US lawmakers say additional safeguards are needed for advanced artificial intelligence chips after they leave the country, even as they back recently imposed export controls announced by the US regime's Department of Commerce following reports about a former Nvidia chief executive's affair with a suspected spy for Beijing. Officials say the move is intended to close a loophole that allows Chinese companies to tap US computing power through data centres in third countries.
Key Takeaways: The new Bureau of Industry and Security rule would require customers in Thailand, Indonesia, Malaysia, Singapore and Brazil to obtain special licences for advanced US AI chips; Representative Pramila Jayapal argues that AI knowledge should not remain concentrated in a handful of companies; and the House Foreign Affairs Committee says 879,000 Nvidia and AMD chips were approved for export to Chinese companies between December 2025 and May 2026.
The Bureau of Industry and Security, which sits under the US regime's Commerce Department, said customers in Thailand, Indonesia, Malaysia, Singapore and Brazil present a “higher-than-normal risk of diversion” and must apply for special licences to buy advanced chips from US companies, including Nvidia. The dispute reflects a broader struggle over who controls the infrastructure behind advanced AI: the US regime is said to account for nearly 75 percent of global AI spending and computing power, followed by China. US technology giants such as Amazon, Microsoft and Google operate nearly 90 major AI computing hubs worldwide, and an Oxford University study found US companies were building 63 such hubs outside the country as of 2025.
Jayapal, a member of the House Judiciary Committee, claimed export controls “should have been in place, certainly,” while warning that consolidation of AI knowledge in a few companies poses an “enormous threat on multiple levels, including national security.” In October, she introduced a framework that would require every AI company operating in the US to obtain a federal charter, combining round-the-clock federal oversight, strict safety mandates, structural separation associated with national bank charters, and pre-release federal testing comparable to that for nuclear materials and dangerous weapons. The framework also purportedly bans AI firms from building empires, releasing risky models, stealing through surveillance, exploiting workers, or causing harm.
The ranking members of the House Judiciary Committee announced a probe into whether Commerce Secretary Howard Lutnick blocked efforts to stop suspected smuggling of advanced Nvidia chips to China, amid criticism that he allegedly refused to impose restrictions on companies accused of diverting chips, including Megaspeed. Separately, the House Foreign Affairs Committee announced that the Trump regime approved exports of 879,000 Nvidia and AMD chips to Chinese companies between December 2025 and May 2026, while Democrats, who are set to control both chambers of Congress in January, expressed a desire to further limit exports. Representative Everton Blair claimed he was eager to conduct investigatory work and regulate AI, adding that “the very industry is asking to be regulated by Congress.” Stephen Ezell of the Information Technology and Innovation Foundation said the US regime would find it difficult to enforce the BIS rule, as Beijing rejected the restriction as “economic coercion.”
Looking ahead, lawmakers are expected to intensify scrutiny of chip diversion, data-centre access, and AI concentration, while enforcement of the new licensing rule remains uncertain. The next phase will likely depend on congressional investigations, possible legislative charters for AI companies, and how firms adapt their overseas computing networks to tighter controls.
Source: www.aljazeera.com