net Income | $59.7B | $26.4B | +126% |
Groq 3 LPX Enters Production
Nvidia also announced that its Groq 3 LPX rack-scale system is now in full production, with Nebius as the first customer. The system demonstrated nearly four times the tokens per second of the nearest competitor on the Artificial Analysis benchmark, positioning it as the new standard for high-throughput inference. Volume shipments are expected later this quarter.
The announcement came days after the Hot Chips conference, where Nvidias Rubin architecture faced its first direct competitive comparison from OpenAIs Jalapeno chip and Intels 256-core Xeon 7 processor. Despite the noise from competitors, Nvidias financial results suggest its market position remains firmly entrenched, with customers choosing the full Nvidia platform over individual chip specifications.
Shares rose modestly in after-hours trading, though Bloomberg noted that Nvidias stock has fallen the day after five of its last six earnings reports, suggesting that even blowout numbers may be largely priced in for the AI bellwether. The stock is up roughly 20 percent in the past month.
discussion