Defense does not fund your idea. It publishes topics describing what it needs and buys against them. Getting that backwards is the single most common wasted submission in the whole SBIR program.
The most useful sentence about DoD SBIR is that the Department is a customer, not a philanthropist.
It publishes topics — specific descriptions of capabilities it needs — and buys work against them. You do not bring your idea and ask for support. You find the topic your capability genuinely answers, and you propose to answer it.
This is the opposite of the NSF model, where no topics exist and you propose whatever you like. Companies that treat the two as interchangeable produce the single most common wasted submission in the SBIR program: a well-written proposal describing an excellent technology that does not answer the topic it was submitted against. Reviewers see it constantly, and it scores badly every time.
Most DoD BAA cycles open with a pre-release period in which you can contact the topic authors directly and ask what they actually mean. Once the BAA opens for submission, that communication closes.
First-time applicants routinely discover the cycle during the open period, after the window has shut, and write their proposal by guessing at intent. Experienced ones plan their year around pre-release dates.
If you take one operational thing from this page: find the pre-release dates for the components you care about, put them in a calendar, and use them. The difference in proposal quality between a company that asked the topic author what problem they were really trying to solve and one that did not is visible in the score.
Proposals go through the Defense SBIR/STTR Innovation Portal. Submissions by any other route are disregarded — there is no email fallback and no exception.
DSIP sits on top of the general federal registration stack, and all of it takes time:
None of this is difficult; all of it is slow, and all of it is a hard gate. Starting in the week before a BAA closes is how companies miss cycles they were otherwise ready for.
Proposals are assembled in volumes — cover sheet, technical volume, cost volume, the Company Commercialization Report, and supporting documents. The CCR discloses your prior SBIR awards and what came of them. It is worth knowing that this record is visible: a company with a long award history and thin commercialization looks exactly like the pattern the 2026 reauthorization’s new proposal caps were written to address.
On topics explicitly designated DP2, you can propose for Phase II without holding a prior DoD Phase I — but only if you can document that the feasibility described in the topic’s Phase I section has already been established, through work funded from any source.
This is a genuine shortcut for a company arriving with relevant prior R&D, and a wasted cycle for a company hoping the reviewers will accept enthusiasm in place of evidence. The documentation requirement is the whole test.
Companies without a defense-relevant capability. The topics define the space. If nothing you do answers a topic, there is no route in, and waiting for a topic that fits is a legitimate strategy.
Companies that cannot survive a funding gap. The interval between phases is unfunded, and defense contracting timelines are not fast.
Companies with foreign ownership, foreign investment, or foreign research ties they have not examined. The 2026 reauthorization made national security due diligence mandatory across the program, covering foreign affiliations, sources of investment capital, licensing arrangements and relationships with entities in countries of concern. In defense this scrutiny was always heaviest; it is now statutory.
DSIP now brands itself under “Department of War.” Executive Order 14347, signed September 5, 2025, authorized that as a secondary title for non-statutory communications; the legal name remains the Department of Defense, because only Congress can change a statutory title. Nothing about the SBIR process changed with the branding — but if you are searching for current guidance, you will now find it under both names, which is worth knowing before you conclude a page is out of date.
DoD publishes what it wants. Your job is to find the topic your capability genuinely answers — not to reshape a topic around the product you already have. Reviewers see the latter constantly.
Most DoD BAA cycles include a pre-release window where you can question topic authors directly. After that window closes, communication is cut off. This window is the highest-leverage part of the entire process and most first-time applicants miss it.
The Defense SBIR/STTR Innovation Portal is the only accepted route — proposals submitted by any other means are disregarded. First-time registration takes time, and it sits on top of SAM.gov and the SBA Company Registry.
Submissions are structured in volumes — cover sheet, technical volume, cost volume, the Company Commercialization Report, and supporting documents. The CCR is a track record disclosure and is scored; companies with prior awards and no commercialization are visibly so.
DP2 skips Phase I but requires documented evidence that feasibility is already established. It is a genuine shortcut for companies with prior relevant work and a waste of a cycle for anyone else.
The sole-source authority attached to Phase III is the commercially valuable output. Structuring Phase I and II work with a transition path to a program of record is what separates companies that build a defense business from companies that collect awards.
