When the Chinese start-up DeepSeek released its latest artificial intelligence model last month, it edged Beijing closer to a future that it has spent years trying to build.<\/p>\n
In a small but meaningful break from American technology, DeepSeek said for the first time that its new model<\/a> had been optimized to run on chips made by the Chinese tech giant Huawei. This was a milestone in China\u2019s long-running effort<\/a> to develop advanced technologies at home and reduce its reliance on Western innovation.<\/p>\n While most of the world\u2019s leading A.I. systems still rely on semiconductors from the U.S. chip-making giant Nvidia, Chinese A.I. firms are increasingly turning to homegrown alternatives.<\/p>\n The timing of DeepSeek\u2019s announcement \u2014 before this week\u2019s scheduled summit between President Trump and Xi Jinping, China\u2019s leader \u2014 gives Beijing fresh confidence entering trade talks that U.S. export controls on Nvidia chips have not derailed China\u2019s A.I. development.<\/p>\n Any meaningful shift by China away from American A.I. technology could limit the impact of U.S. export controls and deprive Washington of a critical source of leverage over Beijing. That prospect gained urgency since DeepSeek\u2019s A.I. technology rattled the U.S. tech industry and turned the company into a potent symbol of China\u2019s drive for technological self-sufficiency.<\/p>\n Before last year\u2019s meeting between the two leaders, Mr. Trump said he planned to discuss Nvidia\u2019s most powerful A.I. chips with Mr. Xi, fueling speculation that the United States might ease restrictions on the technology.<\/p>\n But after years of Washington\u2019s preventing Chinese companies<\/a> from buying certain advanced technology products, firms like DeepSeek and Moonshot AI are starting to design their A.I. systems around the constraints rather than waiting for them to disappear. That includes exploring how their models can run on a broader range of processors beyond Nvidia\u2019s.<\/p>\n \u201cU.S. export controls are not freezing China\u2019s A.I. development,\u201d said Wei Sun, a principal A.I. analyst at Counterpoint Research in Beijing. \u201cThey are forcing China to build an alternative stack.\u201d<\/p>\n DeepSeek has said its latest model can use Huawei chips for inference, the process that allows an A.I. system to respond more quickly and accurately to users. Inference generally requires less computing power than training, the demanding process of teaching a model how to function. DeepSeek still relied on Nvidia chips to train its system, according to two people in the semiconductor industry who were not authorized to comment publicly on the matter.<\/p>\n It was not immediately clear how DeepSeek gained access to those chips, though Chinese companies can still remotely use Nvidia chips housed in data centers outside China. DeepSeek did not respond to a request for comment.<\/p>\n Huawei has said it plans to release a chip for training this year. But it also said it would take another year after that before its products could match the performance of Nvidia\u2019s current offerings.<\/p>\n The growing split between Chinese and American A.I. infrastructure is a consequence that Jensen Huang, Nvidia\u2019s chief executive, has long warned would result from rigid export controls.<\/p>\n