The Trump administration has introduced tougher restrictions on advanced artificial intelligence chips, targeting a loophole that allowed Chinese companies to access American computing technology through overseas data centres.
Announced on November 16, 2026, the new rule aims to prevent Chinese firms from bypassing export controls by renting powerful AI computing infrastructure outside China's borders.
Although Washington restricts direct exports of advanced Nvidia processors to China, companies have continued accessing similar computing capabilities through data centres in Southeast Asia.
The restrictions target overseas facilities supplying substantial computing power to entities headquartered in countries including China, Russia and Iran, citing national security concerns.
Under the revised framework, advanced chip exports involving Indonesia, Malaysia, Singapore, Thailand and Brazil will face additional licensing requirements because authorities consider these destinations vulnerable to diversion.
Data centre operators must strengthen customer verification and due diligence, while affected businesses will have approximately 90 days before the new restrictions take effect.
The announcement follows controversy surrounding Singapore-based Megaspeed, which faces allegations involving restricted Nvidia chips and possible diversion to China. The company has denied wrongdoing.
The controversy also exposed disagreements within Washington. Treasury Secretary Scott Bessent supported stronger restrictions, while Commerce Secretary Howard Lutnick reportedly favoured a less burdensome regulatory approach.
The timing is significant, coming ahead of President Donald Trump's planned meeting with Chinese President Xi Jinping, where technology competition and trade tensions could feature prominently.
For Nvidia and global cloud providers, tighter controls could increase compliance costs, complicate international operations and restrict access to commercially important customers.
However, excessive restrictions could accelerate China's investments in domestic semiconductors, alternative AI infrastructure and technological self-reliance, potentially weakening America's long-term commercial influence.
The larger implication extends beyond semiconductor exports. Washington is increasingly treating remote computing access as strategically sensitive as physical chip ownership. Effective enforcement will require transparent licensing, reliable customer verification and international cooperation. The emerging AI rivalry is therefore shifting from controlling hardware shipments to governing who can access computing power worldwide.















