The Federal Grant Contraction Stopped. The Recomposition Didn't: FY2026 Closed at 2,156 Notices With DOJ Up 132% and HHS Down 33%

Fiscal year 2026 closed on September 30, and the federal grant notice count is in. Agencies posted 2,156 notices on Grants.gov over the year. That is up about 5 percent from FY2025's 2,044 — and still down roughly 51 percent from FY2024's 4,386.

Two separate stories live inside those three numbers, and most of the commentary has been telling only the first one.

The first story is that the contraction has bottomed. For most of the last eighteen months the operative narrative in grant-seeking organizations has been decline — posted opportunities fell by more than half year over year, windows compressed from months to weeks, and canceled competitions had to be rebuilt from scratch. That decline is real and the volume loss is permanent so far. But FY2026 did not continue falling. It ticked up 5 percent, and September 2026 alone posted 151 notices against September 2025's 98 — a 54 percent year-over-year increase heading into the fiscal transition. As of October 13, 2026, there were 863 notices open, with 174 scheduled to close within 30 days.

The second story is the one with strategic consequences: the federal grant market did not shrink uniformly and then stabilize. It shrank, and then it was rebuilt with a different composition.

The agency table is the whole argument

Read that table as a reallocation rather than a set of independent budget outcomes. HHS shed 231 notices. State, Justice, and Defense together added 306. The federal government did not post meaningfully fewer notices in FY2026 than FY2025 — it posted them from different buildings.

HHS at 472 is the number that will reshape the most organizational pipelines. For a generation, HHS has been the default anchor of the nonprofit and academic federal pipeline: NIH, HRSA, ACF, SAMHSA, CDC, AHRQ. A 33 percent one-year decline in notice count from the largest grantmaking department in the government is not a rounding error in anyone's forecast. It compounds effects already visible in the research ecosystem — NIH's multi-year award squeeze and AHRQ's effective halt in grantmaking have both been observable at the award level — and it means an organization whose pipeline was 70 percent HHS in FY2024 is now fishing in a pond with a third fewer visible doors.

DOJ at +132 percent is the fastest-moving line in federal grantmaking. The department more than doubled its notice count, from 124 to 288. That expansion has been visible all year in the specific competitions it produced: JAG local formula awards, multi-solicitation BJA releases covering reentry and behavioral health, houses-of-worship security, and COPS hiring. For organizations working in behavioral health, reentry, victim services, youth programming, or community violence intervention, DOJ is now a larger source of posted opportunities than it has been in over a decade — and critically, many of those programs fund work that HHS used to fund under different framing.

State at +43 percent reflects a foreign-assistance apparatus being rebuilt through a different channel. With USAID dissolved and its functions redistributed, notice volume migrated to State. The organizations best positioned are those already registered and audited for international work; the notices are there, but the compliance bar is a State Department bar.

Defense at +18 percent is the quiet, consistent grower. DoD research, instrumentation, and small-business lines have expanded steadily, from SBIR/STTR releases running on a tightened FY27 calendar to university instrumentation programs. The growth is less dramatic than DOJ's but it comes with a structural advantage: DoD opportunity calendars are published further in advance than almost any other department's.

What a 51 percent smaller market with a rearranged mix actually does to you

The temptation is to read "notices stabilized, up 5 percent" as permission to resume a FY2024 pipeline strategy. That reading is wrong in three specific ways.

First, a stable count at half the historical level means the per-notice stakes doubled. The same organizations are chasing roughly half the doors. Competition per posted opportunity is structurally higher than it was two years ago, regardless of whether the count is flat or up 5 percent. An organization that won one in four applications in FY2024 and now wins one in six has not gotten worse at grant writing.

Second, 174 of 863 open notices closing within 30 days is a 20 percent monthly churn rate. That is the operational number. One fifth of everything currently open will be gone within a month. Compressed windows were the defining feature of the contraction and they did not relax when the count stabilized. A pipeline process built on quarterly review cycles cannot catch a 20 percent monthly churn. Weekly monitoring is now the floor, not a best practice.

Third, the recomposition means your historical hit rate is no longer predictive. Past performance in an agency's competitions is the single best predictor of future performance — reviewers, scoring rubrics, and program officer relationships all persist. An organization with twelve years of HRSA history has a genuine advantage in HRSA competitions and essentially none in a first DOJ OJP application, where the portal is different (JustGrants, not just Grants.gov), the two-deadline structure catches first-timers, and the compliance certifications are unfamiliar. Pipeline diversification across agencies is not free. It costs a learning cycle, and the cycle has to be paid before the funding arrives.

The forecast-versus-posted gap is still the main trap

One pattern from the contraction persists and deserves restating, because it generates more wasted effort than any other dynamic in the current environment: forecasted opportunities have held up far better than posted ones. A forecast entry is a statement of intent. It carries an estimated post date, an estimated award ceiling, and no obligation. Through the contraction, forecasts rose while posted notices collapsed, which meant organizations building plans off forecast records were planning against money that had not yet been committed to a competition.

With HHS down 33 percent, this trap is most dangerous in exactly the portfolio where organizations are most anxious. A forecast for an HHS program that posted reliably every year through FY2024 is not evidence that it will post in FY2027. Treat forecast records as watch items and posted notices as pipeline items, and never let a forecast occupy a staffing slot that a posted notice could use.

Two calendar notes for this week

Grants.gov is offline October 17 through October 19. The production system went down Saturday, October 17, 2026 at 12:01 AM ET and returns Monday, October 19, 2026 at 6:00 AM ET. This is the routine third-weekend-of-the-month maintenance window, not an incident. But with 174 notices closing within 30 days and a 20 percent monthly churn rate, a 54-hour outage in the middle of October is a real scheduling fact. Anything due Monday the 19th or Tuesday the 20th should have been submitted Friday. If you are holding a submission for a weekend push, you cannot push this weekend.

The two-system year is still in effect. Grants.gov and Simpler.Grants.gov are running in parallel, with a substantial share of competitions discoverable on one and submitted through the other. Searching only the legacy interface will under-count what is actually open — which, in a market with half as many doors, is an expensive mistake.

How to rebuild a pipeline for the recomposed market

Audit your pipeline by agency, not by program area. Write down the percentage of your last three years' applications that went to each department. If HHS is above 50 percent, you are over-indexed on the one major department that contracted, and the fix is a deliberate agency diversification plan with a named target.

Pick one growth-agency beachhead and pay the learning cost on purpose. Do not scatter across DOJ, State, and DoD simultaneously. Pick the one whose mission genuinely overlaps your work, read three recent solicitations in full, register in the required portal, and apply once with the expectation that the first attempt is tuition. DOJ's +132 percent expansion is the largest single opening, and its OJP programs touch behavioral health, youth services, victim services, and community safety — portfolios that many HHS-oriented organizations already run.

Reframe existing programming rather than inventing new programming. The recomposition moved the funding frame, not the underlying need. A reentry behavioral health program is fundable at DOJ under a public-safety frame and was fundable at SAMHSA under a treatment frame. The work is the same. The theory of change, the outcome metrics, and the partner list in the narrative are what change.

Build your calendar around the 174 number. Roughly a fifth of open notices turn over monthly. Set a standing weekly review, maintain a written go/no-go rule so marginal opportunities get declined fast, and keep a reusable core package — organizational capacity narrative, audited financials, logic model, key personnel bios — so that a three-week window is workable.

Stop waiting for FY2024 volume to return. Three fiscal years of data now point the same way: 4,386, then 2,044, then 2,156. The market has found a level, and that level is about half of what it was. The 5 percent uptick is a floor, not a recovery. Plan against 2,000-ish notices a year with a different agency mix and a 20 percent monthly churn, because that is the market that exists.

The contraction story is over. The organizations that struggle in FY2027 will not be the ones that failed to notice the decline — almost everyone noticed. They will be the ones that waited for the old composition to come back.