China Races to Build AI Chip Supply Chain Amid U.S. Export Curbs

China is pushing to build its own artificial intelligence chip ecosystem after U.S. export controls blocked access to advanced semiconductors and chipmaking tools. These restrictions have cut China off from buying powerful AI chips made by companies like Nvidia, and from key manufacturing equipment made in the U.S., Netherlands, and Japan.

In response, China has invested billions of dollars to strengthen its local chip industry. Huawei has emerged as a leading player, with its Ascend 910B and 910C processors showing improvement, though still behind Nvidia’s latest models. Experts say Huawei has reduced the performance gap but remains about one generation behind.

To manufacture chips, Chinese companies rely on local foundries like SMIC. While SMIC can produce 7-nanometer chips, it lacks access to extreme ultraviolet (EUV) lithography machines needed to make the most advanced chips. These tools are made by ASML in the Netherlands and are restricted under U.S. export rules.

China has been trying to work around these limits by using older equipment, but production yields are low. Domestic companies like SiCarrier Technologies are exploring new methods, but experts say catching up could take years.

In memory chips, which are essential for AI computing, China is also lagging behind. South Korea’s SK Hynix and Samsung dominate the market, and have begun restricting sales to China under U.S. pressure. Chinese chipmaker CXMT is working to develop high-bandwidth memory, but large-scale production is still far off.

Despite progress, analysts say China’s AI chip industry still depends heavily on foreign technologies. The country’s efforts reflect a major shift toward self-reliance, but challenges remain in scaling and competing globally.