Two months ago, the Trump administration unveiled its so-called “skinny budget,” which envisioned cutting funds for the National Institutes of Health by 18 percent, or $5.8 billion. Scientists were appalled. As my colleague Adrienne LaFrance reported, one doctor said that the cuts “would set off a lost generation in American science.”

The bulked-up version of the President’s budget for fiscal year 2018, which will be released next week, may not allay those fears. According to two sources within the NIH who were briefed on the issue, the administration may pursue a new strategy in its quest for cuts, by proposing a 10 percent cap on the NIH’s indirect costs—the money it gives to grantees to support administration, equipment, libraries, IT, lighting, heating, electricity, and other overhead.

“It’s going to make every single university president across the country call their representative,” says one of the sources, who agreed to speak on condition of anonymity.

The White House denied that a final decision on the cap had been made. “The budget process is a complex one with many moving parts,” a spokesperson wrote in a statement to The Atlantic. “It would be premature for us to comment—or anyone to report—on any aspect of this ever-changing, internal discussion before the publication of the document.”

It’s not surprising that the administration is considering a cap. In the wake of the skinny budget, Secretary of Health and Human Services Tom Price defended the cuts by arguing that indirect costs represented “inefficiencies”—money going towards “something other than the research that’s being done.” As Science reported in March, the NIH doled out $6.4 billion in indirect costs in fiscal year 2016, which was 38 percent of the $16.9 billion it spent directly on research. If the 10 percent ceiling had been installed, indirect costs would have been capped at roughly $1.7 billion, representing a saving of $4.7 billion.

On the face of it, that sounds like a positive thing.  Indirect costs have long been controversial, with critics arguing that they incentivize institutions to spend money on wasteful administration and lavish purchases instead of actual research. Universities negotiate indirect costs with the HHS  when they are awarded NIH grants, and while administrative costs are capped at 26 percent, others are not. In 2014, an investigation by Nature revealed large disparities in indirect costs, with some universities receiving rates of just 20 percent while others got windfalls of 85 percent.

But defenders note that these costs are not optional, and even the generous amounts received from the NIH aren’t enough to cover the price of running a research lab. While universities could stand to streamline their spending, instituting a 10 percent cap is like using a sledgehammer to attack a problem that demands a scalpel. “We could find efficiencies,” adds Mary Woolley, president of Research!America, a non-profit that advocates for biomedical research, “but you’re not looking for efficiencies when you say we’ll cut off three or four of your appendages.”

“Even if you wanted to do this, you don’t do it in one year,” says Harold Varmus, a Nobel laureate and former NIH director. “It would be a tremendous blow for many of our research institutions and ignores the real cost of doing research. If you really want to disembowel a source of learning and ingenuity in America, this is what you do.”

“Instead of having an informed process where they get people together and talk about how to build efficiencies, they’re just backing into the numbers that the President put forward,” says one of the sources at the NIH. “They’re trying to pretend that they’ll cut something without actually cutting science.” The source revealed that officers at NIH were briefed on the upcoming proposal by a senior official earlier this week. The official did not reveal details about the overall NIH budget, but implied that the 10 percent cap would result in a similar-sized cut to what was proposed in the skinny budget.

The administration will likely argue that it is following the lead of private funding agencies like the Bill and Melinda Gates Foundation, which already impose a 10 percent limit on indirect costs for grants awarded to U.S. universities. But Woolley says, “There is absolutely no question that those grants require sacrifice.” They force universities to find other ways of paying whatever indirect costs are not covered. Indeed, it’s only possible to accept such grants because the uncapped support from the NIH buoys up so much of America’s scientific infrastructure. If that support was to be capped, lights would literally have to go off.

The proposed cap is “unlikely to go anywhere,” Woolley adds. “All presidents may propose what they will, but it’s Congress who decides.” Figures compiled by Research!America show that congressional appropriators rarely take the President’s suggestion at face value, regardless of which party is in power. Indeed, the current Congress recently decided to completely ignore Trump’s suggestion to slice $1.2 billion for fiscal year 2017. Instead, they awarded the agency an extra $2 billion—the second such annual increase in a row, and a continuing sign of the strong bipartisan support that biomedical research has long enjoyed.