Despite U.S. limitations on the shipment of two of Nvidia's top chips to China, Nvidia Corp Chief Executive Jensen Huang stated on Wednesday that he still sees a sizable market for Nvidia's data center chips in the nation.
Nvidia announced on Sept. 1 that the U.S. government has instructed it to halt exporting its A100 and H100 chips to China, which could have a negative impact on the company's revenues by up to US$400 million in the current fiscal quarter. The two devices, which are used in data centers to accelerate artificial intelligence activities like natural language processing, are Nvidia's two fastest CPUs.
Huang explained the limitations revealed earlier this month at a press conference following the company's autumn product launch. He stated the limitations have defined levels for both the performance of a chip and the processor's capacity to connect additional chips.
According to him, there is "a large space for us" in the Chinese market due to the laws. "The vast majority of our customers are not affected by the specification," Huang said. "So our expectation is that for the United States and also for China, we will have a large number of products that are architecturally compatible, that are within the limits and that require no license at all."
Huang explained these chips are a part of bigger chip lines with "large number of products" that can still be sold in China. Huang made this statement during the news conference. Huang added that Nvidia would approach the US government for permits if Chinese clients wanted its top chips.
"You could surmise that the goal is not to reduce or hamper our business," Huang said. "The goal is to know who it is that would need capabilities above this limit and give the United States the opportunity to make a decision about whether that level of technology should be available to others."
Nvidia informed investors on August 26 that the US government had informed it that a new license requirement was intended to address a risk that certain products would be "used in, or diverted to, a military end use' or "military end users" in China and Russia." The US trade restrictions on AI chips affect both Nvidia and AMD.
The limitation is anticipated to have an effect on Nvidia A100 and H100 Tensor Core GPU chips for machine learning acceleration utilized in data center operations like natural language processing.