America's Seed Fund pays small businesses to do federal R&D without taking equity. The program expired in September 2025, went dark for six months, and was reauthorized through 2031 in April 2026 — with significant changes to how it works.
Open programs in the OpenGrants database whose listings mention sbir / sttr. Refreshed on every site build.
If you last looked at SBIR before autumn 2025, the program you remember was interrupted.
Statutory authority for SBIR and STTR expired on September 30, 2025. During the lapse, agencies could not issue new solicitations or commit new funds. It lasted about six months. On April 13, 2026, the Small Business Innovation and Economic Security Act of 2026 was signed, reauthorizing both programs through September 30, 2031.
This matters beyond trivia. Solicitation calendars were disrupted, agency pipelines backed up, and several rules changed. Advice written before late 2025 — which is most of the SBIR advice on the internet — describes a program with different limits and different compliance requirements than the one you are applying to now.
Strategic Breakthrough Awards. A new follow-on category for agencies spending over $100 million a year on SBIR, worth up to $30 million per award, structured either as a single award or a sequence of milestone-triggered payments over a maximum 48-month performance period, with matching-fund requirements. Award execution must happen within 90 days of proposal receipt — an unusually tight statutory clock.
Proposal caps from FY2027. Agencies will set their own limits on how many proposals a company may submit, per company, per solicitation, or per topic. This is the response to the long-running “SBIR mills” complaint — firms that win many awards through sheer submission volume without commercializing. If you submit a few proposals a year this changes nothing; if your model is volume, it changes everything.
Mandatory national security due diligence. Agencies must now examine foreign affiliations, where your investment capital came from, your technology licensing arrangements, and any relationships with entities in countries of concern. Companies on designated watch lists are excluded. Agencies denying an application on security grounds must provide the basis. Practically: your cap table and your research collaborations are now part of the review.
Codified technical assistance. $6,500 per Phase I project and $50,000 per Phase II project, which agencies may include in or add to the award.
Phase I buys the answer to a risk question. Can this work? Agencies may award up to $323,090 without SBA approval, though most fund considerably below that ceiling. The most common proposal failure is describing a product when the agency is buying a feasibility experiment.
Phase II buys development. Up to $2,153,927 without SBA approval, typically over 24 months, normally following a successful Phase I with the same agency. Some agencies accept Direct to Phase II proposals on designated topics if you can document that feasibility was already established.
Phase III buys nothing — and is the most valuable phase. It carries no SBIR money. It is any work deriving from your SBIR that is funded from other sources, and it comes with sole-source contracting authority: an agency can contract with you directly, without competition, on work traceable to your SBIR. Companies that understand this from the start structure their Phase I and II work differently from those that discover it at the end.
Companies wanting general operating money. SBIR funds R&D against a technical risk. It is not growth capital, and the compliance overhead is real.
Non-U.S. companies, and companies with ownership structures that will not survive the new screening. Majority U.S. ownership and control is a statutory requirement, and foreign investment ties now receive mandatory scrutiny.
Anyone who cannot survive the funding gaps. The interval between phases is unfunded and has killed otherwise healthy companies. Plan the gap before you need it.
Companies whose technical lead is a university professor — not a disqualification, but it points you to STTR rather than SBIR, since STTR permits the PI to sit at the research institution.
The most consequential early decision is which of the eleven agencies to approach, and the deciding factor is not award size but how the agency solicits.
Open-call agencies — NSF and NIH being the clearest examples — let you propose your own idea against broad program interests. If your innovation is yours and does not map to a government requirement, this is your route.
Topic-driven agencies — the Department of Defense components above all — publish specific topics describing what they want and buy against them. Submitting an off-topic idea to a topic-driven solicitation is the single most common wasted SBIR submission.
Both routes are covered in more depth in the NSF and DoD guides, which describe two genuinely different processes rather than two versions of the same one.
Generally under 500 employees including affiliates, U.S.-based, and more than 50% owned and controlled by U.S. citizens or permanent residents. Some agencies permit majority ownership by investment companies under a separate authority — check the specific agency before assuming.
NSF and NIH run open calls where you propose your own idea. DoD publishes specific topics and buys against them. Applying to a topic-driven agency with an off-topic idea is the most common wasted submission.
SAM.gov with a UEI, the SBIR company registration, and then agency-specific systems — eRA Commons for NIH, DSIP for DoD, Research.gov for NSF. This takes weeks, not days, and every one is a hard gate.
Phase I buys an answer to a technical risk question. Proposals that describe a product roadmap instead of a feasibility experiment score poorly, however exciting the product.
The interval between Phase I ending and Phase II starting has sunk many companies. Agencies run bridge mechanisms of varying quality; know your agency's before you need it.
Phase III has no SBIR money but carries sole-source contracting authority traceable to your earlier awards. For defense-adjacent companies this is where the real revenue is.
