Nvidia’s latest results put the AI infrastructure boom in plain view. The chip giant reported record profits in its most recent quarter on Wednesday, with demand for AI compute continuing to drive the company’s growth.
The numbers also show how much of Nvidia’s business now runs through data centers, where cloud providers and AI companies need GPUs, networking products, and other systems to generate and deliver tokens at scale.
AI demand is still pulling Nvidia higher
Nvidia reported $68 billion in revenue in the most recent quarter, up 73% from the prior year. Of that total, $62 billion came from the company’s data center business.
That data center figure matters because it shows where the company’s growth is concentrated. Nvidia also broke the segment into two major pieces: $51 billion in compute revenue, largely GPUs, and $11 billion in networking products like NVLink.
For the full year, Nvidia reported $215 billion in revenue. The results reinforce the company’s position as both a chip giant and the world’s most valuable company, with AI infrastructure demand sitting at the center of its business.
CEO Jensen Huang described the demand environment in unusually direct terms on a call with analysts after the results.
"The demand for tokens in the world has gone completely exponential," CEO Jensen Huang said. "I think we’re all seeing that, to the point where even our six-year-old GPUs in the cloud are completely consumed and the pricing is going up."
That comment frames Nvidia’s quarter around a simple idea: AI services need compute to produce tokens, and demand for that output is putting pressure on available hardware, including older GPUs already deployed in cloud environments.
Data center revenue defines the quarter
The most important detail in Nvidia earnings is not only the headline revenue number. It is the shape of the revenue.
With $62 billion of quarterly revenue tied to data centers, Nvidia’s results show that the company’s AI compute business is not a side story. It is the main engine.
The split between compute and networking also adds useful context. The $51 billion compute figure reflects the central role of GPUs in AI workloads. The $11 billion networking figure shows that the systems around those GPUs, including products like NVLink, are also a major part of the business.
That distinction is important because large AI deployments are not only about buying individual chips. They require infrastructure that can connect compute resources and move data efficiently. Nvidia’s reported results show revenue flowing from both sides of that equation.
China remains a complicated market
One notable absence from the quarter was revenue from chip exports to China. Nvidia did not report any revenue from those exports, even after the recent lifting of export restrictions by the U.S. government.
Colette Kress, Nvidia’s chief financial officer, said small amounts of H200 products for China-based customers had received U.S. government approval, but that approval has not yet turned into sales.
"While small amounts of H200 products for China-based customers were approved by the U.S. government, they have yet to generate any revenue, and we do not know whether any imports will be allowed into China," Colette Kress, the company’s chief financial officer, said.
Kress also pointed to a longer-term competitive risk in China. She said competitors there, helped by recent IPOs, are advancing and may eventually affect the structure of the global AI industry. The source described the remark as an apparent reference to Moore Threads’ IPO in December.
For Nvidia, that means the China issue is not only about whether approved products can generate revenue in the near term. It is also about whether domestic Chinese competitors could become more disruptive over time.
OpenAI talks are still not final
During the investor call, Huang also discussed Nvidia’s pending investment in OpenAI, which has been reported at $30 billion. He said the companies were still working toward a partnership agreement.
“We continue to work with OpenAI toward a partnership agreement. We believe we are close,” Huang said.
Huang also referenced partnerships with Anthropic, Meta, and Elon Musk’s xAI. Those relationships show how Nvidia is positioned across several major AI companies and platforms.
At the same time, Nvidia’s statements filed with the U.S. Securities and Exchange Commission on Wednesday emphasized that there was "no assurance" an investment would take place. That language keeps the OpenAI deal in the category of a possible partnership, not a completed transaction.
Capex concerns meet Nvidia’s revenue argument
The quarter also comes as investors and analysts continue to examine tech companies’ capex commitments. The concern is straightforward: companies are spending heavily on AI infrastructure, and the market wants to know whether that spending can translate into sustainable revenue.
Huang’s answer was that AI compute investments are becoming directly tied to revenue generation. In his view, compute is not just a cost center in the new AI market. It is the capacity needed to create the token output that AI businesses sell or use.
"In this new world of AI, compute is revenue. Without compute, there’s no way to generate tokens. Without tokens, there’s no way to grow revenues," Huang said. "We’ve reached the inflection point and we’re generating profitable tokens that are productive for customers and profitable for the cloud service providers"
That argument is central to Nvidia’s position. If demand for AI tokens keeps rising, then the need for GPUs, networking products, and data center systems remains closely linked to growth across the AI market.
Nvidia’s latest quarter does not resolve every open question. China revenue is still absent, the OpenAI investment is not assured, and capex sustainability remains a major debate. But the reported results show that, for now, AI compute demand is still producing record numbers for Nvidia.