The Trump administration has proposed a new fee for H-1B visas, while weighing a $100,000 fee on Optional Practical Training, the program that allows international graduates to work in the U.S. for one to three years after finishing their degrees. Roughly 419,000 people were working on OPT in 2024. For many of them, it was likely a major factor in choosing an American university over a school in Canada, the U.K., or Australia.
It is ironic that just weeks ago, headlines lamented the fact that one of the most talented international students we’ve seen in recent years chose to return to China, the U.S.’ biggest competitor. Yang Zhilin then went on to set up Moonshot AI, whose Kimi K3 open model has drawn comparisons to models from OpenAI and Anthropic. Why didn’t he build his company in the U.S.? After all, Yang got his Ph.D. at Carnegie Mellon University. He interned at Google and Meta, and received a job offer from Apple. But Yang wanted to build his own company, and our visa system didn’t have a way to keep him here in the way that he wanted.
Some experts have proposed stapling a green card to every AI Ph.D. diploma — and they’re not wrong, but this idea and the OPT assume that the brightest minds want to work for our big tech companies first, then build after learning from the giants for several years. That’s not the case anymore.
Work experience isn’t as relevant when AI has flattened and accelerated everything. Founders are starting companies in their mid-20s — straight out of or even before finishing their Ph.D. The population capable of starting a frontier AI company is small, and more than two-thirds of these individuals are foreign. That cohort isn’t likely to want to spend five years or more in someone else’s company waiting for a green card, watching their idea age out.
40% The share of Nobel Prizes won by Americans in chemistry, medicine, and physics since 2000 that have gone to immigrants in the country.
And what does the U.S. do? It tightens and slows down its immigration system. In addition to the $100,000 for OPTs, the window for students to remain in the country after finishing a degree has been cut from 60 days to 30 to combat “rampant visa abuse.” Which means students will spend the last year of their Ph.D. stressing about finding a job quickly, instead of thinking about how to commercialize their ideas.
The Trump administration is also trying to end birthright citizenship, which is another draw for immigrants. Officials even floated the idea of barring foreign employees at U.S. labs from frontier-model access before the industry pushed back.
Other nations have taken notice of these actions, and are using them to their own advantage. Canada opened a program for U.S. H-1B holders and had to close it within a day since it was oversubscribed. Australia, France, and others are openly running recruitment campaigns aimed at researchers based in the U.S.
China spotted an opportunity even earlier: It launched the Thousand Talents Plan in 2008 and pulled back more than 7,000 people within a decade. China has recently increased funding for research while the U.S. has cut science funding, and that is proving to be a big incentive for some. Omar Yaghi, who was born into a refugee family in Jordan and came to the U.S. as a teenager to study, shared the Nobel Prize in chemistry last year. He recently left his faculty post at the University of California, Berkeley, for a position at Tsinghua University in Beijing, where he will lead an institute using AI to accelerate the discovery of new materials.
Yaghi was just one of many immigrant Nobel laureates in the U.S. Immigrants have been awarded 40% of the Nobel Prizes won by Americans in chemistry, medicine, and physics since 2000, according to a recent report by the National Foundation for American Policy, a research organization.
The high number of immigrant Nobel Prize winners reflects “a greater openness to immigration that has helped make the U.S. the leading global destination for research in many different science and technology fields, including computer and information sciences, cancer research and others,” the NFAP said.
If we persist in making it harder for international students to come to and stay in the U.S., this may no longer be true. Today, the Pentagon’s Office of Strategic Capital, set up in 2022, will lend a company between $10 million and $150 million to build technologies deemed to be critical to national security. What we also need is an Office of Strategic Human Capital that can identify brilliant international AI students in critical fields and say, here is your long-term American visa, here is $10 million, go build.
In September, the U.S. and China are scheduled to hold talks on AI. Chip export control is sure to be a hot topic of discussion. What they will not admit is that you can’t export-control where a person decides to build a company. For decades, the U.S. was the first choice, but this is not the case anymore.
Immigration reform will take years, and the damage that is being done now will be hard to undo. For decades, for people like me and my family, immigration to the U.S. was a one-way door: You came, you stayed, you built here. The door swings both ways now. Yang Zhilin and Omar Yaghi are just the most visible examples of this. They will not be the last, and if the U.S. does not act, China will pull ahead in the race for AI talent.
Facts Only
* The Trump administration proposed a new fee for H-1B visas.
* A $100,000 fee is being weighed for Optional Practical Training (OPT).
* Roughly 419,000 people were working on OPT in 2024.
* Some international students used OPT as a factor in choosing U.S. universities over Canadian, U.K., or Australian schools.
* Yang Zhilin established Moonshot AI, whose Kimi K3 open model draws comparisons to OpenAI and Anthropic.
* Yang Zhilin received a Ph.D. from Carnegie Mellon University and interned at Google, Meta, and received an offer from Apple.
* Experts have proposed stapling a green card to every AI Ph.D. diploma.
* Work experience is considered less relevant as AI has accelerated development.
* The population capable of starting frontier AI companies is small, and over two-thirds are foreign.
* The window for students to remain in the U.S. after finishing a degree has been cut from 60 days to 30 days to combat visa abuse.
* The Trump administration is trying to end birthright citizenship.
* The Pentagon’s Office of Strategic Capital will lend funds ($10 million to $150 million) for critical national security technologies.
* Omar Yaghi received the Nobel Prize in chemistry and left a faculty post at UC Berkeley for Tsinghua University in Beijing.
Executive Summary
Full Take
The narrative juxtaposes established immigration pathways and perceived meritocracy with the rapidly shifting dynamics of the AI talent landscape. The core tension lies between the traditional model—where immigration facilitates talent attraction, leading to high-level research and corporate innovation—and a newly emerging reality where the speed of technological change renders fixed immigration rules obsolete. The text suggests that the prioritization of an extended post-graduation stay (OPT) is based on an outdated assumption: that brilliant minds require several years within established corporate structures before pursuing independent ventures. The argument shifts from valuing tenure in major tech firms to valuing immediate entrepreneurial capacity, especially among foreign AI talent who are increasingly being incentivized elsewhere or are simply seeking routes for mobility.
The pattern reveals a systemic friction where institutional policy (immigration law) moves at a glacial pace while technological and demographic forces move exponentially. The suggested solution—linking immigration status directly to high-value skills (like AI research) rather than just educational attainment—points toward decoupling talent mobility from restrictive employment-based visa structures. The reference to immigrant Nobel laureates further suggests that the U.S.'s historical advantage in attracting scientific talent is now being re-evaluated in light of alternative global incentives, such as China's talent programs, which are demonstrably effective at mobilizing research. The implication is that restricting the flow of highly mobile, high-skill human capital risks forfeiting leadership in the next technological frontier by failing to secure the necessary expertise domestically.
What if the focus shifted from managing the entry and stay of individuals to strategically positioning the nation as the global destination for AI innovation? If the primary constraint on talent mobility is systemic rather than administrative, then restructuring immigration policy could become an instrument for strategic national development, rather than purely a mechanism for control or deterrence. What specific mechanisms must be established to ensure that high-potential international researchers are incentivized to establish their ventures within the U.S., especially when global competitors are actively pursuing similar talent streams? What is the long-term cost of maintaining a system designed around sequential integration versus one designed around immediate, strategic deployment of cutting-edge human capital?
Sentinel — Human
The text presents a compelling argument linking U.S. immigration policies directly to global competition in the AI talent space, supported by specific historical context and personal examples rather than pure statistical recitation.
