Unlock Editor’s Digest for free
Roula Khalaf, editor-in-chief of the FT, selects her favorite stories in this weekly newsletter.
Nvidia’s claim that its business can circumvent the effects of U.S. controls on semiconductor exports to China is expected to be tested when it reports quarterly results on Tuesday.
Investors have so far dismissed fears that Nvidia will suffer from tightening US restrictions on sales to China of advanced processors suitable for the development of large artificial intelligence systems. The chipmaker said the controls would not have a “significant near-term impact” on its business.
Since the Biden administration’s announcement in mid-October, Nvidia’s market valuation has lost and regained approximately $200 billion – the equivalent of Intel’s entire market capitalization – reaching an all-time high last week that valued the company at $1.2 billion.
Morgan Stanley analysts said in a research note last week that the impact of China’s restrictions was “the most asked earnings question” among Nvidia investors.
China accounts for as much as 25 percent of sales for Nvidia’s data center business, but the company confidently said last month that it does not yet expect a significant fallout from the new U.S. rules.
“Given the strong demand for our products around the world, we do not expect the additional restrictions to have a significant near-term impact on our financial results,” it said.
The results to be announced Tuesday cover the three months ending at the end of October, so the results themselves likely won’t reflect much of the rules change. But investors will carefully examine forward guidance.
Nvidia has attempted to circumvent U.S. restrictions on its chip sales in China with new processors that circumvent performance restrictions.
Earlier this year, it launched a modified version of its flagship H100 processor for Chinese customers, called the H800, following the first US restrictions announced in October 2022. Then, this month, the Financial Times reported that Nvidia had again adapted its products, with a new product for China called H20.
The company has not officially announced the H20 but has released its specifications to potential customers in China. The details have left observers at odds over whether Nvidia has been able to meet U.S. requirements while satisfying Chinese customer demand for AI chips.
Analysts and senior executives at two Chinese cloud computing providers said these details suggested the H20 may not provide enough computing power to effectively train AI systems equivalent to OpenAI’s latest model, GPT-4 . The chips would not allow Chinese companies to remain competitive with their U.S. counterparts, these people said — a crucial goal of the Biden administration.
Executives from Tencent and Alibaba suggested during their quarterly earnings calls last week that Chinese tech groups may not be able to rely on Nvidia’s new chips to drive their models. ‘AI. Both said they plan to focus more on domestic alternatives to Nvidia, with Huawei seen as the most likely beneficiary.
Eddie Yongming Wu, Alibaba’s chief executive, said he expected to see “multiple different chips used (by) multiple different vendors, meeting the demand for AI computing power in the Chinese market.” .
Charlie Chai, a Shanghai-based 86Research analyst, said it was “fundamentally difficult” to use the H20 to train large AI models, limited by a performance ceiling on a single chip.
Nvidia has long opposed U.S. controls on the grounds that keeping U.S. companies in China would only fuel the advance of local chipmakers.
Analysts and industry professionals say it would take years for Chinese tech companies to adapt their systems to use domestic chips instead of Nvidia’s. However, this process has been accelerated by the strengthening of export controls by the US government.
“Now everyone is forced to adapt Huawei’s chips and software,” Chai said, adding that Nvidia “may have underestimated Huawei’s capabilities.”
Still, not all analysts are convinced that Nvidia’s business will suffer, at least over the next year.
Morgan Stanley said it was “somewhat surprised to report that the near-term impact of the new controls is minimal”, in part because – as Nvidia itself indicated – the demand for its AI processors was well ahead of supply in other parts of the world. .
Dylan Patel, chief analyst at chip consultancy SemiAnalysis, said that despite public comments from Chinese companies, they had ordered a “very significant quantity” of H20 chips, which, despite their limitations, were still more capable than the Nvidia A100 chips that OpenAI used to train. previous generations of GPT. He estimated that Chinese customers would spend about $15 billion on H20-based systems next year.
“Hundreds of thousands of H20s are going to be manufactured and sold,” Patel said. “It’s not optimal, but it’s the best chip (for AI development) that China can buy.”
Nvidia declined to comment.
Gn En bus