Jensen Huang, CEO of Nvidia, which faces tight restrictions on selling its chips to China amid the Sino-U.S. battle over trade and technology, said he was disappointed that China has launched an anti-trust investigation and discouraged domestic companies from buying Nvidia chips especially adapted for sales in China. Santa Clara, California-based Nvidia, the world’s most valuable company, is restricted from exporting to China its most advanced chips, which are in demand for developing artificial intelligence (AI). Last week, Chinese regulators accused the company of breaches of anti-trust law stemming from a 2020 acquisition of an Israeli technology company. Adding to the problems, the Financial Times reported that China’s internet regulator was banning domestic tech companies from buying an Nvidia chip model tailored for the local market. Huang said he did not have a reaction, but added, “I think that we could only be in service of a market if the country wants us to be”.
China was the world’s second-biggest AI computing market and Nvidia had contributed more than most companies, Huang said. “I’m disappointed with what I see, but they have larger agendas to work out, you know, between China and the United States, and I’m understanding of that, and we’re patient about it,” he said. Huang said the company would continue to be “supportive” of both governments as they “sort through these geopolitical policies”, adding there was “not very much anxiety there”. CEO Huang and President Trump talked about the latest developments at the state banquet for the U.S. President in London. Jenssen Huang was visiting London at the same time as Trump to unveil new investments, including a deal to supply tens of thousands of its processor chips for data centers that will be part of the UK arm of Stargate, a Trump-backed AI infrastructure project led by OpenAI, the South China Morning Post reports.
Meanwhile, Alibaba Group Holding’s semiconductor design unit T-Head has developed an artificial intelligence chip with capabilities that are on par with Nvidia’s H20 graphics processing unit (GPU). T-Head’s PPU card, with 96 gigabytes of high-bandwidth memory per unit, matched Nvidia’s H20 and surpassed Huawei Technologies’ Ascend 910B, and has a 400-watt power consumption, which is lower than what the H20 needed.
