Chinese tech companies train AI models in foreign data centers

"In doing so, they are attempting to maintain access to more powerful Nvidia chips, which are now virtually unavailable in China itself due to tighter US export controls. This is according to the Financial Times, based on sources with direct knowledge of the situation. According to the report, since April, when the US further restricted sales of Nvidia's H20 chip, there has been a steady increase in AI model training at offshore locations."

"The company managed to build up a large stock of Nvidia chips before the US export bans and can therefore still use domestic training facilities. In addition, DeepSeek is working with Chinese chip manufacturers led by Huawei to optimize and develop a new generation of domestic AI chips. This development underscores how geopolitical tensions surrounding advanced chips are reshaping the global AI infrastructure."

Chinese tech companies increasingly move AI model training to data centers outside China to access more powerful Nvidia chips restricted by US export controls since April's H20 restrictions. Southeast Asia has become a key region where companies lease compute from non-Chinese-managed data centers; Alibaba and ByteDance use such arrangements to deploy chips physically outside China. DeepSeek is an exception with pre-ban stockpiles and continues domestic training while working with Huawei-led domestic chip makers. Geopolitical tensions over advanced chips are reshaping global AI infrastructure. Demand for LLM training capacity is outpacing domestic supply as multiple companies build larger models concurrently.

Read at Techzine Global