Amid the continued surge in global demand for AI computing power, chip giant and one of the world's most valuable companies, NVIDIA, announced its latest quarterly financial results on Wednesday, delivering a record profit and revenue performance. According to AIbase, the company achieved revenue of $6.8 billion in a single quarter, an increase of 73% year-on-year; of which, the data center business contributed $6.2 billion, accounting for more than 90%. Annual revenue reached $21.5 billion, setting a new high.

From a structural perspective, NVIDIA has disclosed for the first time more detailed breakdowns of its data center revenue: $5.1 billion from GPU-based computing services, and $1.1 billion from networking products including NVLink. The strong growth is driven by the continuous expansion of demand for generative AI training and inference.
CEO Jensen Huang stated directly during the earnings call: "The global demand for tokens has grown exponentially." He pointed out that even GPUs deployed in the cloud six years ago are now fully occupied, with prices continuing to rise, indicating that the supply of computing power remains tight.
Notably, despite the recent removal of some export restrictions on chips to China by the U.S. government, NVIDIA did not record any revenue from the Chinese market this quarter. CFO Colette Kress mentioned that although a small number of H200 units were approved for export to Chinese customers, no revenue has been generated yet, and there is still uncertainty about whether future imported products will be allowed into China. She also noted that, driven by recent IPOs, Chinese competitors such as Moore Threads are accelerating their progress, which may have long-term impacts on the global AI industry structure.
In terms of capital, Huang also mentioned NVIDIA's planned investment of approximately $3 billion in OpenAI, stating that both parties are nearing an agreement. However, the documents submitted to the U.S. Securities and Exchange Commission also emphasized that "it cannot be guaranteed" that the transaction will ultimately be completed. Additionally, he mentioned that collaborations with Anthropic, Meta, and xAI continue to deepen.
Responding to market concerns about the sustainability of capital spending by tech companies, Huang said that the logic of the AI era has changed: "In the new world of artificial intelligence, computing is revenue. Without computing, you can't generate tokens; without tokens, you can't increase revenue." He believes the industry has passed a key turning point, with computing power directly transforming into profitable business outcomes.
With the wave of generative AI, NVIDIA is further solidifying its leading position in the global computing infrastructure sector. The competition surrounding computing power supply, international market competition, and ecosystem investment strategies will become core variables in the next phase of industry development.
