Nvidia posted $96.2 billion in revenue last quarter and guided to $108 billion for the next one, putting the chipmaker within a single quarter of joining the exclusive club of companies that book more than $100 billion every 90 days. Data center revenue alone hit $89 billion, more than doubling year-over-year. Profits also more than doubled, to $59.7 billion. The sequential jump from the prior quarter was over $10 billion.
Only three companies have ever cleared $100 billion in a single quarter: Amazon, Apple, and Alphabet. Nvidia, whose entire business model a decade ago was selling graphics cards to gamers, is now roughly one earnings report away from making it four. The forward guidance implies another $12 billion sequential step-up, which is itself larger than the total quarterly revenue of most public tech companies.
The mix tells the story of where AI capex is actually landing. Data center accounts for $89 billion of the $96.2 billion total — roughly 92% of the business. That segment doubled year-over-year, driven by hyperscaler and frontier-lab demand for GPUs to train and serve models from OpenAI, Anthropic, Google, Meta, and Microsoft. Nvidia's chips remain the default substrate for every serious AI workload, and the company is capturing an outsized share of the dollars flowing into the buildout.
Key facts
- 01Nvidia reported $96.2B in quarterly revenue, a $10B jump from the prior quarter.
- 02Data center revenue hit $89B, more than doubling year-over-year.
- 03Profits more than doubled to $59.7B in the quarter.
- 04Nvidia guided to $108B in revenue next quarter, on track to cross $100B.
- 05Edge computing, including consumer gaming, brought in $7.2B, up 27% year-over-year.
The consumer side, which Nvidia now categorizes as "edge computing" and includes its gaming GPU business, brought in $7.2 billion. That's up 27% year-over-year, but Nvidia acknowledged that it experienced "slower consumer PC sales that were tempered by elevated memory and systems prices." Component shortages are pushing GeForce prices higher, and gamers are increasingly the segment getting squeezed as the company allocates capacity to data center customers willing to pay data-center prices.
Nvidia warned ahead of the earnings report that its AI chips are about to get more expensive as well. That is a notable admission at a moment when hyperscalers are already spending record sums on infrastructure and when several of Nvidia's largest customers — Microsoft, Meta, Google, and increasingly Anthropic — are developing custom silicon precisely to reduce their exposure to Nvidia pricing.
Anthropic disclosed earlier this month that it is developing custom AI chips for Claude, joining a list of frontier labs and cloud providers trying to build alternatives. OpenAI has been publicly touting its own Jalapeño inference chip. None of that has dented Nvidia's numbers yet — the $89 billion data center print suggests the custom-silicon threat remains a multi-year project rather than a near-term revenue drag.
The gap between data center and everything else keeps widening. At $89 billion versus $7.2 billion, Nvidia's AI infrastructure business is now more than 12 times the size of its consumer graphics business in a single quarter. A decade ago that ratio was inverted. The company's identity as a gaming hardware brand is now a rounding error against its identity as the pick-and-shovel supplier of the AI economy.
The counterweight worth naming: guidance of $108 billion assumes that hyperscaler AI capex holds, that the custom-silicon programs at Microsoft, Meta, Google, and Anthropic don't ship in volume soon, and that component pricing for HBM memory and advanced packaging doesn't blow out further. Nvidia's own admission of "elevated memory and systems prices" is a small crack in that story. If any of those variables move against the company, the $12 billion sequential step-up becomes considerably harder to underwrite.
Nvidia is at this point a proxy for the entire AI capex cycle. When a single company's data center segment doubles year-over-year to $89 billion, it means the hyperscalers and labs writing those checks are still convinced that the returns on frontier model training and inference will justify the spend. The moment that conviction cracks — at any one of the top five customers — Nvidia's guidance becomes the most-watched number in tech. For now, the trajectory points at $100 billion a quarter, and then past it.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
