Recent US decisions on exports of advanced semiconductors have left observers struggling to answer basic questions: how, exactly, does Washington now see China? Is Beijing a competitor to be constrained, a rival to be managed, or a market that remains — uncomfortably — indispensable? 

A series of recent moves fail to offer clear answers. In January 2026, the Trump administration shifted toward a case-by-case licensing approach for certain advanced AI chips to China, allowing some previously restricted exports. Weeks later, it imposed a 25% tariff on some advanced computing chips, invoking national security rules. Around the same time, the White House granted Taiwan Semiconductor Manufacturing Company an license to ship American chipmaking tools to its fabs in China.  

The recent Supreme Court ruling limiting the president’s ability to impose certain tariffs adds to the confusion. Without a consistent tariff tool to wield, export controls could reemerge as one of the executive branch’s most important sources of leverage. This may already be happening, as firms react to reported plans for the US government to act as the “gatekeeper for the AI industry”. 

For much of the past decade, Washington framed export controls as a way to deny China access to the most advanced technologies, buying time to preserve American and allied advantage. The approach relied on prohibitions and the assumption that chokepoints could be enforced without severe blowback.  

Experience has eroded confidence in the strategy. Semiconductor supply chains remain interconnected, capital-intensive, and geographically distributed. Attempts to sever them have proven costly and destabilizing. The economic reality is that decoupling (full, clean, and enforceable) is impossible.  

The TSMC license is emblematic. Allowing the world’s most important contract chipmaker to continue shipping American equipment to its China fabs represents a concession to reality. If TMSC was blocked, US toolmakers would suffer. Industries dependent on Chinese chips, from autos to vacuum cleaner manufacturers, would struggle. During the COVID-19 pandemic, chip shortages caused massive economic pain. 

A similar logic underpins the revised AI chip licensing regime. Export controls have not disappeared; they have become conditional and administratively dense. Companies must certify performance thresholds, disclose end users, submit to third-party testing, and demonstrate that foreign sales will not divert supply from US customers. 

Controls once drew bright lines between permitted and prohibited technologies, between civilian and military use, and between allies and adversaries. Today’s framework blurs those lines, making them hard to explain, and easy to contest. If tariffs can no longer function as a sweeping fallback tool, the burden on this more intricate regulatory architecture will increase. 

Congress has noticed — and is unhappy. Skepticism toward easing export controls on China remains bipartisan and appears to be hardening. Lawmakers argue that even commercially constrained chips can be repurposed for military or intelligence use, particularly in a system as opaque as China’s.  

Pending legislation would expand congressional oversight of licenses, extend controls to cloud services, prioritize domestic buyers, and limit executive discretion. The tension between the White House and Congress looks to become a defining feature of US tech policy. The Supreme Court’s intervention strengthens Congress’s hand in this tug of war, affirming the powers of legislative authorization.  

This White House acts on the assumption that China’s technological rise cannot be stopped, only shaped. Its priority is reducing reliance on adversaries, reshoring critical capacity, and insulating domestic supply chains from external shocks. Tariffs, not export bans, had become a favored instrument. Yet, with sweeping tariffs now legally constrained, this preference may prove harder to sustain, forcing a reevaluation of export controls. 

Congress, by contrast, continues to treat denial as both feasible and necessary. That divergence creates uncertainty for firms, allies, and markets. It weakens the credibility of the overall strategy. Judicial scrutiny adds a third constraint, limiting executive power. 

China will continue to advance regardless of US preferences. Its trajectory is constrained by inefficiencies, demographics, and external pressure, not on American permission. Export controls may slow certain progress but cannot reverse it, as China’s response to recent measures — halting or discouraging purchases of certain US chips — demonstrates.  

US policy requires a clear articulation of goals. Is the aim to slow China, to compete alongside it, or to ensure US resilience regardless of outcomes?  Current policy gestures toward both objectives without fully committing to either.  

Elly Rostoum is a Senior Resident Fellow with the Center for European Policy Analysis (CEPA).  

Bandwidth is CEPA’s online journal dedicated to advancing transatlantic cooperation on tech policy. All opinions expressed on Bandwidth are those of the author alone and may not represent those of the institutions they represent or the Center for European Policy Analysis. CEPA maintains a strict intellectual independence policy across all its projects and publications.

The post US Chips and China: A Confusing Cocktail  appeared first on CEPA.