We’re learning a lot about how government can shape our lives by watching the second Trump Administration dismantle it. One lesson is that government’s capacity to do good runs on information no less than on funding and regulations. From weather and economic forecasts to the census to predictions of other countries’ military capabilities to vaccine monitoring, data and ideas generated inside and outside of the federal government have guided decisions in a world of profound complexity. But as the young men of Elon Musk’s DOGE figured out quickly, information is also a point of vulnerability for the entire workings of government, and it can be exploited by those like Musk and Trump who seek to disable government, concentrate its power, or redirect that power to private profit.

Dozens of small federal agencies devoted to information and ideas have been gutted; expert advisory commissions disbanded; and grants for libraries, museums, and scientific and health research cut off without review. Indicators such as the National Assessment of Educational Progress, which always had strong conservative support, have been cancelled, pared back, or delayed, often because contracts were arbitrarily canceled, advisory panels dissolved, and key staff fired.

Much of the loosely connected galaxy of information and data that guides policy falls outside the formal boundaries of government, in a pluralistic set of institutions that are independent of the administration or political parties. Along with universities, independent policy research organizations—think tanks—are key to the system of knowledge production and policy ideas, particularly in the United States. Every think tank, aside from the few that maintain an allegiance to the current Administration, now faces a test: How do they not only survive, but remain relevant when the assumptions and processes under which they were born have been wiped away? How can their capacities be put to good use at a moment when the idea of informed decision-making is itself under attack, when little matters other than the raw and often arbitrary exercise of power?

Most think tanks are nonprofits, so their immediate risks resemble those that confront higher education and other civil society organizations. Some receive federal grants for research projects and may have already received “stop-work orders” on those grants. Their tax-exempt status may be challenged, or they may be subjected to intrusive investigations by the Internal Revenue Service (IRS); they may be called before congressional subcommittees such as the one chaired by Representative Marjorie Taylor Greene; they may be sued for efforts to address racial disparities. Their funders, often large private foundations like the Ford and Gates Foundations (denounced by Vice President JD Vance as “cancers on American society”), face similar threats, which may provoke excessive caution. At an extreme, the Administration may seek to take control of independent think tanks, as it did the congressionally chartered Wilson Center and the U.S. Institute for Peace.

In response to these risks of loss of funding or autonomy, think tanks’ first obligation is to follow the principle that elite law firms and Columbia University learned too late: There’s no advantage or safety in bargaining with this Administration or conceding wrongdoing where none exists. While there’s little sense among think tanks of belonging to a sector—there is no “think tank association,” for example—they have to see a threat to one as a threat to all, even if the organization under attack represents a different ideological approach.

Beyond these shared legal and financial risks, think tanks face a deeper test. Not just their individual autonomy, but their role in a larger system of knowledge and idea production is at risk. For many institutions, their relationship with government—whether providing technical advice, conducting program evaluations, developing new policy ideas, or finding trends in public data—defines some or most of their purpose. Especially for older, technocratic organizations, the legislative and executive branches of government have been their primary constituency—for example, they might provide Congress or agencies with assessments of the likely effects of a new tax policy or regulation. That audience, or at least the majority that controls the federal government, is not interested in the analyses and recommendations of think tanks—with the exception of a handful of MAGA-identified institutions, such as the America First Policy Institute—and might not be for a while.