Science in the United States is in a state of crisis, induced by a government whose hostility to it is without precedent. The Trump administration is determined to make the national research enterprise much smaller and to radically transform what remains, with the goal of placing it under direct political control in support of an oligarchic agenda. While the administration hasn’t yet been able to achieve this goal fully, it has already done great damage and is in the process of doing more, particularly to fields such as my own—atmospheric and climate science—that are seen as unhelpful to this agenda. The consequences will be long-lasting.
It can be difficult to understand what’s happening, though, even for those paying attention. When I’ve spoken with colleagues overseas in recent months, nearly all have asked me something along the lines of: “We are reading the news from the United States. It looks like a total disaster, particularly for science and scientists, and even more for those who study climate or anything close to it. Yet, when we communicate with our colleagues over there, they seem to be working more or less normally, rather than being in a state of panicked distress. How can both of these things be true? What is going on?”
I tell my overseas colleagues that both impressions are correct, and that in fact the cognitive dissonance is even stranger from the inside.
For many of us, daily life and work are still going on, in some reasonable facsimile of what they looked like before Donald Trump was elected president for the second time. And yet again, at the same time, the foundations of American science are being profoundly shaken, if not dismantled entirely.
Research funding is being sharply cut in practice, even though Congress passed agency budgets close to those of previous years. Institutions and practices are being replaced by entirely different ones, or with nothing. The changes are entirely politically driven and are being made rapidly, with little or no input from the scientific community.
The consequences of these dramatic and unwelcome changes are difficult to predict in detail. The degree of long-term damage will depend on the 2026 and, especially, 2028 elections. But the harm will outlast the Trump administration, and some of it may be permanent.
And yet it’s true that today, many scientists at many institutions are still doing their work.
Research is being performed. Data is being collected and analyzed, experiments and simulations are being performed, papers written, students educated.
Maybe this means that our institutions are more resilient than they might have looked a year-and-a-half ago, and that American science is likely to emerge from this period in some semblance of its former condition. On the other hand, maybe we’re just doing science, metaphorically, on the deck of a ship that’s heading straight for an iceberg but just hasn’t quite hit it yet.
In the DOGE period, early in the second Trump administration, Elon Musk and his team fired as many scientists as they could, as quickly as they could. Agency budgets were slashed, buildings closed.
The combined viciousness and stupidity of the DOGE phase are evident in one example that is literally very close to home for me. In May 2025, the NASA Goddard Institute for Space Studies was forced to leave its building, just a block from where I live. The scientists at the NASA Goddard Institute for Space Studies were told to work remotely. But the federal government rents the building from Columbia University, and the lease doesn’t expire until 2031, so the government (via the General Services Administration) is still paying rent. Eventually, the scientists were told that they had to report to work in person; to make that possible, NASA is now renting other space on the campuses of several different universities in the city, including Columbia (Cokinos 2026). Yet the original building remains empty, the government paying rent to keep it that way. This is what this administration considers “eliminating fraud and waste.”
(Editor’s note: When asked for comment by the Bulletin of the Atomic Scientists in mid-October 2026, a NASA spokesperson said the agency canceled its Armstrong Hall lease and temporarily put employees on remote work agreements, and “Those employees now have new physical workspaces as part of recent agreements NASA has formalized with Columbia University and New York University.”)
As 2025 proceeded, it started to seem possible that every major government lab that was doing climate science might close. As the NASA Goddard Institute for Space Studies was going through the torment with its building, the president’s budget had also proposed eliminating NOAA’s Office of Oceanic and Atmospheric Research. This would have shuttered many labs critical to US weather and climate science, including the Geophysical Fluid Dynamics Laboratory in Princeton—an outgrowth of the original development of numerical weather prediction in the 1950s on some of the first digital computers, and later, the generator of a Nobel Prize to its pioneering climate modeler, Syukuro Manabe. And in December, the Trump administration announced its intent to close the National Center for Atmospheric Research in Boulder, Colorado.
The NASA Goddard Institute for Space Studies, the Geophysical Fluid Dynamics Laboratory, and the National Center for Atmospheric Research each do many different things, but one thing they all do is develop, maintain, and run their own Earth system models. Together they contribute three distinct sets of models and results to the Coupled Model Intercomparison Projects—international collaborations that are essential to global climate science as well as every Assessment Report of the Intergovernmental Panel on Climate Change. This, no doubt, put targets on them in the eyes of the Trump administration. Closing any one of these three labs would have been a huge blow to the fields of atmosphere, ocean, and climate science; to close all three in one year would have been inconceivable.
It hasn’t happened—not yet, anyway. All three labs are still there. I know scientists at all three, and as far as I can tell (from what is admittedly a very selective and anecdotal sample) they don’t seem to be in quite the state of existential shock and dread that they were a year ago. I suppose the simple fact that life has gone on for another year, and that the Trump administration’s attention has drifted to bigger crises it has manufactured—Iran and trade wars—is evidence of the natural human tendency to assume that tomorrow will be similar to yesterday and today, and thus that the labs will survive. But the threats haven’t passed, and probably won’t as long as Trump is in office. The fate of the National Center for Atmospheric Research is tied up in lawsuits, and the NASA Goddard Institute for Space Studies remains in exile.
And then there’s the universities. My own, Columbia University, was the first to be attacked, early in 2025. The excuse was the Gaza protests and the university’s alleged tolerance of anti-semitism. The weapon, never used before in this indiscriminate (and almost certainly illegal) way, was the freezing of research funding—mostly to the medical school, simply because that’s where the most funding goes, but we in the other sciences lost some too. Some, but not all, was restored when our university administration signed a deal with the Trump administration, the shame of which we will live with for a long time.
Since then, we are no longer seeing grants canceled or suspended on any large scale. Instead, the number of new grants being awarded to all universities has just been severely reduced.
The Trump administration has issued President’s Budgets featuring catastrophic cuts to the science agencies, including a 57 percent cut to the National Science Foundation in 2026 compared to 2025, and 55 percent in 2027 compared to 2026. Congress has ignored those and funded the agencies at levels close to previous years, so the administration has changed tactics. Now it is intervening directly in the agencies to transform American science: How the science is done, on what topics, and who does it.
The National Science Foundation (NSF), the nation’s core agency for scientific research, currently has no director; nor does it have any members on its governing National Science Board, after Trump fired them all in April. (Like NASA Goddard Institute for Space Studies, the NSF was also evicted from its building last year, though their building had recently undergone major and costly renovations specifically to the agency’s specifications, and after a year of exile the NSF is now in a much smaller building.) One of the fired board members told Inside Higher Ed (Quinn 2026) that the Office of Management and Budget, led by Project 2025 author Russell Vought, is now “effectively running the agency.”
The consequences of the hostile takeover of the National Science Foundation are clear. In fiscal year 2026 (just ended as of the time of this writing), the agency held back $1.4 billion—i.e., it refused to spend the money on research, despite being told to do so by Congress. This amount is about a fifth of the budget allocated by Congress, and this refusal to act clearly violated Congress’ intent and infringes on its constitutional power of the purse. The number of grants was cut more, with only 6,218 grants awarded, compared to an average of over 11,000 awards in the period 2021-24. For the geosciences, the cuts are worse: 51 percent fewer grants, 32 percent less funding compared to the 2021-24 period. Besides the cuts in total funding, the even sharper cuts in the number of grants (implying an increase in average grant size but many fewer research projects) reflect a push towards centralization, which facilitates easier political control.
And the Trump administration intends for these numbers to get worse. As reported in the September 11 issue of Nature, all planned NSF geoscience funding for the coming year is to occur under a single call for proposals, as opposed to many each year in the past (Garisto 2026). A National Science Foundation Program Solicitation says that “The agency projects that this funding opportunity for earth, atmospheric and ocean sciences will be allotted about $110 million and that this amount will support 200–350 grants (NSF 2026). In 2023, the solicitations for these three areas totalled (sic) more than $450 million and funded some 1,000 awards.”
The administration’s approach now is less chaotic and more deliberate than it was in the DOGE period. There’s now a veneer of intellectual argument behind the raw exercise of power, as expressed by Trump’s science adviser, Michael Kratsios, in a recent report entitled Science: A New Golden Age.
That report does say some true things. For example: Scientists spend too much time writing proposals, peer review can be biased against big new ideas, and universities are structurally not great at user-driven research.
But the remedies it proposes are much worse than the disease. For example, the report calls for shifting funding away from basic research (and away from universities) and towards applied work with clear pathways to commercialization—which neglects the fact that basic R&D is the very foundation (Sobel 2026) on which applied work rests. (To understand the difference between basic R&D and applied work, consider this: When Albert Einstein hypothesized in 1905 that sometimes light acted as a series of discrete packets instead of as a continuous wave, and that a packet of light could strike a material and cause it to emit an electron—an observation for which later he won the Nobel Prize in Physics—that was basic R&D. It laid the theoretical groundwork for what happened in 1954, when a trio of scientists at Bell Labs figured out how to use that understanding to create a silicon-based wafer that would convert about six percent of the sunlight that fell on it into usable energy—the first practical photovoltaic cell, a technology subsequently used to power satellites, businesses, and people’s homes. What happened at Bell Labs was an example of applied work on the way to commercialization. But Einstein’s basic R&D had to come first.)
Nevertheless, with this presidential administration, the cart is put before the horse: Commercialization first, then understanding the forces you are dealing with. Consequently, AI and quantum computing are in; understanding the universe for its own sake—or even worse, understanding our own planet for the purpose of improving its inhabitants’ collective well-being in the face of dramatic human-induced change—is out.
And worst of all, by making all funding subject to review—and possible claw back at any time, even mid-project—they make science a pawn of political appointees. This political control is clearly the point, and a radical disruption from NSF’s long history of being run largely by scientists.
The scientific community has had no significant input into any of the massive changes at either the National Science Foundation or the other science agencies. Kratsios himself has a Bachelor’s degree in political science and no experience as a working scientist. He does have experience as a venture capitalist working with billionaire conservative activist Peter Thiel. This administration’s actions, together with Kratios’ report, suggest that they have been strongly influenced by Thiel, by billionaire Trump adviser Marc Andreesen (Tiku 2025), and by other venture capitalists—who apparently either don’t know or don’t care (Chitkara 2026) that their own success was made possible by the infrastructure and workforce that were built by decades of government-funded science since World War II.
So, to my European colleagues’ question: How have research universities been managing the new hostile environment in which we find ourselves? As far as I can tell, one day at a time.
Departments in my field are reducing the number of graduate students they admit, leaning harder on any non-government funding sources they may have—some philanthropies have stepped up, and while they can’t come close to replacing the government, they’ve made a meaningful difference—and hoping for better times.
But the worst probably hasn’t come yet. A typical grant lasts three years, sometimes extended by one or two years if not all the funds have been spent, so to some degree we’re all running on the fumes of pre-Trump grants, as many fewer come in. And the funding cuts are coming with other attacks—particularly new restrictions on visas for foreign students—that will make it even more difficult to do our work.
Nor is it just about students. Grant funding supports the salaries of more-senior researchers as well—particularly those on “soft money,” meaning that their continued employment depends on grant funding. The wave of firings under DOGE may have ended, but it seems almost certain that we will lose many US-based academic and government scientists in the next couple of years. They will be forced either to leave science or to leave the country, due to lack of funding. At the same time, fewer new PhDs will be trained to replace them.
I fear that much of the damage will be long-lasting. Even if the 2026 midterm elections were to turn the House and Senate over to those who want to restore the post-World War II compact between science, funding, and the federal government, the agencies will remain under the direct control of the Executive branch, and thus subject to the same kinds of actions we are already seeing. A Democratic president in 2028 would almost certainly make a bigger difference, but wouldn’t be able to fix everything immediately. Many experienced people will have left the agencies. Programs will need to be rebuilt, practices re-established, new grant competitions opened and applied for—but all with fewer staff and less institutional memory.
But most of all, the trust in the system that had been shared by everyone in it will have been fundamentally diminished. The brain drain is already underway; many institutions have already lost talented people, senior and junior, to opportunities overseas.
And now that we know how much it is possible for a hostile government to badly break the structures of US science, it will be that much harder to recruit the most talented people into our institutions, whether from home or abroad.
I want to believe that things will eventually be okay. I want to close by saying that the traditions that have made American science so successful—generous funding, awarded based on merit via peer review; academic freedom; openness to foreign talent; and most of all, systems of funding and executing science that are largely run by scientists, rather than by politicians—are resilient enough, and important enough to the nation’s well-being, that they will become the norms again.
But I don’t really know. The successes of American science in the 20th and early 21st centuries, and the resources and political consensus that enabled them, are without historical equal. Nothing lasts forever.
On the other hand, the last year-and-a-half has convinced me anew of the importance of what we do—we academics, we scientists, we climate scientists.
There have been times when I have doubted it. Working in climate science, through decades when the country has mostly taken little or no action on climate despite relentless warnings from us scientists, has at times made me question the value of our work. But if what we were doing weren’t important, this government wouldn’t be trying so hard to stop it.
Science is the practice of trying to understand the external world on its own terms as dispassionately as possible. Universities are primary sites for scholarship, not just in the sciences but also in other forms of inquiry that broaden our understanding of the world and ourselves. The goal of the authoritarian project, on the other hand, is to control citizens’ understanding of reality. Science and universities, as independent sources of understanding and authority, are its targets.
The flaws of our actual institutions notwithstanding, their existence and independence are necessary to a free society. Just doing our work, apparently, is a form of resistance.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Disclosure Statement
No potential conflict of interest was reported by the author.
Source: thebulletin.org
