THE YEARS since the second world war have been good to American science. The country’s researchers sent man to the moon, invented the modern computer and led the sequencing of the human genome. Its government funded basic research in academia and national labs, which could then pass on findings with commercial promise to firms.

Now there are signs the innovation machine is breaking down. The system’s critics fret that America’s scientific infrastructure encourages groupthink, since a small number of committees distribute a large share of funding for basic research. They worry that universities are ill-suited to the interdisciplinary work that much of 21st-century science requires. And America is no longer an undisputed scientific superpower: China’s research and development (R&D) spending has caught up with Uncle Sam’s in purchasing-power terms, and is growing faster. In 2024 China overtook America in the Nature Index, a measure of publications in the most prestigious scientific journals.

Advocates of “metascience”, the study of scientific discovery, have spent years pointing out the foibles of the system. Lately they have been putting their ideas into practice. Private institutions, many of which were founded within the past five years, say they are building the future of science. The history of American innovation suggests they may be right.

Many of the most exciting American discoveries of the 20th century were not dreamt up on university campuses, but within corporate research institutes. Scientists at Bell Labs (the research arm of AT&T), Xerox PARC (the copy-machine maker’s lab) and IBM Research came up with such marvels as the transistor, and won a handful of Nobel prizes in so doing.

Yet for all their successes, private labs failed to justify their upkeep. A paper published in 2004 by William Nordhaus, an economist at Yale University, estimated that, in the post-war period and across the whole economy, just 2.2% of value from innovation flowed to firms. In response companies transitioned to “R&D by M&A,” buying startups spun out from universities rather than funding their own labs.

Corporate labs’ retreat led basic research to become much more reliant on government-backed efforts. In the view of private-science advocates, that is lamentable. The concentration of science funding means fewer innovative ideas can benefit society at large, says Caleb Watney of the Institute for Progress, a think-tank. To wit, a study of 100,000 National Institutes of Health (NIH) grant applications, conducted in 2025 by Pierre Azoulay of the Massachusetts Institute of Technology and Wesley Greenblatt of Harvard Medical School, found that projects which took out-of-consensus approaches were less likely to have their funding renewed.

And seeking funding is onerous. According to one survey, researchers spend about 44% of their time doing administrative tasks like writing proposals and reports.

To illustrate these issues, many cite the example of Katalin Kariko, a biologist whose work provided much of the science underpinning mRNA vaccines, such as those for covid-19. For decades funders, notably the NIH, rejected her proposals, considering them too risky. She happened to stay in science, but many others in her position would not, says Stuart Buck of the Good Science Project, an advocacy group.

Such concerns have made their way from metascientists’ lips to powerful ears. On July 21st the American government released a report arguing that universities should not be “the only home of America’s most innovative scientific research”. Instead funding agencies should act more like venture capitalists, betting on a wide variety of institutions.

As part of this push, the National Science Foundation (NSF) is set to allocate $1.5bn to fund its first round of “X-labs”, new organisations that “advance breakthrough science outside of traditional institutions”. A bipartisan bill in Congress would create a similar initiative at the NIH, which has a much larger budget, notes Josh Harder, a Democratic representative and one of the bill’s sponsors. Other changes would see federal agencies offer “golden tickets” to oddball projects, or begin to give long-term money to individuals rather than projects.

Such ideas have discomfited scientists because of their source. As it attempts to spark “a new golden age” for American science, the Trump administration wants to fund it less. The White House’s proposed budget for fiscal year 2027 suggests halving the NSF’s budget. The worry is that the combination of less funding spread across more institutions will mean university labs lose out.

Another entrant to the world of private science is Episteme, a company launched in 2025 by Sam Altman of OpenAI and Louis Andre, an elusive researcher. Rather than demanding pre-set targets or papers, Episteme hopes to give its lead researchers near-total freedom. While the select few it funds often have an approach in mind to solving their chosen problem, Episteme offers them the flexibility to adapt. “If an idea fails, which in science is something that happens very often, you should be able to pivot and try a different angle,” says Mr Andre.

Many important questions fall outside that scope and will have to remain within universities. Not every attempt at private science has flourished. Arena BioWorks, a private biomedical institute founded in 2023, shut down after less than two years. But whereas Arena’s funders reportedly expected quick payouts, Mr Andre says his backers, who have funded Episteme from their personal fortunes, are committed for the long haul. They are not betting on one specific area of research, either. “If a programme fails, it’s okay. We can recycle it and repurpose the resources into something else,” he reckons.

Between federal support, independent experiments and a coming wave of AI-fuelled cash, private science is “seeing the stars align,” says Mr Marblestone. Historians of science and technology argue the post-war system cemented American hegemony. In the bet on private science the stakes are high, both for boffins and for America. ■