USGS Just Cut Its 104(g) Water Grant Ceiling By Two-Thirds And Pointed The Whole Competition At Data Centers

Academic PIs accustomed to NIH's modular budgets have until 5 p.m. Eastern on November 12, 2026 to compete for one of four $120,000 awards under the U.S. Geological Survey's Water Resources Research Act 104(g) national competition, posted as announcement G27AS00164 on grants.gov.

A $462,500 Pool, Four Awards, And A Ceiling That Fell By Two-Thirds

The FY2026 solicitation — formally titled "Water Resources Research Act Program National Competitive Grants FY2026 (104g General)" and posted on grants.gov as opportunity 364012 under announcement number G27AS00164 — carries an estimated total program funding of $462,500, an expected four awards, and a per-award ceiling of $120,000. Divide the pool by the award count and you get $115,625. That arithmetic matters: USGS is signaling it intends to fund at or near the ceiling rather than spread the money across a dozen seed-sized projects. Four PIs will be funded. Everyone else writes for free.

The ceiling is the number that should stop you. The prior-year general competition, announcement G25AS00485, ran with a $310,000 award ceiling — and at least one state institute advertised project budgets capped at $300,700 after its administrative coordination fee came off the top. FY2026's $120,000 is roughly a 61% reduction in what a single 104(g) general project can command. The program did not get cheaper to apply to. It got cheaper to win, which is a different and worse thing.

For a researcher calibrating against NIH norms, the scale is jarring. A $120,000 ceiling spread across a project period of up to three years is less total direct-and-indirect support than a single modular R01 year. This is not a program that funds a lab. It funds a focused question, a graduate student, and a defensible dataset.

Your Real Deadline Is Your State Institute's, And In Arizona It Has Already Passed

Here is the structural feature that eliminates more otherwise-fundable PIs than peer review does: you cannot submit to 104(g) yourself.

Under the Water Resources Research Act, proposals must be filed through a state or territorial Water Resources Research Institute by an investigator at a U.S. institution of higher learning. Your institute is the applicant of record. You are a subrecipient of a process you do not control. That means every institute sets its own internal deadline well ahead of the federal one, and those internal deadlines are the ones that actually bind you.

The spread is wide and it is live right now. Arizona's Water Resources Research Center set its cutoff at 5 p.m. MST on October 27, 2026, and opened eligibility to faculty and research personnel at any of the state's three public universities. The Massachusetts Water Resources Research Center at UMass Amherst set November 1, 2026 at 11:59 p.m. ET, explicitly to give staff time "to review, coordinate, and prepare proposals for submission to USGS by the federal deadline." Massachusetts also requires that your budget carry funds for WRRC staff coordination if the project is funded — a line item that comes out of your $120,000, not in addition to it.

So the honest deadline calendar for a PI reading this on October 4 is somewhere between three and four weeks, not six. If you have not already emailed your state institute director, that is the single highest-value action available to you today, ahead of writing a word of science. Institutes also vary in how many proposals they will forward, and a director who learns about your project on October 25 has no reason to prioritize it over one they helped shape in September.

Why USGS Is Buying Data-Center Water Research This Year

The FY2026 general track is not a blank topical canvas. It is scoped — and the scope is the most newsworthy thing in the solicitation. The competition runs under the heading "104G General — Water Use and Data Centers," and it targets water use associated with data centers and other water-intensive industries. The program language asks for work that improves methodologies capable of more reliably characterizing industrial water withdrawals, cooling demands, and local impacts.

Compare that to the prior cycle, which solicited research on the economic value of the USGS streamgage network and on advancing hydrologic models using AI and machine learning. In one year the program moved from "help us value and model our own data infrastructure" to "tell us how much water the AI buildout is actually consuming, and where."

That pivot is a direct response to a measurement gap. Industrial self-supply withdrawals are among the weakest categories in national water-use accounting; cooling demand at hyperscale facilities is frequently proprietary, estimated from nameplate capacity, or reported at annual aggregate scale that is useless for assessing a stressed local aquifer. A program that funds four projects at $120,000 cannot resolve that gap. It can fund the methods papers that the next generation of national water-use estimates will cite.

The practical read for PIs: a proposal that treats data centers as a motivating anecdote and then does generic water-quality work will score badly. The reviewers are buying methodology for industrial withdrawal and cooling characterization. Bring a measurement approach, a validation strategy, and a plausible path to transferability beyond one facility.

The 1:1 Match Is Where NIH-Trained PIs Lose This Competition

104(g) requires a 1:1 match of non-federal to federal dollars, typically demonstrated across the project period. This is the requirement that most reliably surprises investigators whose funding instincts were formed at NIH, where cost sharing is generally neither required nor, on most mechanisms, even permitted to influence review.

Read the requirement literally: a $120,000 federal award obligates you to document $120,000 in non-federal contribution. Not federal pass-through. Not another agency's grant. Institutional funds, state appropriations, foundation dollars, cost-shared faculty or student effort, documented in-kind, and committed utility or industry partner contributions are the usual sources. Your sponsored programs office knows which of these your institution will actually certify, and they will need lead time to do it — which collapses back into the internal-deadline problem above.

This is also the point where a well-chosen partner stops being a nicety. A data-center operator, municipal water utility, or state water agency that can commit documented in-kind staff time, metering access, or facility data does two things at once: it closes your match and it makes the methodology credible. USGS encourages substantial collaboration between the agency and applicants, and collaborative proposals are expected to describe each party's role in detail. A letter that vaguely endorses your work is not a role description.

Match requirements reward PIs who treat budget architecture as part of the science case rather than an administrative afterthought — the same discipline that separates winning from losing applications at private research funders. The William T. Grant Foundation's research grants program rewards a closely related instinct: proving an intervention works rather than merely studying it. Here, the analogous move is proving you can measure something the federal government currently cannot.

The PFAS Track Is Missing And The Invasive Species Track Is Gone

104(g) has historically run as a family of national competitions: a general track, a PFAS water-quality track, and an aquatic invasive species track. FY2026 has thinned out. As of early October 2026, the PFAS notice of funding opportunity had not yet been posted, and institutes are reporting that no aquatic invasive species competition is available this cycle.

That changes the calculus in two directions. If your work is PFAS-adjacent, resist the urge to retrofit it into the data-center general track — a scoped solicitation is a bad home for an off-topic proposal, and a PFAS NOFO may still post. If your work is invasive-species-adjacent, there is no 104(g) home for it in FY2026 at all, and the realistic alternatives are the state institute 104(b) base-grant competitions your institute runs with its own annual allocation.

It also means the general track is likely to absorb displaced applicants. Four awards against a field swollen by researchers who would ordinarily have filed elsewhere is a worse ratio than the raw numbers suggest. Precision beats ambition here: a tightly scoped single-question methods proposal that lands squarely on industrial withdrawal measurement will beat a sprawling three-aim program that reads like a repurposed R01.

Email Your Institute Director Before You Draft Anything

Email your state Water Resources Research Institute director today — before drafting — and ask two questions: what is our internal deadline, and what non-federal match will this institution certify? Those answers determine whether a G27AS00164 proposal is feasible for you at all. Then pressure-test the rest of your water-research pipeline: search active water resources research funding on Granted to see which 104(b) state competitions, NSF hydrology programs, and agency water-quality solicitations are open alongside this one, so a near-miss on four national awards is not the end of your fiscal year.