FYI: Science Policy News
FYI
/
Article

NSF Graduate Research Fellowships Expected to Fall Next Year

AUG 28, 2026
The program solicitation estimates a sharp decline in awards for 2027 despite recent signals of support from the White House.
AIP_Clare_Zhang_800x1000.jpg
Science Policy Reporter, FYI FYI
NSF VA HQ May 2026

The front entrance of the new NSF headquarters in Alexandria, Virginia.

NSF

The National Science Foundation plans to cut back its prestigious graduate research fellowship awards next year, despite recent White House praise for the program. The agency expects to award 1,500 fellowships, which provide three years of support for outstanding STEM graduate students, in a major downturn after an all-time high of nearly 2,600 awards in 2026.

The agency solicitation adds to the confusing saga of the NSF Graduate Research Fellowship Program in the second Trump administration. Awardees plummeted in 2025, then rebounded to record numbers in 2026. The White House meanwhile requested deep cuts to the program for fiscal years 2026 and 2027, but last month proposed expanding the program in the Office of Science and Technology Policy’s report on the future of the US research enterprise.

The July report praised the GRFP as an example of “bet[ting] on people, not just projects.” However, it also said that “more work remains to improve the selection mechanism and further empower students to choose their universities and principal investigators.”

The solicitation also newly excludes research on STEM Education and Learning or in the Social, Behavioral, and Economic Sciences Directorate from fellowship eligibility, adding to eligibility restrictions the agency imposed last year that took applicants by surprise.

NSF issued the solicitation on Aug. 25, leaving applicants with fewer than 60 days to submit their proposals — 30 days short of NSF’s self-imposed minimum.

The 1,500 estimated awards are consistent with the president’s budget request for fiscal year 2027, and the final number of awards will be subject to congressional appropriations, an NSF spokesperson said.

“I think 1,500 awards would be terrible,” former GRFP Program Officer Susan Brennan said. “But I think, basically, they have to go by the president’s budget at this stage. So, I think the best thing the community can and should do is to advocate to have 2,500 awards at least.”

NSF needs leadership “that will maintain a set of standards and not be reactive to whatever political wind is blowing,” Brennan added. NSF has lacked a director since April 2025.

In fiscal year 2026, Trump only requested enough money for around 800 GRFP awards. But the Senate Appropriations Committee’s bill came out before NSF’s solicitation that year, proposing flat funding to support 2,000 fellowships. NSF ended up exceeding even the Senate committee’s target, making nearly 2,600 awards by April 2026.

This time around, Senate appropriators have not yet released their fiscal year 2027 funding proposal for NSF. The House Appropriations Committee approved its bill in May but did not specify an amount for GRFP. Both chambers will soon return from their August recess, the Senate on Sept. 14 and the House on Aug. 31.

NSF Graduate Awards chart depicting past up-and-down patterns

A chart depicting GRFP awards since 2017.

AIP

Concerns over the number of GRFP awards spiked in 2025 when the agency had planned for 2,300 fellowships but selected just 1,000, followed by another 500 two months later. Before the additional awards were announced, the agency’s governing board announced a pilot program to co-fund additional GRFP awards with industry partners. The board did not respond to subsequent inquiries about the pilot program, and Trump dismissed all the board members in April without appointing replacements. NSF recently announced a separate pilot program to co-fund four-year PhDs with universities and industry.

Even if the number of awards ends up being higher than 1,500 in 2027, seeing that estimate might discourage prospective applicants, said former GRFP Program Officer Gisèle Muller-Parker. “That’s going to depress the diversity of applicants and the fields that they come from,” she added.

Removing STEM Education

The latest solicitation further tightens program requirements by excluding STEM Education and Learning as an eligible field of study. Last year, NSF began excluding second-year PhD students from applying, with no advance warning. It also eliminated clinical psychology students from consideration. The agency maintained those changes for this year.

Typically, fewer than 1% of GRFP awards fall under STEM Education and Learning, Brennan said. However, “1% of 2,500 is still a number that could result in interesting research and change, and to remove that kind of seed corn opportunity could well be significant,” she added. “I think if a student, let’s say from physics, has an idea for how to introduce quantum in junior high, they could end up having more impact than, let’s say, someone doing traditional chemistry as a GRFP fellow.”

STEM education research has already taken a huge hit at NSF and appears poised for further cuts. Nearly half of the more than 1,600 research grants NSF terminated last year were studying STEM education. This year, NSF awards overall and within STEM education are both down about 40% compared to the average. And though the agency announced broad funding solicitations across its directorates last week, none of them are for STEM education.

The NSF spokesperson said the agency routinely reviews GRFP fields of study and makes updates based on changes in the scientific landscape and agency priorities. In response to a question on whether NSF would continue to fund STEM education research broadly, the spokesperson said only that “NSF remains committed to funding meritorious research and development projects that advance scientific discovery and innovation while fostering our nation’s future STEM workforce as outlined in recent White House Office of Science and Technology Policy guidance.”

GRFP applicants could apply under other fields of study while still including educational components in their proposal, Brennan said. “But here’s the problem: reviewers from mechanical engineering, physics, chemistry — they’re equipped to judge research projects in their fields. They’re not equipped to judge education research, and most of them are somewhat biased against it,” she added.

The solicitation also removes the Social, Behavioral, and Economic Sciences Directorate from its list of participating directorates. Some fields within SBE are still eligible, including psychology and social sciences, “but there’s no directorate home for them anymore,” Brennan said. Trump’s fiscal year 2027 budget request proposes eliminating the SBE Directorate.

Beyond eligibility rules, Brennan and Muller-Parker say the agency has shown bias toward certain fields in its recent GRFP awardees. The 2025 awards heavily favored applications related to quantum information science, artificial intelligence, and machine learning, according to their analysis. Then in 2026, dozens of applications, including many in the life sciences, were returned without review.

The eligibility changes from this year and last came without warning from NSF. Muller-Parker said the agency needs a communication plan for the GRFP that includes calls for input on proposed changes. “If they had put out a call, they would have received answers which would inform the policy decision,” she added. “Instead, what happens? Everyone is waiting for the solicitation to come out, and then everyone is shocked when there are these changes, and we don’t know how they came about.”

Related Topics
More from FYI
FYI
/
Article
The Senate’s short-term spending bill would temporarily stop OMB from finalizing its grantmaking rule, if the House adopts it.
FYI
/
Article
The commission adopted the rule without addressing concerns about proposed satellites’ impacts on scientific observations.
FYI
/
Article
The exodus of non-US citizens has “certainly” been influenced by science funding and visa policy changes under Trump, a new AIP report says.
FYI
/
Article
Some of the most important decision-makers in science policy are facing voters in primaries and general elections this year.
/
Article
The newest spacecraft from NASA started as a spy telescope. It’s just one example of the intelligence world offering resources to astronomers.
/
Article
Although atoms in an optical lattice typically have only local interactions, periodically shaking them can effectively induce random long-range couplings.
/
Article
With quantum state tomography, researchers in the booming field of quantum information science are able to characterize essential operations and devices.

Related Organizations