The AI arms race gets personal

Artificial Intelligence and tariffs have become the pivotal battlegrounds in the escalating rivalry between the United States and China, and the U.S. Department of Commerce (DOC) redrew a significant line of engagement this week by formally rescinding the Biden administration’s “Artificial Intelligence Diffusion Rule”, mere days before its May 15th effective date. The rescission signals a strategic pivot in how America intends to wage this high-stakes tech war. This manoeuvre, set against the persistent drumbeat of the US-China tariff disputes, casts a long shadow of uncertainty and opportunity over tech companies and global supply chains desperately trying to navigate the fight for AI supremacy.

The now-axed “Artificial Intelligence Diffusion Rule,” introduced under former President Joe Biden in January, was a complex attempt to manage the global proliferation of advanced AI chips and, by extension, AI capabilities. The stated aim, per a White House press release, was to “provide clarity to allied and partner nations” and streamline chip licensing. Critically, however, it established new, widespread restrictions on exporting U.S.-made AI chips.

Biden’s framework was essentially a global triage system. Closest allies like Japan and South Korea were largely spared. Adversaries, notably China and Russia, already under tight restrictions for advanced AI chips, faced even tougher limitations, especially concerning sophisticated “closed” AI models. The most significant impact was on a vast third tier of nations like Israel and Mexico. These nations would have faced caps on AI chip imports (reportedly around 50,000 GPUs each, with avenues for increase), a measure explicitly designed to prevent China and Russia from sourcing critical technology through them. Industry giants like Nvidia voiced strong opposition, calling the proposed rules “unprecedented and misguided” and warning they would “derail” global innovation. This regulatory structure was intended to build on previous guidance from October 2022 and October 2023.

The Trump administration has decisively swept this tiered system aside. The DOC’s announcement of the revocation included plans for a “replacement rule in the future, likely focusing on direct negotiations with countries as opposed to blanket restrictions,” according to a Bloomberg report. This reflects a different philosophy of engagement as the U.S. Secretary of Commerce for Industry and Security, Jeffrey Kessler, declared, “The Trump Administration will pursue a bold, inclusive strategy to American AI technology with trusted foreign countries around the world, while keeping the technology out of the hands of our adversaries.” He didn’t mince words, labelling the Biden approach an “ill-conceived and counterproductive” imposition.

Instead of Biden’s broad rulebook, the DOC issued immediate, pointed guidance: a stark reminder that using Huawei’s Ascend AI chips anywhere globally violates U.S. export rules; a caution about the dire consequences of allowing U.S. AI chips to be used in training AI models within China; and recommendations for fortifying chip supply chains against diversion tactics. This indicates a preference for direct, forceful interventions and targeting specific entities and practices rather than implementing a complex, universal regulatory scheme.

New battle tactics

This shift in AI export control is more a change in battlefield tactics than a retreat from the strategic objectives. Both administrations share the core goal: cementing U.S. dominance in AI and critically, preventing China from leveraging advanced technology to surpass American military and economic power. Biden’s strategy aimed to manage AI diffusion through a systematic, if cumbersome, international framework. The Trump administration seems to be gearing up for a more flexible, arguably more aggressive and transactional engagement, relying on bespoke agreements and sharp, punitive actions against those, like Huawei, deemed central to the adversary’s capabilities. The explicit condemnation of using US AI chips to train models in China underscores this direct confrontational stance.

The rhetoric of a “bold, inclusive strategy with trusted foreign countries” hints at a different approach to alliance-building in this tech conflict, but the unwavering focus on keeping advanced tech from “adversaries” shows the underlying contest with China remains paramount.

The economic front in the AI war

This recalibration of AI export controls cannot be viewed in isolation. It is intrinsically linked to the ongoing tariff war, another major front in the US-China battle. Significant tariffs, initiated under the previous Trump administration and largely continued by Biden, were already a defining feature of the economic landscape. U.S. Tariffs on Chinese goods, and retaliatory Chinese tariffs on U.S. products, have directly impacted the tech sector by increasing the cost of components, servers, and finished goods essential for AI development and infrastructure.

In the current term, Mr. Trump has further solidified the hard line on China, deploying tariffs as a coercive tool in trade and technology negotiations. Apart from tariffs, the current administration has increased compliance on chipmakers. Most recently, Nvidia was required to go through a licensing route to ship its H20 GPUs to China. The AI chip was a dumbed-down version of the powerful H100 and H200 GPUs that were tailor-made for China.

This type of policy shift creates a dual pressure point for the tech industry: restricted access to and movement of AI technology, compounded by the increased financial burden of tariffs. Tech companies are, in essence, navigating a minefield where both the AI know-how and the hardware it runs on are subject to intense geopolitical manoeuvring and economic warfare.

Navigating the new battleground

For tech companies and their global supply chains, this integrated battlefront of AI controls and tariffs signals continued, if not amplified, turbulence. In the near term adapting to ambiguity will be important as geopolitics move from a defined set of rules to a future that is based on direct negotiations. While some may find new avenues through bespoke agreements, the lack of a clear, universal framework makes strategic planning challenging.

Secondly, countries previously in Biden’s “third tier” might perceive an opening, but they will also likely face intensified diplomatic and economic pressure as the U.S. prosecutes its “trusted partner” strategy in the shadow of the overarching US-China rivalry.

Looking to the long term, the intertwined forces of AI export controls and tariffs will likely compel companies to de-risk and diversify their supply chains. “Friend-shoring” and regionalisation efforts will intensify as businesses seek to insulate themselves from the fallout of this battle. This could lead to increasingly separate tech ecosystems and a more splintered technological world.

As these ecosystems evolve, China will only accelerate its already massive investment in domestic AI capabilities—from semiconductor design and manufacturing to foundational AI models. The U.S. trade policies are acting as a catalyst for China’s self-reliance drive even as a U.S. foreign policy increasingly built on direct deals for technology access will likely trigger a more dynamic and competitive scramble among nations to align themselves strategically, caught between the technological gravity of Washington and Beijing.

This shift in U.S. AI export policy is not a standalone event but a tactical adjustment on this sprawling battlefield. While the objective of ensuring American AI leadership and containing China’s technological advance remains unchanged, for the global tech industry, this means bracing for continued geopolitical turbulence where innovation, market access, and supply chain integrity are all caught in the crosshairs of this superpower contest.