Despite stringent US export controls targeting Nvidia’s advanced AI chips, China’s technology sector continues to accelerate its artificial intelligence development through strategic overseas operations. Industry experts reveal that Chinese firms are effectively circumventing restrictions by accessing high-performance computing resources in Southeast Asian data centers.
The controversial export regulations, initially implemented by the Biden administration and subsequently modified under Trump, were designed to limit China’s access to cutting-edge AI hardware. However, according to Gary Wojtaszek, director at GDS Holdings and seasoned data center executive, these measures contain fundamental limitations. “The US chip export rules are pretty nebulous. In China, you can’t import chips, but you can export your data to train your model. Then you import it back,” Wojtaszek explained, comparing the phenomenon to water naturally finding its level.
This technological end-run occurs primarily through two channels: Chinese companies either purchase computing power directly from Southeast Asian data centers or install their own server infrastructure in facilities across Singapore, Malaysia, Indonesia, Thailand, and beyond. This allows them to train sophisticated AI models overseas before deploying the finished algorithms domestically.
The pattern has become so established that major Chinese tech giants including Alibaba and ByteDance now routinely conduct AI training operations in regional hubs like Singapore and Johor, Malaysia. In one documented case, Chinese engineers transported hard drives loaded with training data to Kuala Lumpur, where they rented approximately 300 Nvidia AI servers—a maneuver that technically violated no existing laws.
Complementing this overseas strategy, Beijing has simultaneously encouraged domestic substitution, urging technology firms to prioritize homegrown alternatives like Huawei’s AI processors. This dual-track approach has enabled China to narrow the technological gap with the United States, leveraging its advantages in energy production and massive data pools for AI training.
The financial impact on Nvidia has been unexpectedly positive, with the company reporting $215.9 billion in revenue for the fiscal year ending January 2026—a 65% year-on-year increase. Remarkably, Singapore emerged as Nvidia’s second-largest market, accounting for approximately one-fifth of total revenue, largely driven by Chinese demand.
However, new legislative developments threaten to close these loopholes. The recently passed Remote Access Security Act would expand US authority to restrict foreign access to sensitive technologies through cloud services and remote computing. If enacted, using US AI chips via overseas data centers could become subject to the same controls as physical exports.
Concurrently, Southeast Asia’s data center market is experiencing unprecedented growth, partially fueled by declining preference for Hong Kong as a computing hub. According to Doug Adams, CEO of NTT Global Data Centers, connectivity issues and restricted access to Western AI services have diminished Hong Kong’s appeal, with customers increasingly choosing alternatives across the region.