US President Donald Trump recently cancelled a planned executive order on Artificial Intelligence (AI), citing concerns that it might hinder America's competitive advantage over China. This order, which had faced multiple delays, was effectively halted due to direct intervention from prominent industry leaders, including Elon Musk of xAI, Meta CEO Mark Zuckerberg, and venture capitalist David Sacks, a former AI and cryptocurrency advisor to Trump.

The now-scrapped executive order was not designed as a sweeping regulatory framework. Instead, it proposed a voluntary mechanism allowing AI developers to engage with federal agencies and submit advanced models for security review up to 90 days before public release. It intentionally avoided establishing a licensing regime or mandatory hold periods. Despite its voluntary nature, this level of oversight was seemingly deemed excessive by the influential industry figures who lobbied the White House. Trump himself stated he postponed it because he "didn’t like certain aspects of it" and worried it "could have been a blocker," a surprising stance from a president who generally prioritizes AI for jobs and national security.

This episode has further underscored a significant regulatory vacuum in the United States. While some governance architecture exists through piecemeal executive orders, agency guidance, and voluntary agreements – such as the federal Centre for AI Standards and Innovation's evaluation agreements with major tech companies – comprehensive AI legislation remains absent. The Trump administration had previously urged Congress to establish a single national standard for AI, preempting state-level laws, but no action has been taken.

The contrast with China's approach to AI governance is stark. Beijing's State Council outlined plans in May to accelerate comprehensive AI legislation by 2026, and the National People’s Congress has listed AI legislation for review for the third consecutive year. China has also implemented new rules requiring AI companies to establish internal ethics review committees. While Washington postpones policy, Beijing actively writes rules, highlighting a widening gap in regulatory foresight.

Thursday’s events clarified that in the current US administration, the effective veto on AI regulation rests with a small, influential group of industry principals with direct access to the president. Elon Musk, whose xAI competes directly with other major AI firms, has a vested interest in maintaining an open regulatory field. Similarly, Mark Zuckerberg's Meta champions open-source AI development. Even after formally leaving his advisory role, David Sacks evidently retains significant influence over executive action. This situation is further complicated by reports that OpenAI is simultaneously securing White House backing for state-level AI regulations, a move that appears inconsistent with the administration's broader push against state-imposed AI laws, suggesting a deeper problem in policy coherence.

Trump's rationale for cancelling the order – protecting the US lead over China – is a consistent theme in his administration's AI policy decisions. China, in turn, closely observes Washington's internal debates. Following a recent Trump-Xi summit, an agreement was reached to launch an intergovernmental dialogue on AI. Beijing will have noted that Washington’s internal discussion on even voluntary AI oversight was ultimately resolved not by policymakers, but by the very companies poised to benefit most from a lack of guardrails. Experts like Lizzi C. Lee from the Asia Society Policy Institute's Centre for China Analysis emphasize that both the US and China grapple with the fundamental question of where the regulatory frontier for frontier AI should be, particularly as models become more autonomous and critical to cybersecurity. She posits that a more crucial race is on governance and safety, not merely who possesses the most advanced models, but who can effectively govern powerful AI without stifling innovation. Despite arguments that AI safety and regulation can coexist with innovation, this perspective was insufficient to prevent the executive order's cancellation, leaving the future of US AI governance uncertain.