Americas

  • United States
Anirban Ghoshal
Senior Writer

Nvidia’s new chips for China to be compliant with US curbs: Jensen Huang

News
Dec 06, 20233 mins
CPUs and ProcessorsTechnology Industry

Nvidia’s AI-focused H20 GPUs bypass US restrictions on China’s silicon access, including limits on-chip performance and density.

Nvidia Jensen Huang Computex
Credit: Nvidia

Nvidia’s new chips designed for the Chinese market are expected to be compliant with the latest restrictions around chip exports set forth by the Biden administration, according to company CEO Jensen Huang.

“Nvidia has been working very closely with the US government to create products that comply with its regulations,” Huang was quoted as saying in a Reuters news report.

Huang said that the company intends to come out with a new set of products that do not breach the restrictions.

Huang’s remarks come just weeks after it delayed the launch of its made-for-China chips, before which news reports surfaced that the chipmaker could be trying to circumvent US restrictions that aim to prevent Chinese companies from stringing together lesser-powered GPUs to build compute capacity for AI-based workloads. 

According to a news report from Chinastarmarket, which cites sources in the industrial supply chain put in place by Nvidia to manufacture GPUs, Nvidia was in the process of developing the H20, L20, and L2 chips, which were expected to go on sale from November 16.

The new restrictions by the US government include a limit on total chip performance at 4,800 TPP (total performance) and a performance density of 5.92.

While the total chip performance limitation blocks the sale of high-powered GPUs — such as the A100 and the H100, which are sold in China as the A800 and the H800 with modified bidirectional bandwidth rates — the performance density threshold of a chip was set by the Biden administration to block the loophole of Chinese enterprises and organizations buying GPUs with less performance power and stringing them together to build compute capacity for AI-based workloads.

Performance density, according to experts, is the total performance power of a chip divided by its die area.

As per the latest export restrictions, any chip with a performance density of 5.92 or above stands for an absolute export ban. Companies need to apply for a license for chips with performance densities between 3.2 and 5.92.

Due to these restrictions, Nvidia expects to see a drop in revenue from its Chinese market, the company said during its earnings call in November. 

“We expect that our sales to these destinations will decline significantly in the fourth quarter, though we believe will be more than offset by strong growth in other regions. The US government designed the regulation to allow the US industry to provide data center compute products to markets worldwide, including China,” Colette Kress, chief financial officer at Nvidia, was quoted as saying in a transcript from The Motley Fool.

However, the top executive said that though China accounted for 20% of its revenue, gauging the impact of the restrictions would be difficult presently.

Huang, according to a Bloomberg report, also said that he saw Huawei and Intel as “formidable” rivals in the chipmaking space.

The CEO was visiting Singapore to lead talks about the city developing its large language model and some potential investments.