Nvidia has posted another blockbuster quarter, fueled by the seemingly insatiable appetite for AI computing. The chipmaker — one of the world’s most valuable companies — says demand is stretching across the entire stack, from cutting-edge accelerators to older GPUs still running in cloud fleets.
On the company’s earnings call, CEO Jensen Huang framed the moment in unusually blunt terms: “The demand for tokens in the world has gone completely exponential,” he said, adding that even Nvidia’s “six-year-old GPUs in the cloud are completely consumed” as pricing rises.
For the quarter, Nvidia reported $68 billion in revenue, up 73% from a year earlier. The company said $62 billion of that total came from its data center segment — the business that has become Nvidia’s AI engine room.
Nvidia also provided more detail on what’s driving that data center number. It split the segment into $51 billion in compute revenue (largely GPUs) and $11 billion in networking, including products like NVLink.
Zooming out, Nvidia reported $215 billion in revenue for the full year, underscoring how quickly AI infrastructure spending has reshaped the company’s financial profile — and the broader semiconductor industry’s priorities.
China remains a complicated part of the story. Nvidia did not report revenue from chip exports to China, and CFO Colette Kress said that while small amounts of H200 products for China-based customers were approved by the U.S. government, they have not yet produced revenue — and the company still lacks clarity on what will ultimately be allowed.
Kress also warned that domestic Chinese competitors are gaining momentum, pointing to a recent wave of IPO activity — including Moore Threads’ IPO in December — as a sign that local players may increasingly challenge the global AI hardware landscape over the long term.
Huang used the call to address Nvidia’s ongoing talks with OpenAI as well. “We continue to work with OpenAI toward a partnership agreement. We believe we are close,” he said, also referencing partnerships with Anthropic, Meta, and Elon Musk’s xAI. Still, Nvidia’s filings emphasized there is “no assurance” that an investment will happen, despite reports that the figure under discussion could be as high as $30 billion.
Finally, Huang pushed back on concerns that Big Tech’s AI spending spree might be unsustainable. His argument was simple: in the AI economy, compute isn’t just a cost center — it’s the product.
“In this new world of AI, compute is revenue,” Huang said. “Without compute, there’s no way to generate tokens. Without tokens, there’s no way to grow revenues,” adding that the industry has reached an inflection point where “profitable tokens” can be productive for customers and profitable for cloud providers.