NIH’s FY 2026 budget reserves half of its competing research project grant allocation for awards that fully fund every out-year as part of the initial obligation. That accounting change is the clearest signal yet that multi-year grant awards are not one thing. They are two legally different instruments wearing the same name, and only one of them puts the whole number on the table the day you sign.

The Headline Number on Your Award Letter Is Actually Two Numbers

Open a Notice of Award for a three-year, $900,000 project and you will usually find something like $300,000 in the current-year obligation field and $600,000 described as future support, anticipated commitments, or recommended future years. Those two figures have completely different legal weight, and organizations routinely treat them as one.

The federal rulebook draws the line explicitly. Under 2 CFR 200.1, the period of performance is “the time interval between the start and end date of a Federal award, which may include one or more budget periods,” and the definition adds that naming it “does not commit the Federal agency to fund the award beyond the currently approved budget period.” A budget period, by contrast, is the funded slice you are actually authorized to obligate against. When an award anticipates multiple budget periods, 2 CFR 200.211 requires the agency to state that later periods are subject to availability of funds, program authority, satisfactory performance, and compliance with the award terms.

Agencies restate this in their own guidance in blunter language. The NIH Grants Policy Statement describes out-year amounts as projections “based on the information available at the time of the initial award,” says they are not guarantees, and states that they “create no legal obligation to provide funding beyond the ending date of the current budget period.” That sentence is the whole argument. Under the incremental model, a three-year award is closer to one funded year plus two years of stated intent than to a signed three-year contract.

How to Tell Which Instrument You Actually Got

You do not need a lawyer to classify your award. You need four fields off the Notice of Award, read in this order.

Compare the budget period dates to the period of performance dates. If the budget period is twelve months and the period of performance is thirty-six, you hold an incrementally funded award with two future non-competing continuations ahead of it. If the two date ranges are identical and longer than a year, you hold a forward-funded award and the money is already obligated.

Read the obligation field, not the total. The amount recorded as obligated in the current action is the only figure backed by appropriated funds today. Everything in a future-support table is a planning number.

Find the continuation trigger. Incrementally funded awards make the next payment conditional on something you must file, typically an annual progress report submitted before the current budget period ends. Miss that filing and the mechanism that releases year two never fires, regardless of how well the project is going.

Check the carryover clause. Automatic carryover lets unspent money roll from one budget period into the next without asking. Restricted carryover requires written approval, and if the request is late or denied, the unspent balance from that period generally reverts. Most NIH awards run on automatic carryover; Department of Education awards commonly do not. Carryover terms can also change between continuation awards, so the clause you read in year one is not necessarily the clause governing year two.

Four fields, five minutes, and you know whether the second and third columns of your budget are money or weather. If you are still comparing programs before you get to this point, the structure of the award is worth checking against the federal grant programs you are shortlisting, because two grants with identical headline totals can carry very different amounts of real commitment. Filtering a grant database search by project length alone will not surface this difference; it lives in the award terms, not the listing.

NIH Is Moving Half Its Money Into the Other Column

The split between these two instruments used to be stable and boring. It is not anymore. NIH has historically funded research project grants incrementally, obligating each year’s commitment from that year’s appropriation and reserving forward funding for narrow cases such as construction, short projects, and certain Office of the Director awards.

Beginning in FY 2025 and continuing in FY 2026, NIH began reserving half of its competing research project grant allocation for awards that fully fund their out-year commitments up front. The agency’s stated reason is budget flexibility: fully funding a project at the start stops future appropriations from being encumbered by projects that started years earlier. The AAMC’s tracking of NIH forward funding notes that the FY 2027 congressional justification goes further and proposes fully funding out-year commitments for all competing research project grants.

The tradeoff shows up in award counts. Congressional Research Service analysis of NIH funding attributes a projected drop to 4,312 new competing grants under the FY 2026 request partly to this policy, which produces fewer but larger awards. Congress has been watching the same math: the FY 2026 appropriations act included a provision capping the amount NIH may obligate for multiyear awards at the FY 2025 level, with “multiyear award” defined to include grants, cooperative agreements, and contracts.

For a single applicant this is a real change in risk profile, not a bookkeeping footnote. A forward-funded award removes annual appropriations exposure and the continuation-approval step. It also concentrates competition into fewer slots and starts a five-year clock: under 31 U.S.C. 1552(a), agencies must close fixed-year appropriation accounts and cancel remaining balances by September 30 of the fifth fiscal year after the year of availability, which can limit carryover and cut off the ability to extend a final budget period on a fully funded award. Anyone tracking this shift should also watch the wider rulemaking covered in the grant policy and industry news feed, since OMB’s proposed rewrite of 2 CFR Part 200 carries a targeted October 1 effective date.

What Breaks When Year Two Arrives Light

The reason this classification matters is that organizations spend multi-year grant awards against the headline number. Three failure modes follow, and all of them are budget failures rather than program failures.

The first is payroll. A three-year award with two years of projected support invites a three-year hire, but only the first year is obligated. If the continuation lands below the commitment level, the gap does not show up as a smaller grant. It shows up as a person whose salary is no longer covered halfway through a fiscal year, and personnel is the hardest line to unwind quickly.

The second is subawards. Pass-through entities that issue three-year subawards against an incrementally funded prime have promised money the prime does not yet hold. The cleanest fix is to mirror the structure downstream: issue subawards by budget period with continuation language that matches the prime’s, so the risk sits in the same place in both documents.

The third is the continuation filing itself. Non-competing continuation awards are usually conditioned on an annual progress report, and NIH’s FY 2026 fiscal policy notice confirms that continuations are generally issued at the commitment level shown on the Notice of Award but remain subject to appropriations, satisfactory performance, and compliance. The commitment level is the ceiling, not the floor, and the report is what unlocks it. Treat that due date with the same seriousness as the original application deadline, because functionally it is one.

Rebuild the Forecast Around Obligated Dollars

The fix is a two-column budget. Column one holds obligated dollars: the current budget period obligation plus, on a forward-funded award, the entire remaining balance. Column two holds projected dollars: every out-year figure the agency has described as anticipated. Fixed commitments (salaries, leases, insurance, executed subawards) may only be underwritten from column one. Variable and deferrable costs (equipment purchases, travel, optional evaluation work) can be planned from column two and cut without breaking anything if the money moves.

Two habits keep the model honest. Re-forecast on the date each continuation award is issued rather than at the start of your fiscal year, because that is when a projection converts into an obligation or fails to. And discount column two for anything with visible political or appropriations exposure instead of carrying it at face value. An organization running several awards should know, on any given day, what share of its funded pipeline is legally obligated. For most grant-dependent nonprofits, that number is meaningfully lower than the total they report to their board. Diversifying across award structures, not just funders, is a real hedge here, and the nonprofit grants hub is a reasonable starting point for building a mix that does not depend on every out-year landing.

Frequently Asked Questions

Are multi-year grant awards legally binding for the full amount?

Usually not. Under 2 CFR 200.1 and 200.211, identifying a multi-year period of performance does not commit the agency to fund beyond the currently approved budget period, and agencies must state that later periods depend on available funds, program authority, performance, and compliance. The exception is a forward-funded or multi-year funded award, where the entire amount is obligated in one action and the budget period matches the project period.

What is the difference between a budget period and a project period?

The project period, called the period of performance in the Uniform Guidance, is the full authorized span of the award. A budget period is a funded slice inside it, usually twelve months, during which you may incur obligations against the funds awarded. A three-year award ordinarily has one period of performance and three budget periods. The distinction controls carryover, continuation filings, and what your accounting system may post against the award.

Can an agency reduce a promised out-year amount?

Yes. Out-year figures on the Notice of Award are projections contingent on appropriations, satisfactory progress, and the continued interest of the government. A reduction is not a termination and generally does not carry the same appeal rights, which is why a light continuation is often harder to contest than an outright cancellation. The practical protection is a budget that never depended on the full projection in the first place.

Does a no-cost extension add money to a multi-year award?

No. A no-cost extension moves the end date of the period of performance without adding funds. Many recipients can take one 12-month extension under expanded authority with advance notification rather than prior approval; anything beyond that, or anything involving a scope change, needs the agency’s written approval. On a forward-funded award, the five-year appropriation cancellation rule can limit whether a final period can be extended at all.

How should a small nonprofit budget against a three-year award?

Underwrite fixed costs from the obligated amount only, and place discretionary spending in the projected column. Mirror the prime award’s budget-period structure in any subawards you issue. Put the annual progress report due date on the calendar with the same weight as an application deadline, and re-forecast the moment each continuation award arrives rather than on a fixed internal schedule.

Bottom Line

Multi-year grant awards are worth pursuing, and the forward-funded model is genuinely better for recipients when you can get it: no annual appropriations exposure, no continuation-approval gate, and full spending certainty from day one. But the word “multi-year” tells you about duration, not about commitment, and the two travel separately.

The specific action worth taking this week is narrow. Pull every active award your organization holds, record two numbers for each (obligated to date and projected remaining), and total the second column. That total is your unfunded forecast. If it exceeds what you could absorb in a single year, the exposure is not theoretical, and the answer is usually a wider mix of award structures rather than more of the same instrument. When you are ready to build that mix deliberately, OpenGrants’ managed grant writing and strategy services can help structure a portfolio where no single continuation decision can take the organization down with it.