Top Chinese technology giants are increasingly shifting their artificial intelligence development outside the country to work around U.S. limits on advanced semiconductor access, according to a Thursday report from the Financial Times.
People familiar with the situation told the outlet that major firms, including Alibaba and ByteDance, are conducting the training of their newest large language models in overseas data facilities, especially those located in South-east Asia.
This move enables them to tap into Nvidia’s high-end chips—hardware they can no longer easily obtain within China due to escalating U.S. export restrictions.
Related story: Micron to Halt Server Chip Sales to Chinese Data Centers
Firms Move Operations as Measures Tighten
The U.S. has steadily expanded limitations on the sale of high-performance semiconductors and related technologies to China, citing national security concerns and the goal of slowing China’s progress in strategic sectors, including AI.
As the Financial Times noted, these rules have pushed some of China’s most prominent technology companies to relocate critical parts of their AI development pipelines to regions unaffected by Washington’s controls.
“Alibaba and ByteDance are among the tech groups training their latest large language models in data centres across South-east Asia,” the report said, citing two people with direct knowledge of the matter.
Offshore Training Becomes a Strategic Workaround
By shifting their most computationally intensive AI training tasks abroad, these firms are still able to rely on Nvidia’s powerful GPUs—widely regarded as essential for building competitive AI systems.
Although Nvidia has crafted downgraded chip versions that comply with U.S. regulations, Chinese companies remain motivated to access the more advanced models to keep pace with global innovation.
The outlet explained that this approach allows the companies to continue developing next-generation AI technologies despite newly introduced U.S. measures aimed at constraining China’s technological advancement.
U.S. officials have stated that the restrictions are intended to prevent advanced computing resources from aiding the Chinese military or other state-affiliated entities, while still permitting broader commercial technology flow.
However, the report suggests the policy may be unintentionally prompting Chinese tech leaders to expand their international footprints rather than slowing their progress.
Establishing AI training capacity abroad enables them to operate within legal boundaries while still leveraging top-tier hardware unavailable domestically under current rules.
This growing trend underscores how global—and geopolitically sensitive—AI development has become, as companies navigate regulatory barriers while racing to build increasingly capable AI models.
Read next: Dell Projects Stronger Growth as Demand for AI-powered Servers Surges