GMAsia
    🇨🇳China·AI News·21 Sept 2026·via SCMP·Covered by 6 sources

    As the US weighs restrictions on cloud computing, how will China’s AI sector adapt?

    Chinese AI developers cannot legally acquire the most powerful AI chips. They have maintained pace with global rivals by training models across borders through cloud computing. Firms establish proxy entities in foreign jurisdictions to lease server space in data centers. These centers are located in Southeast Asia, Japan, and the Middle East. Contracts signed by non-Chinese entities allow training on Nvidia processors outside US authority.

    Nexa's Summary

    Washington's potential cloud network restrictions would create China's most severe AI choke point since the chip bans. Chinese firms have circumvented US restrictions for years. They use proxy entities to lease server space in data centers across Southeast Asia, Japan, and the Middle East. This allows access to high-end Nvidia processors for model training.

    The test for Southeast Asian cloud providers will be enforcement. Their data centers have been critical for Chinese firms. New US rules could force these providers to choose between US and Chinese business. This would impact regional data center investments and cloud infrastructure growth. Singapore and Malaysia could see reduced demand from Chinese clients.

    The thing to watch is the specific language of any new US cloud computing restrictions. If the rules target indirect access, Chinese firms will face a hard stop. This would accelerate domestic chip development and alternative cloud solutions within China. It would also increase pressure on regional cloud providers to comply.

    Original reporting by SCMPWe don't republish, read the full story →

    Related reading

    6 stories