Nvidia leads sales after report of several US AI chip edges

(Bloomberg) — Nvidia Corp . led declines in tech stocks after a report that Washington could close loopholes in sales to China of powerful chips used to train artificial intelligence, potentially reducing sales in the world’s top semiconductor market.

Most read from Bloomberg

Nvidia, which gets about a fifth of its revenue from China, fell as much as 3.2% in after-hours trading in New York. Rival Advanced Micro Devices Inc. fell about 3%. The two lead the market for chips that are crucial to the development of generative AI models like ChatGPT. In China, traders unloaded a number of AI-related stocks, pushing Inspur Electronic Information Industry Co. and Unisplendour Corp. ̵[ads1]1; both major hardware suppliers – down 10%.

Nvidia this year designed less capable chips that fall below thresholds that require a license from the Commerce Department before exporting to China or other countries of concern. Washington is now considering action as soon as next month to expand curbs to include the lower-power semiconductors, the Wall Street Journal reported, citing anonymous sources.

Such a move underscores the Biden administration’s determination to limit China’s technological progress and could create tensions between the two countries. The US is increasingly concerned about Beijing’s technological ambitions, including around the use of AI in military and scientific advances that could tip the geopolitical balance.

While that is likely to hurt Nvidia and AMD’s business with the world’s No. 2 economy, the two chipmakers remain at the forefront of a surge in AI development that is driving investment from the United States to Europe and China. From Microsoft Corp. to Baidu Inc. and ChatGPT developer OpenAI, companies around the world are buying their products to train the next generation of artificial intelligence services.

Read more: Billionaires and bureaucrats mobilize China for AI race with US

U.S. officials are also considering whether to limit the leasing of cloud services to Chinese AI companies, which have used such platforms to train their models, the Journal reported, citing people familiar with the discussions. Inc. and Microsoft are among the world’s largest cloud service providers. Nvidia declined to comment on the Journal report.

“Chinese AI firms may also be able to obtain dedicated AI chips from third-party countries. So I think it will be difficult for the US to enforce the regulations,” said Robert Lea, an analyst at Bloomberg Intelligence. “Although further restrictions could delay development of AI from Chinese firms, I don’t see much long-term impact as Chinese firms take an increasingly innovative approach to solutions.”

Read more: Nvidia became ChatGPT’s brain and joined the $1 Billion Club

The world’s most valuable chipmaker has more than 80% market share for so-called data center accelerator chips, and had been operating under rules that required approval for shipments to China of A100 and new H100 parts. It was able to partially mitigate the impact on its finances by selling a modified version of the A100 that is slower to access data and therefore did not trigger the restriction.

An AI accelerator is a graphics processor, or GPU, tailored to train AI models by feeding them tons of data. It is better suited for such tasks than a general purpose CPU because the architecture can perform parallel work in large volumes. Nvidia was the first to come up with a language for making GPUs perform AI tasks, giving it a big lead over rivals such as AMD and Intel Corp.

AMD was also subject to the first rules announced last August. While Nvidia’s much smaller rival emerged largely unscathed from the crackdown, Nvidia said in September that the restrictions would cost it $400 million in that quarter.

The report “may have hurt sentiment today,” said Vey-Sern Ling, chief executive of Union Bancaire Privee.

(Updates with Asian market reaction from second part)

Most read from Bloomberg Businessweek

©2023 Bloomberg LP

Source link

Back to top button