OpenAI CEO Sam Altman has said in an interview that the U.S. could be underestimating China’s progress and capability in artificial intelligence. CNBC quoted Altman saying that export controls alone likely aren’t a reliable solution and that he’s worried about America’s biggest rival since the Cold War.
“There’s inference capacity, where China probably can build faster. There’s research, there’s product; a lot of layers to the whole thing,” Sam told CNBC. “I don’t think it’ll be as simple as: Is the U.S. or China ahead?”
Altman is the second CEO of a major tech company that has questioned the effectiveness of export controls. Nvidia CEO Jensen Huang has said time and again that this policy was a failure, claiming that stopping China from getting AI chips like the Nvidia H20 and AMD MI308 is stifling the U.S.’s economic and technological leadership. It’s argued that despite the numerous bans, China was still able to develop advanced AI LLMs, like DeepSeek.
“My instinct is that doesn’t work,” Altman added. “You can export-control one thing, but maybe not the right thing… maybe people build fabs or find other workarounds.”
Even if Washington fully banned the export of all AI chips to Beijing, Chinese companies would still be able to get their hands on them through the black market. Just last quarter, it’s alleged that at least a billion dollars’ worth of banned Nvidia GPUs have entered China, with some smugglers even advertising that they can get their hands on the B300 even before it has officially launched.
Despite this, Beijing is encouraging its companies to move away from Nvidia. Chinese state media branded the H20 as unsafe and outdated, instead urging companies to rely on domestic hardware. This might be a difficult task, though, especially as Nvidia’s product stack is what gives it an advantage over its competitors. And even though its homegrown chips aren’t as powerful as what Nvidia offers, it can overcome this limitation through sheer brute force. This won’t be a problem for the country, too, as it has ample electricity production to cover the power demand of its burgeoning AI industry — something that the U.S. is having trouble keeping up with.
Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.
