A Nobel Laureate Left Berkeley for Beijing on the Fourth of July. The U.S. Has Spent $675,000 Training Each Scientist Now Heading for the Door.
Omar Yaghi took the 2025 Nobel Prize in Chemistry to Tsinghua University. He is the most visible departure in a brain drain that, by the numbers, is costing the United States far more than the funding cuts that caused it.
Six hundred seventy-five thousand dollars. That is the approximate public investment the United States makes to train a single research scientist through a doctoral program and post-doctoral fellowship — counting federal grants that cover stipends, tuition waivers, laboratory overhead, and the institutional infrastructure that keeps world-class universities running. It takes seven to eleven years. Most of the money comes from NIH and NSF grants that flow through the same pipeline the current administration has proposed cutting by 41% and 55%, respectively.
On July 4, 2026, Omar Yaghi walked away from that investment.
Yaghi shared the 2025 Nobel Prize in Chemistry for inventing metal-organic frameworks — materials with internal surface areas so vast that a single gram can have the surface area of a football field. His MOFs harvest drinking water from desert air, capture carbon dioxide from industrial exhaust, and store hydrogen for clean energy. He left the James and Neeltje Tretter Professorship at UC Berkeley, one of the most prestigious chairs in American chemistry, for Tsinghua University in Beijing, where he will lead a new AI-driven materials research center.
He is not an outlier. He is a data point on a curve that has steepened every quarter since January 2025, and the economics of that curve are worse than the headlines about “brain drain” have conveyed. Much worse.
Three Datasets Nobody Combined
Three datasets tell the story when you read them together, which nobody has done publicly.
First, the demand signal. A Nature poll of more than 1,600 U.S. researchers found that 75% are considering leaving the country. Among postdocs — the people who will staff tomorrow’s labs — that figure was 79.4%. Among PhD students, 75%. These are not hypothetical intentions from people who will never act. The Nature’s Careers jobs board showed a 32% increase in applications from U.S. scientists to overseas positions in Q1 2025 compared to a year earlier, and a 68% spike in March alone for views of non-U.S. job postings. Applications from European and Chinese scientists to American positions collapsed.
Second, the supply response. Consider the European Research Council, just one of dozens of international funding bodies, saw early-career grant applications from U.S.-based scientists rise from 60 in 2024 to 169 in 2026 — nearly a threefold increase. Senior researcher applications went from 23 to 114 in the same period. Nearly five times as many established American scientists applied to a single European funder in two years.
Third, the economic value of what is leaving. Through the United for Medical Research coalition, NIH reported that its $36.9 billion in FY2024 grants generated $94.5 billion in economic activity, supporting more than 408,000 jobs nationwide. That is $2.56 in economic activity for every dollar invested. Every $100 million in NIH funding produces 76 patents, and those patents generate an estimated $598 million in further research and development. A dollar of basic research stimulates $8.38 in industry R&D within eight years.
These datasets exist in separate reports, cited in separate news stories, read by separate audiences. When you put them in the same spreadsheet, the picture is unambiguous.
What a Departing Scientist Actually Costs
Start with the training investment. A typical biomedical PhD requires five to seven years of federal funding: stipend ($37,000–$41,000 at NIH’s current NRSA rates), tuition remission ($30,000–$50,000 at research universities), and institutional overhead (the indirect cost rate that covers facilities, administration, and compliance, averaging 55% at major research universities). That totals $325,000–$490,000 for the doctoral phase alone, nearly all of it federally funded.
Postdoctoral training adds another two to four years at $55,000–$65,000 in salary plus the same overhead multiplier: $170,000–$400,000 more. Median public investment per researcher through the pipeline: approximately $675,000. Some cost less, many cost far more, but $675,000 is a defensible midpoint based on NIH pay scales and published indirect cost data.
Now calculate what that investment returns. Dividing total NIH-generated economic activity ($94.5 billion) by supported jobs (408,000) yields roughly $232,000 per position per year in direct economic contribution. Over a 25-year research career — the median span for an academic scientist who begins an independent lab at 35 and retires at 60 — that is $5.8 million per researcher in direct economic activity alone, before accounting for the $8.38 multiplier that compounds basic discoveries into industry R&D over a decade.
Benjamin Jones at Northwestern and Lawrence Summers at Harvard calculated the broader number in “A Calculation of the Social Returns to Innovation”: each dollar invested in R&D yields approximately $5 in GDP returns, a figure Jones calls “remarkably good” and “an investment where you put in $1 and get $5 back.”
The Exit Bill
Assume conservatively that 5% of the 408,000 NIH-supported researchers actually leave — a fraction of the 75% who say they are considering it, but consistent with the fivefold increase in ERC applications from senior scientists, which represents real job-search behavior rather than poll sentiment.
| Metric | Value |
|---|---|
| Researchers departing (5% of 408,000) | 20,400 |
| Sunk training investment ($675K each) | $13.8 billion |
| Annual economic activity lost ($232K each) | $4.7 billion/year |
| 25-year career value (direct) | $118.3 billion |
| Industry R&D multiplier ($8.38 per $1 basic research) | Compounds annually |
$13.8 billion in sunk training costs, written off. $4.7 billion per year in lost economic activity, recurring. Over the career span of those researchers, more than $118 billion in direct economic contribution walks out the door. And those are the conservative numbers — the NIH-only researchers at a 5% departure rate. Proposed cuts also hit NSF ($3.9 billion, −55%), NASA ($3.8 billion, −34%), and DOE ($3.2 billion, −16%) as well, each with their own research workforces.
Compare that to the proposed savings. Under the FY2026 budget request, civilian R&D spending would fall by $48.7 billion. But NIH alone generates $2.56 in activity per dollar. Cut $19 billion from NIH, lose $48.6 billion in economic activity. Net savings after accounting for economic impact: negative $29.6 billion. Add the brain drain costs, and the fiscal arithmetic becomes not merely unproductive but self-destructive.
Who Is Picking Up the Tab
Every researcher the United States trains and then exports arrives at their destination pre-assembled. No country on the receiving end pays training costs. They inherit the network effects — the collaborations, the institutional knowledge, the student pipelines — that took decades and billions of dollars to build. A global bidding war is already open.
| Country/Region | Recruitment Budget | Per-Researcher Offer |
|---|---|---|
| Canada | C$1.7 billion | Varies by program |
| EU (“Choose Europe”) | €500 million | ERC grants: €1.5–2.5M |
| Austria | €12.5 million (initial) | €500,000 / 2 years |
| France (Aix-Marseille) | €10–15 million | ~€700K per researcher |
| China (provincial/city) | Undisclosed (16× R&D since 2000) | $14K–$140K relocation + lab |
Austria’s Academy of Sciences recruited 25 researchers from Harvard, MIT, and Princeton with €500,000 grants. Its president, Heinz Fassmann, said the quiet part out loud: “Thank Trump for this brain gain.”
China’s play is more strategic. Tsinghua did not merely offer Yaghi a position; it offered him a new AI-driven research center with resources to compress materials development cycles “by orders of magnitude.” China’s R&D spending has increased sixteenfold since 2000, making it the world’s second-largest investor in research after the United States. Provincial and city governments have added their own bounties: Taizhou offers $14,000 to any university graduate who relocates, Gulin County provides $42,000 to PhD holders, and Hunan Province pays $140,000 to doctoral students arriving from abroad.
Sara Seager, a planetary scientist at MIT, is moving to the University of Toronto in September. “There are many reasons why I’m returning to Canada,” she told Reuters, “and one of them is the budget cuts and also the huge uncertainty in science funding in the U.S.” Yang Dan, an influential neuroscientist at UC Berkeley, has already relocated to Beijing. Ardem Patapoutian, a 2021 Nobel laureate in Physiology or Medicine, was actively recruited by China after his NIH grant was frozen. Terence Tao, the Fields Medalist and UCLA mathematician — widely considered the greatest living mathematician — is weighing his options publicly.
What One Departure Costs
Yaghi’s departure illustrates what the aggregate numbers cannot. He invented an entirely new class of materials. MOFs now have applications in carbon capture, water harvesting, hydrogen storage, natural gas purification, and drug delivery. He trained approximately 200 researchers during his career, nearly half of them Chinese nationals. His work spawned Atoco, a startup building water-harvesting devices for arid regions using bespoke MOFs designed in his Berkeley lab. It connected to BASF, which supplies commercial MOFs for AirJoule’s atmospheric water generators.
Consider the analogy to the Human Genome Project: a $300 million public investment that generated, by NIH’s own accounting, $265 billion per year in total economic impact and 850,000 jobs. One foundational research program created an entire industry. MOFs are not yet at that scale, but their application space — carbon capture alone is projected to be a $50 billion annual market by 2030 — is growing along a similar trajectory.
All future innovations from Yaghi’s lab now flow through Tsinghua. His next 200 trainees will be Chinese, not American. Patents will be Chinese. Startups will be Chinese. And the industrial applications — the carbon capture systems, the water harvesters, the hydrogen storage that could underpin a clean-energy economy — will be built in China first, by people who learned from the scientist America trained for forty years and then pushed away.
Against Panic
Brain drain fears have a history of being overstated. During sequestration in 2013, similar warnings circulated, and the U.S. research enterprise survived largely intact because Congress reversed the worst cuts within two years. American universities retain structural advantages — deeper infrastructure, more diverse funding sources, a venture capital ecosystem unmatched anywhere else — that make them resilient to political disruption. Administration officials have pointed out that even with proposed cuts, the U.S. would still account for the most global research funding of any single country.
More substantively, the 75% figure comes from a self-selected Nature reader poll, not a probability sample; researchers angry about funding cuts are more likely to respond. The actual departure rate will certainly be lower. Many who “consider” leaving will stay because their spouses have jobs, their children are in school, or because the competitive salaries, equipment budgets, and intellectual freedom available at top American universities remain unmatched. Congress has already partially mitigated the worst proposed cuts: the NSF got $8.8 billion for FY2026 instead of the $3.9 billion the White House requested.
And China has its own problems as a destination. Academic fraud, political interference in research agendas, surveillance of academics, and an intellectual property regime that makes many Western scientists nervous. The EU’s €500 million recruitment program is less than 1% of the total U.S. funding it hopes to poach, a rounding error in terms of absorptive capacity.
Limitations
The $675,000 per-researcher figure relies on published NIH stipend rates and average indirect cost data; actual costs vary widely by institution and field. The $232,000 annual economic contribution per position uses an input-output model that may double-count activity where grant dollars circulate within the same institutional economy. The 5% departure assumption is illustrative, not predictive, and could prove either too high or too low. The Jones/Summers $5 GDP multiplier for R&D investment is a macroeconomic estimate with broad confidence intervals. Competitor recruitment numbers, especially China’s, may be inflated for political signaling. And the most important caveat: academic transitions take years, so the true scale of departures will not be measurable until 2028 at the earliest.
What You Can Do
If you are a university administrator, the ERC application data is the most actionable leading indicator available. Senior researcher applications from the U.S. quintupled in two years. That trajectory means your most productive faculty are already in conversations with overseas institutions, and retention packages offered now cost a fraction of replacement. Quantify the grant revenue, patent income, and indirect cost recovery each at-risk researcher generates, and present the board with the net cost of losing them versus the cost of keeping them. The math will make the case faster than any argument about academic freedom.
If you are a policymaker, understand the asymmetry: cutting $1 from NIH destroys $2.56 in economic activity immediately, and $8.38 in industry R&D over eight years. The proposed $19 billion NIH cut produces a net economic loss of $29.6 billion before counting a single departing researcher. Pair that with the sunk cost of training ($675,000 per person) and the compounding loss as those researchers generate returns abroad, and the fiscal case for restoring funding is not ideological. It is arithmetic.
If you are a researcher, watch the ERC and Canada’s C$1.7 billion program for timeline pressure — recruitment windows close, and the best-funded positions are going now. If you are a taxpayer, understand that the argument for cutting “wasteful” research spending has a testable implication: if the researchers were unproductive, nobody would pay to recruit them. Austria, Canada, France, China, and the European Commission are not spending billions to hire mediocre scientists. They are spending billions because the people America trained are the best in the world, and they are, for the first time in modern history, available.