NEW YORK, March 17 (Reuters) – Nvidia has won approval from Beijing to sell its second most powerful artificial intelligence chips to China and is also preparing a version of the Groq AI chip that could be sold to the Chinese market, sources familiar with the matter said.
The long-awaited regulatory approval paves the way for the US chipmaker to resume sales of H200 chips, which have emerged as a major point of tension in US-China relations, in a market that once generated 13% of Nvidia’s total revenue.
Despite strong demand from Chinese companies and US approval for exports, Beijing’s hesitation to allow imports has been the main barrier to shipments of H200 chips to China.
On Tuesday, Nvidia Chief Executive Jensen Huang said the H200 had been licensed to ‘many customers in China’ and that it had received purchase orders from ‘many’ companies, allowing it to resume production of the chip.
The company had halted production of the chip last year due to growing regulatory hurdles in the US and China, according to a report at the time.
Nvidia had been waiting for licenses from the US and China for months. It has received some U.S. approvals, and a source familiar with the matter said the company has now also received licenses from Beijing for many customers in China.
Continues after advertising
A spokesperson for the Chinese embassy in Washington said it was ‘not aware of the details’ and referred questions to ‘the relevant authorities’.
CNBC also reported on Tuesday that Huang told them the company now has authorization from both the US and China.
In a filing with the US Securities and Exchange Commission late last month, Nvidia said the US had granted a license in February that would allow ‘small quantities of H200 products for specific customers based in China’.
Continues after advertising
NVIDIA PREPARES GROQ CHIP FOR CHINA
Nvidia is also preparing a version of the Groq AI chip that could be sold to the Chinese market, Reuters reported earlier on Tuesday, citing two sources familiar with the matter.
The company plans to use Groq chips for what’s known as inference, where AI systems answer questions, write code or perform tasks for users. In the products Nvidia showed off this week, the company plans to use its upcoming Vera Rubin chips, which cannot be sold in China, in combination with Groq chips.
Continues after advertising
While Nvidia dominates the AI systems training market, it faces much more competition in the inference market. Several large Chinese companies, including AI heavyweights like Baidu, already produce their own inference chips.
The chips being prepared for China are not downgraded versions or made specifically for the Chinese market, one of the sources told Reuters. But the new variant can be adapted to work with other systems, the source said, adding that the Groq chip should be available in May.
Nvidia did not immediately respond to a request for comment.
Continues after advertising
