Three Questions to Emily Benson: Inside the Conflicting Signals of U.S. Export Control Strategy

The Trump administration has sent mixed signals-both tightening and loosening-on semiconductor export controls toward China. How do you assess the overall direction of US policy? 

It can be difficult right now to read the tea leaves in Washington. The news cycle never stops, and there is this sense that major policy announcements are being made almost every day. But yes, it is clear that the Trump administration has been shifting course on chip export controls. For example, they recently lifted the ban on Nvidia AI chip exports to China and reached a deal on TikTok, allowing a U.S. version of the app to continue to exist in the American market. To China hawks, these moves feel like abrupt reversals of previous national security commitments. But it is also important to understand that this shift fits more into a kind of post-Cold War logic, where increasing exports was often seen as commercially beneficial, especially if it created long-term dependencies on US technology.

What explains this change in course? Let’s start with a very domestic factor: political priorities. Even just a year ago, domestic politics were a major driver of how the US approached China. The spy balloon incident, for instance, stirred up public sentiment and pushed policymakers to take a harder line. Human rights concerns were also a major theme under the Biden administration. But right now, the domestic political mood is much more inward-looking: Americans are focused on other issues. That reduces the pressure on the White House to take aggressive stances on something as complex as export controls.

Let me add that beyond political priorities, there are other practical factors at play. Bureaucratic capacity is a real challenge: the Bureau of Industry and Security (BIS) is currently operating with significantly fewer staff than last year, with just 218 people handling around 30,000 license applications. This does not suggest much room for a major policy shift. Moreover, export controls have become far more political. Decisions that were traditionally made by technical staff through analytical processes are now being handled at the political level. 

This is particularly vital given the mid-October tit-for-tat escalation that mirrors the U.S.-China exchange in April. That broader pressure is placing constraints on Washington’s ability to maintain a consistent export control policy at the moment. Maintaining some semblance of the status quo or pursuing de-escalation will require both sides to implicitly agree not to levy new statecraft measures before an arrangement is reached.

There are still a few factors that could push the U.S. back into a tougher posture: another Volt Typhoon-style cyberattack from China on the U.S.; an intensification of export restrictions by China, that would further disrupt U.S. production capacity; the development of groundbreaking new technologies by China, that would shift the military balance in China’s favor; or beyond export controls, an expanded set of restrictions by China against U.S. companies. These concrete steps that China could take could force the U.S. to respond with much greater resources. If Washington reaches that level of anxiety, we could find ourselves back where we were a year ago. But I do not see that happening in the immediate future.

How do you see the administration handling semiconductor technology transfers with allies, particularly in the context of trade negotiations but also more broadly in bilateral relations?

For most of the 20th century, there was actually a fairly cohesive international consensus around which technologies needed to be controlled in the interest of international peace and security. That is what gave rise to multilateral frameworks likethe Wassenaar Arrangement: a broadly accepted set of principles focused on limiting the proliferation of tools, inputs, and technologies that could be used for military purposes and lead to the proliferation of weapons. But that is clearly no longer the case today. If anything, what we are seeing now is a growing absence of consensus on what is worth controlling, which is sparking real debates about whether the existing multilateral regimes are still suited for today’s environment.

We are also seeing a shift in U.S. strategy in this space. In its AI Action Plan, for example, the U.S. signals a clear desire to move away from long, cumbersome multilateral negotiations. Instead, Washington is leaning more toward bilateral engagement, prioritizing one-on-one talks where it can extract greater concessions from partners, in a more transactional approach.

The strategic focus is gradually shifting away from traditional export controls and toward building a kind of common external tariff wall around North America, particularly to protect this area from Chinese imports. And Washington has gone a step further: it has asked both Mexico and Canada to adopt review mechanisms similar to CFIUS, not just based on concerns around overcapacity, but grounded in national security considerations, with a much broader interpretation of what qualifies as a national security risk.

What all this means in practice for semiconductor technology transfers is that we should expect more extraterritorial rule-making by Washington, often with immediate entry into force. It is not uncommon now for a new regulation to be announced one day and enforced the next, with less regard for the implementation capacity of foreign partners.

What advice would you offer Europe on key issues to address, including how Europe should approach transatlantic relations in that regard?

In the short term, a U.S.-China economic détente actually seems quite likely. Similar to April, both Beijing and Washington appeared to backtrack on escalatory rhetoric during a tense exchange of trade measures in mid-October. But the question is how long any stability will really last. That’s why, for the EU, the existential question going forward is: how can Europe avoid becoming collateral damage if U.S.-China talks eventually fall apart? My first piece of advice would be for Europe to recognize its weight in the semiconductor value chain.

First, a tremendous amount of EU companies hold real leverage over key parts of the supply chain. Without them, neither the U.S. nor China could produce the advanced technologies they need. The European Commission has not yet fully succeeded in advertising the extent of Europe’s authority in these supply chains. We need a more nuanced debate around how far the EU can and should go in asserting control over these critical supply chains. There is real value in having that discussion. At the same time, Europe also needs to better prepare for the possibility of Chinese retaliation and develop a clear response strategy. Many American companies were caught off guard by Chinese countermeasures. Europe must take a stronger, deterrence-focused stance to protect European companies. At the end of the day, this should not be solely about supply chain vulnerabilities: the EU does own critical chokepoints it can showcase and use to influence global tech trade dynamics. 

Of course, there are also technical areas where Europe could take a closer look at how it manages export controls: consider whether or not a country-agnostic approach is the right one to take, ensure that major export control decisions targeting China are really grounded in data-driven, fact-checked analysis, or raise the question of whether an ICTS-style regime would be relevant to assess imports posing national security risks, as a complement to export control mechanisms.

What seems both reasonable and healthy is for Europe to push for formal agreements with American firms (mutual assurance doctrines, for instance) and to define a concrete set of potential retaliation measures to ensure those agreements are respected. More broadly, Europe should come to the table with clear, well-defined demands that are firmly tied to its long-term tech sovereignty goals.

Copyright image : CHIP SOMODEVILLA / GETTY IMAGES NORTH AMERICA / Getty Images via AFP
On July 23 in Washington, Donald Trump signs the “Winning the AI Race” executive order.