Nvidia reported $81.6 billion in revenue for the quarter ending April 26, up 85% compared to the same period a year ago. Of that, $75.2 billion came from data centers. For context, Nvidia's entire annual revenue in 2023 was around $27 billion. It now generates three times that in a single quarter.
The company expects $91 billion in revenue for the current quarter. China is effectively zeroed out of that forecast. Export restrictions have blocked sales of its most advanced chips to Chinese customers, and Nvidia said it has generated no revenue from those shipments yet.
The revenue numbers are extraordinary but not surprising to anyone who has been following AI spending. The more interesting story is elsewhere in the filing.
Nvidia's holdings in privately held companies nearly doubled in a single quarter, from $22 billion to $43 billion. The company spent $18.5 billion on these investments in just three months. The previous quarter, it had spent $649 million. That is a 28-fold increase in startup spending from one quarter to the next.
The single largest deal is a $30 billion investment in OpenAI, the company behind ChatGPT. Nvidia also invested in Anthropic and participated in roughly two dozen other private funding rounds in 2026 alone. On top of that, the company committed up to $3.2 billion in glassmaker Corning and up to $2.1 billion in data center operator IREN, neither of which is counted in the $43 billion figure because they are publicly traded.
Jensen Huang, Nvidia's CEO, framed the strategy plainly on an investor call: Nvidia is also ramping up computing infrastructure specifically for Anthropic, a buildout that was essentially zero until recently.
The question that a growing number of analysts are asking is whether this is sustainable. The concern goes like this: Nvidia invests in companies that are losing money. Those companies use Nvidia's investment capital to buy Nvidia chips. Nvidia records that as revenue. A Wedbush Securities analyst described Nvidia's activity as fitting "squarely into the circular investment theme." Short sellers, including Jim Chanos and Michael Burry of "The Big Short" fame, have gone further, drawing comparisons to how Lucent Technologies propped up struggling telecom customers during the dot-com bubble, who then used that money to purchase Lucent equipment.
Nvidia has pushed back directly. The company published a detailed rebuttal noting that its customers pay for chips within 53 days of purchase, and that the investments are strategic, not financing schemes. Other analysts, including those at Futurum Group, agree, arguing that using free cash flow to fund the companies building on your platform is rational business.
Both readings can be true at once. The revenue is real. The chips are being shipped. Payments are being made. At the same time, a significant portion of the money flowing to Nvidia's customers originated from Nvidia itself. As long as AI spending keeps growing at this pace, the loop holds. If it slows, the interdependencies become very visible very quickly.
For anyone running a business that depends on AI tools, the takeaway is straightforward. The companies whose products you use, from ChatGPT to cloud AI services, are being financially supported by the same company that makes the hardware they run on. The AI supply chain is far more vertically concentrated than it appears on the surface, and Nvidia is the thread running through almost all of it.