Nvidia data center revenue nearly doubled to $75.2 billion in its fiscal first quarter, but the stock barely budged. The reason is hiding in plain sight: Nvidia's biggest customers are becoming its fiercest rivals. The chipmaker reported adjusted earnings of $1.87 per share on revenue of $81.62 billion for the quarter ended April, crushing Wall Street estimates of $1.76 and $78.86 billion respectively. Data center revenue surged 92% year over year to $75.2 billion, accounting for more than 90% of total sales.
Gaming, once Nvidia's flagship business, now contributes less than 8%. But shares drifted 1% lower in after-hours trading. The muted reaction reflects a market that has already priced in the beat and is now laser-focused on the risks ahead.
Nvidia acknowledged those risks in a 10-Q filing, warning that customers developing custom ASICs "may not require all of the features and functionality our data center systems provide." Google, Amazon, Meta, and Microsoft are all designing their own AI chips. Meta showcased four custom chips in March that Taiwan Semi will manufacture.
Google confirmed Monday it will create a new AI infrastructure company centered around its TPU chips, with Blackstone investing $5 billion in equity capital.
"If we are unable to successfully compete in this environment, demand for our products could decrease," Nvidia said in the filing.
China revenue has collapsed to zero. CFO Colette Kress said Nvidia shipped no Hopper products to China during the quarter, compared with $4.6 billion in the year-ago period. The US approved H200 exports to Chinese customers, but Beijing has discouraged purchases to protect domestic chip maker Huawei. "We have yet to generate any revenue, and we are uncertain whether any imports will be allowed into the country," Kress said.
CEO Jensen Huang previously estimated the Chinese AI market at $50 billion annually.
Nvidia is restructuring how it reports results to address the shifting market. The company now splits its business into two segments: Data Center and Edge Computing. The Data Center segment covers hyperscalers plus a new ACIE category (AI Clouds, Industrial, Enterprise), which saw revenue triple year over year. Edge Computing includes data processing devices for agentic and physical AI, plus gaming GPUs. That segment delivered $6.4 billion, up 29%. On the call, Huang made the bull case with characteristic bluntness. He defended his earlier forecast that cumulative demand for Nvidia's most advanced chips would reach "at least $1 trillion" through the end of 2027.
"Compute capacity is profits," he said. "This is the way computing is going to work in the future. And if they don't have the compute, they won't have the revenues."
Huang framed agentic AI as the next growth catalyst, predicting "billions of agents" will drive compute demand. Each agent will spawn sub-agents, each requiring GPU and CPU cycles.
"The world has a billion users," Huang said. "The world will have billions of agents."
Nvidia's next-generation Vera Rubin system is the near-term catalyst. Kress confirmed Vera Rubin will ship during fiscal Q3 and ramp in Q4. The system delivers 10 times more performance per watt than its predecessor. Huang called it "even more successful than Grace Blackwell," citing growing inference share and a new partnership with Anthropic.
Hyperscaler revenue hit $38 billion in the quarter, up 12%, accounting for half of data center sales. The other $37 billion came from the ACIE segment.
Sovereign AI crossed $30 billion in fiscal 2026, more than triple the prior year, providing what analysts describe as a hedge against hyperscaler concentration risk.
Nvidia also returned cash to shareholders, authorizing an $80 billion stock buyback and raising its quarterly dividend from 1 cent to 25 cents per share. The company guided Q2 revenue to $91 billion, well ahead of the $87.2 billion consensus. The stock opened 0.5% lower Thursday, a garden-variety fade for a company that has conditioned markets to expect perfection.
