Google Advertisement
WHASINGTON – The United States has moved to sharply restrict the flow of advanced artificial intelligence (AI) chips to China, issuing new export rules that close loopholes exploited by Chinese firms through overseas subsidiaries.
The decision underscores Washington’s determination to slow Beijing’s progress in AI and high-performance computing, even at the expense of American semiconductor giants.
Closing the Loopholes
The Department of Commerce announced that licensing requirements will now apply to Chinese headquartered companies regardless of where their subsidiaries operate.
This change targets firms that had been sourcing restricted chips through Southeast Asian hubs, particularly Malaysia, bypassing earlier bans.
The chips in question include Nvidia’s Blackwell and Rubin processors and AMD’s MI350X accelerators, both essential for training large-scale AI models.
Analysts estimate that hundreds of thousands of units may have reached Chinese linked entities since mid-2025, despite existing restrictions.
The move represents a reversal from the Trump administration’s lax enforcement of the AI Diffusion Rule in 2025, when loopholes allowed shipments to continue largely unchecked.
By tightening controls, Washington aims to prevent China from acquiring the computational power needed to advance frontier AI systems.
China, however, has demonstrated resilience. Domestic firms have begun developing alternative AI architectures that rely on fewer chips, while state-backed initiatives continue to push for semiconductor self-sufficiency.
For U.S. chipmakers, the restrictions carry significant financial consequences. Nvidia and AMD, whose products dominate the AI accelerator market, face billions in potential lost sales.
Southeast Asian distributors may also see disruptions as Washington scrutinizes their role in global supply chains.
While the restrictions aim to safeguard U.S. national security, they also carry risks. If China succeeds in building competitive domestic chips, Washington’s strategy could inadvertently accelerate Beijing’s independence.
Moreover, existing data centers in China that already house restricted chips are not required to remove them, leaving gaps in enforcement.
The new rules mark a decisive step in the tech rivalry between Washington and Beijing, signaling that AI has become the latest battleground in global economic competition.
For the semiconductor industry, the challenge will be balancing compliance with U.S. policy against the loss of one of its largest markets.
