Export Controls Should Advance U.S. Semiconductor Leadership

The United States needs a clear, consistent, and bipartisan strategy for semiconductor export controls.

During the last presidency, the Biden administration imposed increasingly stringent export controls on advanced chips, semiconductor manufacturing equipment, and other components, such as high-bandwidth memory. The restrictions eventually covered not only countries on which the United States has an arms embargo, such as China and North Korea, but also other jurisdictions. The final set of restrictions, known as the AI Diffusion Rule, would have imposed a complex set of restrictions and licensing requirements on AI chip exports to all but 18 of America’s closest allies.

After President Trump took office, his administration rescinded the most onerous export controls, opening the door to greater chip exports by U.S. companies. But it has not yet developed a clear and consistent policy. For example, it has reversed its chip export policy multiple times. The Trump administration banned the sale of Nvidia’s H20 to China in April 2025, reversed that policy in July, and then only started issuing licenses in August after the company agreed to give the federal government 15 percent of its revenue from selling those chips. Likewise, in May, the administration banned the export of chip design software without a license, but then it lifted those restrictions in July.

It is unclear whether these policies aim to serve as a bargaining chip in trade negotiations with China, generate revenue for the federal government, or—as they should—maintain America’s leadership in semiconductors while delaying China’s progress and ensuring its dependence on the U.S. semiconductor ecosystem for as long as possible. Moreover, because the Biden and Trump administrations have pursued these policies with little bipartisan buy-in from Congress, they remain subject to the whims of each presidency. The result is that the U.S. semiconductor industry has no predictable policy to plan against.

The central goal of a U.S. semiconductor export control strategy should be to maintain American leadership and prevent China from closing the gap. The Chinese government is pursuing “self-reliance, national resilience, and ultimately industrial dominance” for its semiconductor industry—meaning it eventually wants to supplant U.S. chipmakers in China and in the global market.

While it’s unrealistic to expect U.S. export restrictions to keep China dependent on U.S. chips forever, the United States can slow China’s progress. In this context, the recent decision to permit more advanced chips in the Chinese market supports these goals. Allowing U.S. chipmakers to compete in the Chinese market benefits them by providing more revenue to invest in R&D to maintain their global leadership, while denying Chinese chipmakers the opportunity to build local scale and compete globally.

China’s complete decoupling from U.S. semiconductor technologies would hurt American techno-economic power. The semiconductor industry relies heavily on scale, as larger markets help offset high initial investments and lower manufacturing costs per unit. Consequently, market leaders are better positioned to reinvest in research and development, advance new generations of semiconductor technology, and sustain their leadership.

Furthermore, advanced chips are essential for the development and deployment of AI, including frontier AI models. Losing leadership in the semiconductor sector would diminish America's competitiveness across all knowledge-intensive industries, from robotics to biotech. And semiconductors serve dual purposes, with both civilian and military uses; America’s leadership in this field is a matter of national security.

The U.S. government should pursue a strategic export control policy that helps the U.S. semiconductor industry while slowing the rise of China’s. This policy should entail reviving the “small yard, high fence” strategy, in which the United States restricts a narrow set of its most advanced semiconductor technologies to its geopolitical rivals while allowing other beneficial economic activity. This policy aligns with much of what the Trump administration is doing now and served as the basis for the Biden administration’s export restrictions, providing a solid basis for bipartisan agreement.

The biggest challenge is deciding how small a yard and how high a fence are necessary, but policymakers should focus on pragmatism over perfection. The Biden administration’s mistake in pursuing this policy was that it kept expanding which chips to exclude from foreign markets and adding new controls to prevent potential evasion—the yard kept getting bigger, and the fence kept getting taller. The Trump administration’s mistake is that it hasn’t decided what size yard or fence it even wants, and it wants a cut of the revenue for anything that goes over the fence.

A simpler, more focused policy would serve U.S. commercial and national security interests better. First, ban U.S. companies from selling any chips to designated foreign organizations that pose a threat to U.S. national security, such as those on the Commerce Department’s Entity List or the Defense Department’s Section 1260H list of Chinese military companies. If these lists are not comprehensive enough to cover organizations supporting the Chinese military, then have Commerce create another one for this purpose. This policy should ensure that U.S. companies are not directly providing U.S. technology to foreign adversaries or their enablers.

Second, impose a fixed waiting period, such as 6 months or 12 months, before U.S. chipmakers accept orders on newly released chips from Chinese companies. This policy would ensure that U.S. companies have the opportunity to be first in line to obtain the most advanced chips, while also providing a clear timeline for U.S. chipmakers to manage sales and exports to a critical foreign market. This policy would also prevent them from having to spend time and money designing special chips just for China’s market.

Third, reject efforts to require U.S. semiconductor firms to add mandatory technical controls to chips, such as geolocalization requirements. Any government-imposed mandates will not only turn off potential customers—including among friendly nations that worry that such controls could be used against them if the United States becomes an unreliable ally—but they will also serve as fodder for the Chinese government to allege these chips have backdoors that present security threats. Instead, the U.S. government can require that firms in countries of concern, as a condition of receiving an export license, agree to activate voluntary location verification technology to provide assurance that they have not diverted their chips.

Finally, cease requests that U.S. chipmakers pay a percentage of export revenue to the U.S. government. If the U.S. government wants to raise revenue, it should do so through traditional tax policy, not ad-hoc deals with private sector companies that need government approval.

The United States should take the opportunity to craft a strategy for chip export controls that sustains American leadership in semiconductors. By working with Congress, the Administration can establish a clear, consistent, and bipartisan chip export control strategy that will support this critical sector and allow the United States to remain ahead of China.