Dear Aventine Readers, 

In this issue we’re looking at what cuts in US federal funding for scientific research mean for researchers and scientific progress down the line. Europe and China are both trying to attract top US talent, and are showing some success. So far, however — and for a variety of reasons — we are not seeing a mass exodus of scientists leaving for other countries. The experts we spoke to are divided on the lasting effects of funding reductions. Some think it could cause a pause in progress that could be recouped; others believe it could lead to a longer term drying up of the US talent pipeline. 

Also in this issue: 

I hope everyone got a breather over the summer. Hello, September! 

Danielle Mattoon
Executive Director

Subscribe

Subscribe to our newsletter and be kept up to date on upcoming Aventine projects

On May 5, European Commission president Ursula von der Leyen made a pointed offer to the world’s scientists. “We can all agree that science has no passport, no gender, no ethnicity or political party,” she said, in a speech at La Sorbonne. “To every researcher, at home or abroad … our message is clear: Choose science, choose Europe.”

She didn’t mention the United States. She didn’t have to.

If enacted, President Trump’s proposed 2026 budget would cut all US federal science funding by more than 20 percent, from $176 billion to $141 billion, according to analysis by the American Association for the Advancement of Sciences. Funding for the National Science Foundation and National Institutes of Health — the backbone of American academic science — is on the chopping block, with the NSF facing a 56 percent cut and the NIH a 39 percent cut. New NIH grants are already harder to secure; automatic renewals are drying up. 

The rest of the world sees an opportunity to poach top talent, and programs across Europe, Asia, Australia and Canada now promise funding for US-based researchers looking to relocate. The EU has made the biggest commitment, with as much as €500 million offered over the next two years for scientists considering a move. The fear in the US is that a mass exodus of talent could decimate domestic research. “[In the Boston area] we have one of the most vibrant, exciting, innovative ecosystems in the world,” said Kenneth Kaitin, a professor of medicine at Tufts University. “All of that is based on academic research that is constantly spilling out new findings.” But, he said, “that's all under threat.”

Aventine spoke with academics facing cuts, labor economists and think tank researchers, who believe that while an exodus of research talent is unlikely to take place in the short term, the US faces the reality of stalled labs, lost experiments and aspiring PhDs who quietly choose alternative paths. Meanwhile, some of its brightest talents will choose to do their work overseas. 

Packing up and leaving

On June 26, Aix-Marseille University welcomed the first eight US researchers under its Safe Place for Science program, offering up to three years of funding and institutional support to help them continue their work. Nearly 300 researchers applied — including academics from NASA, Berkeley and Stanford — for the 20 spots in total that are available.

Interest, at least, in leaving, appears to be widespread. In a Nature poll of 1,600 US-based researchers published in March, 75 percent said they “are considering leaving the United States” following the current disruptions to science. “Don’t want to leave,” one graduate student responded. “But what’s the alternative?”

Among academics, there is an anecdotal sense that people are already starting to make moves. “I have direct evidence of people packing up their bags and moving,” said Kaitin. 

Yet it is still too early to know what the scale of departures might be. “It's true that researchers in the US are hurting in many various ways,” said Gaurav Khanna, a labor economist at the University of California, San Diego.” “[But] we don't have enough data just yet to know whether it's happening or not.” So far, evidence for the assertion that large numbers of researchers are relocating measures only interest, not action — like the Nature survey or web analytics about job searches and applications, which show that interest in overseas roles has increased by as much as 30 percent among US-based researchers.

But relocating is complicated!

Even if anxiety and discontent are growing, moving lives and labs to another country is a daunting prospect that many people would avoid if they could. Language, salary, funding, visas, spousal employment, familial responsibilities and lab infrastructure all play into the decision, and the dynamics of how such factors interact are poorly understood. Many researchers may try to ride out the current situation for a few years, either until the US midterms or the next presidential election, said Khanna, by taking jobs at lower-tier universities or in industry.

The situation is complicated further by the fact that many nations allocating money to attract US researchers are cutting or considering cuts to their research and higher education budgets due to growing economic pressure and shifting political priorities, Khanna pointed out —  making a move less appealing. The Netherlands recently reduced its higher education budget by €1.2 billion, or about $1.4 billion, while France cut its research budget by €930 million, or about $1.1 billion. The EU’s two-year €500 million Choose Europe program for researchers may be generous, but it represents a little over 1 percent of the total annual funding cut that the AAAS predicts could hit US science research. 

There’s also a question about other nations’ capacity to accommodate US researchers. While funding is limited, so too is lab space, equipment and overall supporting infrastructure like housing and childcare. “The capacity for European, Latin American, Canadian, Asian countries to absorb these people in any reasonable amount of time, meaning over the next decade, is limited,” said John Bound, a labor economist at the University of Michigan, “with one exception, and that's China.”

China has invested heavily in scientific infrastructure and research in recent decades. Its spending on research and development has increased 16-fold since 2000, and it is now second only to the US in terms of total investment. While its Young Thousand Talents Program, designed to lure leading researchers from around the world to its shores, was established in 2008, US funding cuts are “certainly accelerating” China’s ability to attract talent, Bound said. A swath of programs at the national, provincial and city levels offer high salaries, relocation expenses, housing, health care and promises of research funding, to both Chinese nationals working overseas and foreign-born researchers. But none of this makes moving to China a straightforward decision. National security and ethical concerns will likely be significant factors, particularly for researchers in sensitive fields such as AI or synthetic biology.

Who leaves, who stays 

Academic migration is slow-moving and hard to track. Richard Freeman, a labor economist at Harvard, expects preliminary signs by early 2026, when federal statistics and research projects begin to reflect changes in the US academic workforce.

In the meantime, several factors will probably influence the way things play out. Foreign-born researchers are most likely to leave the US and return to their home countries, said experts who spoke with Aventine, pushed away by both a lack of funding and the government’s hostility toward immigrants. While it’s difficult to find detailed data on researcher nationality, as of 2022 13 percent of university faculty in the US were Asian, according to the National Center for Education Statistics. It’s also likely that early-career researchers, with fewer family commitments and without an established research group, are more likely to leave than senior colleagues. 

Beyond that, Freeman predicts US graduates will drop out of or not apply for PhD programs, instead opting for careers in industry, and that undergraduates will go into less research-oriented academic majors.

This suggests that instead of a mass exodus, we’re more likely to see a gradual erosion of scientific talent through a combination of steady emigration and a drying up of the pipeline. “It's going to be death by a thousand cuts,” said Emily McGrath, director of workforce policy at the Century Foundation think tank.

The impact on US science

Measures of US scientific output, often based on citation rates or patent activity, are lagging indicators, so it could take years before we see quantifiable effects from the cuts to science funding. In the meantime, there are guesses as to what effects the cuts will have. 

Freeman suspects that foreign-born researchers create disproportionately high-quality and high-impact research compared to their American-born counterparts, an idea he and his team are investigating. If that turns out to be the case, the damage done to US science will likely exceed the number of people lost. 

Another potential outcome could be a reshaping of which problems get solved and why. “What you're going to see is [that] problems that are tackled are those that have commercial benefit,” said McGrath. “And that's not a good thing for society. We don't want to only be tackling problems that have solutions that are sellable.”

And then there’s the unknowable. Kaitin points out that some research projects will be abandoned or never started in the first place, meaning certain breakthroughs just won’t happen. 

There may yet be a reversal of the funding cuts: a change of heart in the White House, a shifting power balance after the midterms, or a new administration at the next presidential election could all change the course of current policy. And there’s little consensus on the window of opportunity in which things could be turned around. Kaitin contends that damage is already done; Khanna thinks that meaningful damage will occur only if the funding situation persists for a decade or so. 

The prominence of the US as the world’s major scientific power hangs in the balance. How the nation responds will shape not only its future as a research superpower, but also the kind of knowledge that is pursued. 

What should you make of OpenAI’s GPT-5?** On August 7, OpenAI unveiled GPT-5, the latest iteration of its most advanced language model. Critics were underwhelmed. While the model delivers improvements in areas like software engineering and scientific reasoning, they’re not as dramatic as many expected. The lack of a step change in performance suggested “that progress on large language models has stalled,” The New Yorker announced. The main criticism: That because GPT-5's performance is only a little better than its predecessors, it shows the limits inherent in generative AI models — that extra data and computing power aren't yielding the results that many people had anticipated. But some experts argue that this critique misses the mark. Though GPT-5 doesn't exhibit the same leaps in user experience as previous models, according to the AI evaluation lab METR, GPT-5's improvements are actually on pace with those of earlier models if you measure progress according to the complexity of tasks completed. METR tracks the complexity of software development tasks that language models can complete. Its data suggest this benchmark roughly doubles every 200 days and GPT-5 sits right on schedule, capable of finishing a task that would take a human developer just over two hours. (If that pace holds, a model capable of handling a full day of human programming work could arrive within two years.) There are other, similarly non-razzle-dazzle upgrades: ChatGPT now selects the most appropriate AI model behind the scenes, removing friction for end users. And it’s cheaper to run than previous top-tier versions, a shift that seems designed to accelerate adoption. GPT-5 may not feel like a huge leap in user experience, providing responses to prompts that don’t feel all that different than those that GPT-4 could produce. But it seems to be a calculated step forward, with the focus on ease of use and low cost potentially paving the way, SemiAnalysis predicts, for the monetization of hundreds of millions of free ChatGPT users out there.

Tooth regrowth could make dentures obsolete. Clinical trials are underway for treatments that may allow humans to regrow teeth, potentially replacing dentures and implants with living tissue. A number of labs are exploring how to coax the body into regenerating teeth, reports New Scientist, an ability that most mammals have lost through evolution. One approach, being developed at Tufts University, creates an environment in which dental cells can grow before implanting them into the jaws of pigs, where they then develop into tooth-like structures. In another effort, at King’s College London, researchers are working on chemical signals to activate dormant cells in the jaw to prompt regrowth, though they haven’t yet managed to produce full teeth. And in Japan, scientists at Kyoto University have developed a therapy that targets the genetic switches controlling tooth growth. Their startup, Toregem Biopharma, began a Phase I clinical trial in 2024 to test the technique in men missing at least one tooth. That final project’s first real goal is to treat congenital edentulism — a condition in which children never develop teeth — potentially sparing them a lifetime of dentures. Researchers hope these methods could also be used to help adults regrow individual teeth. Progress has been slow so far, in part because dental research receives less funding than other areas of medicine. Still, many technical hurdles have been cleared, raising hopes that biological tooth regeneration could become a routine procedure.

A new bioelectronic implant treats rheumatoid arthritis. A device newly approved by the FDA could offer relief to some of the 1.5 million Americans living with rheumatoid arthritis. The inch-long SetPoint System is surgically implanted in the neck and delivers a one-minute daily electrical stimulation to the vagus nerve, a key part of the nervous system that links the brain to major organs. In a yearlong randomized controlled trial involving 242 patients, the therapy reduced inflammation that leads to rheumatoid arthritis and helped more than half of recipients achieve remission or significantly reduced symptoms without the use of drugs. Currently, rheumatoid arthritis is generally treated with steroids or biologics, potent medication based on living organisms, which can cost thousands of dollars per month and have little impact. The price of the implant hasn’t been disclosed, but the company told The New York Times that it could be less expensive than a year’s worth of some rheumatoid arthritis drugs and is designed to last 10 years. Beyond arthritis, the SetPoint system may signal a broader shift toward bioelectronic medicine. Clinical trials are already underway to test vagus nerve stimulation for inflammatory bowel disease in children, lupus and other autoimmune disorders, with further studies planned for multiple sclerosis and Crohn’s disease.

Where Are All the AI Drugs? from Wired
3,900 words, or about 16 minutes

Most drug candidates fail before reaching the market, a problem that makes drug development painfully slow and expensive. Artificial intelligence has promised to fix that, offering smarter molecule design and faster screening. But so far, not a single AI-designed drug has been approved for use. This story dives into the startups racing to change that, and finds glimmers of hope. One company, Exscientia, used AI to narrow what might typically be thousands of potential cancer drug molecules down to just 136, making it easier to manufacture and test candidates in early stage trials. In one case, a terminally ill patient in a safety trial for one of its drugs surprised the company by living far longer than expected. (Because the trial was blind, it’s unclear whether she received the drug or a placebo, but the company was encouraged by the result.) Other startups have similar stories that support a case for cautious optimism. Even so, unless AI-guided drug discovery can produce therapies that get through trials faster, more successfully and more efficiently than traditional candidates, the technology’s impact will remain theoretical.

What if AI made the world’s economic growth explode? from The Economist 
2,800 words, or about 12 minutes

Techno-optimists envision a future in which artificial intelligence and robotics trigger an era of “radical abundance,” where productivity accelerates so dramatically that the economy grows at unprecedented rates. There are reasons to be skeptical of this scenario: The growth described would be like nothing seen before in human history, and AI may never be capable enough to deliver it. But this story sets aside doubt and imagines what the economy would look like if abundance did arrive. In the most extreme scenario — in which output across knowledge work, manufacturing, agriculture and everything else becomes effectively infinite — wages collapse, leaving owners of capital as the only beneficiaries. More likely, the shift would play out unevenly, with bottlenecks in AI’s abilities and adoption shaping the economy in complex ways. Superstar AI experts may earn fortunes while ordinary knowledge workers are put out of work. The fruits of AI-driven production, such as digital entertainment, may be effectively free, while anything involving human labor, such as babysitting, may become incredibly expensive as it can’t be done more efficiently. Abundance might favor capital, but it may also drive up interest rates, making leverage expensive and stock markets even more difficult to predict. The result described is a transformed economy, shaped not just by what AI can do, but by where it can — and can’t — reach.

How to Vaccinate the World from Asterisk
4,900 words, or about 20 minutes

The Serum Institute of India is probably the biggest biotechnology company you’ve never heard of. During the Covid-19 pandemic, it shipped more than two billion vaccine doses to over 170 countries, more than any other company. According to its own estimates, 65 percent of all children on the planet have received at least one of its vaccines. That’s a staggering reach for a company that began in the ’60s on a faltering horse farm in Pune, India, where early experiments involved growing antitoxins in horses. This story traces the Serum Institute’s rise from obscurity to become the world’s largest vaccine manufacturer by volume. Its success is rooted in a series of shrewd choices: It specializes in manufacturing vaccines developed by others, allowing it to avoid risky R&D; it expanded internationally early on to capitalize on selling into foreign markets; it partners with philanthropic funds to offset costs; and it has invested heavily in scale, producing doses for cents rather than dollars. The next chapter, though, is uncertain. Will the Serum Institute remain the world’s low-cost vaccine powerhouse, or will it start competing directly with global pharmaceutical giants to develop its own first-of-their-kind vaccines?