NSF Experiments With New Kind of Science Funding

Uncle Sam wants you to solve big scientific and engineering bottlenecks outside the hallowed walls of academia. On 14 May the U.S. National Science Foundation (NSF) issued a solicitation inviting what it calls “X-Labs,” or independent research organizations, to apply for a total of US $1.5 billion over ten years. The structure of the X-labs solicitation is new for the United States government and closely matches an emerging private funding model for so-called “focused-research organizations” (FROs) that various think-tanks and philanthropies have proposed and tested during the last six years.

A focused-research organization is a team of scientists, engineers, and other technology developers that works on a well-defined problem with a target duration of 3 to 7 years and on budgets in the tens of millions of dollars. Some examples have sought to build an ultrasound-based brain-computer interface, quantify marine CO2 removal, and improve formal verification in mathematics. The funding for these FROs required a team approach to science larger and more agile than a typical academic lab but with a more academic appetite for scientific risk than a commercial venture might have. In some ways, they echo work done by the Defense Advance Research Projects Agency (DARPA), which is known for helping bridge high-risk research and early-stage commercialization of many technologies.

“The NSF’s X-Labs announcement is a welcome signal that the global research community is serious about finding new ways to fund ambitious, high-risk science,” says Pippy James, deputy CEO of the Advanced Research + Invention Agency (ARIA), a United Kingdom government funding body in London that uses a similar model to the X-Labs.

The NSF’s first two X-Lab research areas are scientific instrumentation for sensing and imaging and interconnects and integrated photonics for quantum systems, and the solicitation says the agency will announce additional topics within weeks.

The funding is structured in phases, with $1.5 million per project in the first year, then up to $50 million per project over the next two to three years for selected projects, with a third, more open-ended phase after that. That first-year funding is more than seven times higher than for the typical NSF project of around $200,000.

“Compared to incremental, project-based grants, larger institutional grants and longer-horizon grants let teams take on harder, more infrastructure-heavy problems with the agility to pivot as they learn,” says Jenn Gustetic, director of metascience and R&D policy at the Institute for Progress, a Washington, D.C. think tank that made a proposal last year for how the U.S. government could support more independent research organizations.

The NSF solicitation also requires applicants to demonstrate “substantial” independence from any non-X-Lab institutions such as universities or companies, which the NSF defined in part to mean the ability to make decisions on research direction, partnerships, and staff in days rather than weeks. It would be difficult for a full-time researcher at a university to qualify, for example, which opens the door to industry researchers or academics willing to take extended leave.

What should new money for science look like?

“People have been kind of wanting to do [focused research organizations] in a fairly bipartisan way since 2020,” says Adam Marblestone, an early proponent who now directs Convergent Research, a Cambridge, Mass., non-profit that has spent almost $400 million building a dozen FROs. Some larger goal-directed, rather than principal-investigator-centered, funding has existed for decades in other federal agencies in the form of the Advanced Research Projects Agencies for defense, intelligence, energy, and most recently health. ARPA program managers often took a more hands-on and flexible approach than typical three-year NIH or NSF grant managers could. The IFP published 2 June an Atlas of Innovation comparing many different research funding structures.

The NSF announcement comes against a backdrop of administration requests for dramatic cuts to the agency’s budget, though Congress has generally appropriated stable amounts. NSF has, however, not disbursed all its appropriated funding, because the Trump administration has been feuding with many universities, accusing them of discrimination, and suing them. Most recently, NSF stopped new funding for several prominent universities, Nature reported. MIT’s president in May said that the university has won 10 percent less federal money than the previous year.

The big-ticket nature of the X-labs might make administrators and principal investigators at those universities worry about whether it will impact their own funding. “I don’t think it’s a zero-sum game,” says Erica Goldman, director of policy entrepreneurship at the Federation of American Scientists, a Washington, D.C. science policy think tank, “but the way the timing of the announcements have come out and the rhetoric out there make it very hard to see that.”

“NSF X-Labs is structured to complement the existing system, not displace it — adding an independent institutional type alongside universities, national labs, small businesses and corporate R&D,” Gustetic says. The X-Labs annual sticker price represents less than 2 percent of the agency’s overall budget of $8.75 billion in 2026.

The emerging field of meta-science, which investigates how best to do science, has been debating how governments should build the proper pipelines for converting blue-sky research into returns for all taxpayers. Meta-scientists differ over how FROs should fit into the research system. Some argue that governments can’t expect to apply the vaunted DARPA model to everything. Others write that FROs are a great idea and that the federal government already has a version of them in the form of the Department of Energy’s National Labs, which have spun off science platforms such as the Human Genome Project and the Protein Data Bank.

NSF media affairs head Mike England told IEEE Spectrum that “X-labs creates space and provides funding for new institutions to achieve breakthroughs in scientific discovery, research, and translation, and ultimately help create new platform technologies.” In addition, on 27 May NSF requested information for a new funding idea adjacent to X-Labs it calls Tech Accelerators. These would use NSF money and accelerator expertise to fund and guide “deep-tech” commercialization efforts in agriculture, materials, ocean, and scientific instrumentation.

Other elements of government are also exploring how to incorporate the FRO funding model. In December 2025, U.S. Representative Josh Harder (D-Calif.) introduced a bill that would apply the X-labs model to the National Institutes of Health (NIH). England says that NSF is having conversations with other government agencies about the model and welcomes more agencies to explore it.

“We don’t have great evidence comparing how different funding mechanisms perform—individual project grants, milestone-based contracts, prize challenges, etc.” she says. “X-Labs is a chance to actually learn something about which institutional designs work for which kinds of problems.”

Universities will likely have to adapt to the new model. Mid-career academics may well want to take time away from their university homes to participate in an X-Lab or other FRO project, says Monica Dus, director of the Office of National Laboratories at the University of Michigan and a holder of NSF grants. To welcome them back, universities will need to figure out how to cover teaching duties in their absence, and how to assess commercial experience for tenure or other internal promotions. “Universities should adapt to make sure the research does really reach the people it is meant for,” she says.

Academics may also need to change their approach to succeed, Convergent’s Marblestone says. “When we go to academics at Convergent, it takes a few conversations to plan it, because they don’t always know how they’d manage $50 million and a professional engineering team. You really need a CEO.”

, CEO of the Federation of American Scientists, which published a

, is optimistic about the NSF’s ability to get results from X-Labs. “The team at NSF did a lot of due diligence talking to folks on the outside, not just policy people but people that are building these labs, and integrated some of the key elements into the contours of the solicitation,” he says.