The Trump administration released a sweeping science and technology strategy this week that reveals a fundamental tension in how the US plans to win the AI race against China: it wants to attract the world's best talent while making it harder for foreign researchers to stay in the country. The National Security Science and Technology Strategy, released on Tuesday, explicitly names AI, semiconductors, quantum systems, and hypersonics as critical fields where US leadership is essential for geopolitical competition. Yet the same administration is simultaneously raising barriers to the international talent that fuels innovation in these exact areas.

Why Is the US Restricting Visa Access While Demanding AI Dominance?

The strategy document states that US science and technology leadership "is itself a national security objective," and it calls for "attracting and retaining top-tier global talent in critical national security S&T fields". This language signals urgency; the administration views technological superiority as inseparable from national security in an era of great power competition with China. However, this goal collides directly with recent policy moves that discourage overseas researchers from coming to or staying in the United States.

The administration has proposed a $100,000 fee on H-1B visa applications, a work visa category heavily used by tech companies to hire specialized talent in AI, software engineering, and semiconductor design. Simultaneously, the Department of Homeland Security has replaced duration-of-status visas with a four-year cap for F student visas, J exchange visitor visas, and I journalist visas, effective September 15. This means international students and researchers can no longer stay for the full length of their academic program or research assignment; instead, they face a fixed time limit and must apply for extensions to remain longer.

A broad coalition of labor and education organizations has already filed a lawsuit challenging the visa duration changes, arguing the rule is "arbitrary and capricious" and that the Department of Homeland Security failed to properly consider public comments as required by law. The plaintiffs include colleges, universities, teachers, international students, aerospace workers, and journalists.

What Specific Technologies Is the US Trying to Protect From China?

The National Security Science and Technology Strategy updates the critical and emerging technologies list used in export controls and federal research priorities. The document adds quantum cryptography to the list and removes virtual reality, signaling where the administration believes the real competitive threats lie. Data centers, notably, no longer count as a priority, with private companies pursuing massive buildouts in the US.

Beyond the technologies themselves, the strategy increases restrictions on transferring knowledge and expertise to China in several key areas. These restrictions now apply to:

• Artificial Intelligence: The administration is tightening controls on AI research, models, and training methods that could enhance Chinese capabilities in autonomous systems and decision-making algorithms.
• Quantum Information Systems: Quantum computing poses an existential threat to current encryption, making quantum cryptography and quantum-resistant algorithms critical to national security.
• Semiconductors: Advanced chip design and manufacturing remain central to AI infrastructure, military systems, and economic competitiveness.
• Supercomputers: High-performance computing enables training of large AI models and simulations essential for weapons development and scientific research.
• Hypersonics: Hypersonic weapons and delivery systems represent a new frontier in military technology where the US and China are competing for superiority.

The strategy also calls on US allies to implement similar restrictions, attempting to create a coordinated front against Chinese technology acquisition.

How Can the US Resolve This Talent Paradox?

The contradiction between talent recruitment and visa restrictions creates a practical dilemma for universities, research institutions, and tech companies. International students and researchers are not interchangeable with domestic talent; many fields, particularly AI and advanced computing, have global talent pools where the best researchers may come from anywhere in the world. Restricting their access while demanding technological leadership puts the US in a weaker negotiating position.

The visa duration cap is particularly problematic for long-term research projects. A doctoral student in computer science or physics typically needs five to seven years to complete their degree and contribute meaningfully to research. A four-year cap forces students to choose between leaving the country or navigating a bureaucratic extension process, creating uncertainty that may push top talent toward other countries like Canada, the United Kingdom, or Australia.

The $100,000 H-1B fee, meanwhile, makes hiring foreign talent prohibitively expensive for many companies, especially startups and smaller research labs that cannot absorb such costs. This may inadvertently push AI development and talent concentration toward larger corporations like Google, Microsoft, and Meta, potentially reducing competition and innovation diversity.

"NSF expects award activity to proceed while continuing to support critical research through its rigorous, merit-based review process," said Cassandra Eichner, of the National Science Foundation's Office of Legislative and Public Affairs.

Cassandra Eichner, Office of Legislative and Public Affairs, National Science Foundation

The National Science Foundation has released $1.5 billion in new funding solicitations across five directorates for foundational research in science and engineering, offering a partial counterweight to talent restrictions. These grants will support ambitious, transformative research with award durations up to five years, as well as shorter-term grants to test promising ideas quickly. However, researchers have expressed concern that this funding represents only a small portion of the NSF's $8 billion research budget for fiscal year 2026, and some directorates, including Social, Behavioral, and Economic Sciences and STEM Education, are not mentioned in the funding announcements.

The Pentagon has also ordered 30 American universities to investigate their financial and research ties with China, Russia, and Iran by August 31 or risk losing access to federal funding. This security-focused approach may further chill international collaboration and discourage foreign researchers from participating in US-based projects, even in non-sensitive fields.

The administration's strategy reveals a government grappling with genuine security concerns about technology transfer to China, but implementing those concerns through visa restrictions and talent barriers may ultimately weaken rather than strengthen US AI leadership. The most innovative research ecosystems thrive on diversity, international collaboration, and the free movement of ideas and people. By making it harder for global talent to contribute to American science and technology, the US may inadvertently accelerate the very outcome it fears: a world where China develops advanced AI capabilities independently, without relying on American talent or technology.

" }