SBIR Grants for Startups: The Complete 2026 Guide

The SBIR program was born from a specific frustration: federal R&D dollars were almost entirely captured by large defense contractors and universities, with almost nothing reaching early-stage startups. The 1982 law changed the math by mandating a set-aside. Today that set-aside produces more pre-seed and seed-stage funding for technology startups than any single venture capital firm in the country.

By the numbers: in FY2022, SBIR deployed $4.4 billion across roughly 6,000 Phase I and Phase II awards. The related STTR program (which requires a research institution partner) deployed an additional $662 million. Source: SBA FY2022 SBIR/STTR Annual Report (sbir.gov) DoD is the largest single participant at approximately $2.3 billion annually. NIH is second at roughly $1.0 billion. NSF, DOE, and NASA each run between $100 million and $200 million per year.

What "non-dilutive" actually means in SBIR context

SBIR awards are grants (at NIH, NSF, DOE, NASA) or contracts (at DoD). In neither case does the agency take equity in your company. You do not give up a board seat, preferred shares, or any ownership stake. The government gets certain IP data rights during the award period and for a defined window after, but your core intellectual property remains yours. This is the defining advantage of SBIR over venture capital: you retain full ownership of the company while receiving six figures to prove your technology.

The Phase III vision and why it matters

Many founders treat Phase III as vague future aspiration. DoD evaluators treat it as a scoring criterion right now, in Phase I. When the Army or Air Force funds your Phase I, they expect a credible theory of how this technology transitions into a procurement contract or dual-use commercial product. Companies that plan Phase III from Phase I outperform those that plan it later. The most successful SBIR companies have already identified their defense customer or commercial buyer before writing their first proposal.

The role of the SBA

The Small Business Administration sets SBIR policy but does not run any individual program. It sets the statutory caps (Phase I $323,090, Phase II $2,153,927 as of October 2024 Source: SBA SBIR/STTR Policy Directive, inflation-adjusted October 2024 (sbir.gov/about/policies)), enforces the set-aside percentages, maintains the SBIR Company Registry at sbir.gov, and publishes the government-wide SBIR solicitation calendar. Each participating agency runs its own competitions under SBA policy rules. When you read conflicting dollar figures across agency websites, the SBA-set statutory cap is the authoritative ceiling. Individual agency solicitations may cite lower figures from older NOFO documents.

The PI employment rule requires that the Principal Investigator spend more than 50% of their working time at the applicant small business at the time of award and during the project period. This creates a specific challenge for academic founders: a professor who is still primarily employed by a university (typically at least 9 months per year under a faculty appointment) cannot serve as PI on an SBIR award without changing their employment status. STTR exists precisely to solve this problem: the PI can be primarily employed at the research institution, while a co-PI leads the small business side.

What counts as "primary employment"

Primary employment under SBIR means more than 50% of working time, measured over the period of performance. A PI who works 60 hours per week and devotes 35 hours to the startup qualifies. A PI who works 50 hours per week and devotes 24 hours to the startup does not. NIH checks this at award, not just at application. NSF requires at least 20 hours per week of effort at the small business, regardless of what primary employment looks like elsewhere.

The affiliate trap for corporate spinouts

SBA's affiliation rules (13 CFR Part 121) are extensive and specific. Common affiliation traps include: a parent corporation with more than 50% ownership of the applicant, a common investor with more than 50% ownership of two different applicants, and any entity with the power to control the applicant through contractual arrangements. Corporate venture studios and university spinout programs sometimes structure ownership in ways that accidentally create affiliation with the university or parent company, pushing the combined employee count above 500. Have a lawyer review your cap table and governance documents before your first SBIR application if you are a spinout.

VC-backed companies: the agency-by-agency matrix

NSF: majority VC/PE/hedge fund ownership disqualifies at the entity level. No exceptions. DoD: majority VC-owned companies may apply following the 2022 reauthorization changes, but must certify their ownership and may face additional review. NIH: follows similar permissive rules as DoD for the SBIR program specifically. A majority VC-owned company can apply to NIH SBIR but typically cannot apply to NIH STTR (STTR retains the old restriction). Check the specific NOFO you are targeting for the current policy language, as these rules were in active flux as of mid-2026.

The gap between Phase I success and Phase II success is where many good technologies die. Phase I gives you the proof that your concept works. Phase II asks a harder question: can you build something a real customer would pay for? The best Phase I applications already answer this question in the commercialization plan. Reviewers want to see a named potential customer, evidence of market need, and a credible technology development roadmap, not just a list of experiments.

NSF Phase IIA and IIB supplements

NSF offers supplemental awards on top of the base $1,250,000 Phase II. Phase IIA supplements of up to $500,000 are available for companies that have secured matching private investment (a dollar-for-dollar co-investment from a qualified investor). Phase IIB supplements of up to $500,000 are available for companies with a signed strategic partner agreement. Combined, a single NSF awardee can access up to $2,250,000 across Phase I + II + supplements, making NSF substantially more valuable than its Phase I cap suggests.

NIH Phase I to Phase II timing

NIH Phase II applications can be submitted once Phase I is complete and results have been written up. NIH generally expects companies to wait until Phase I is done, though some institutes allow Phase II application during the final months of Phase I. The Phase II application must be submitted to the same NIH institute that funded Phase I, unless you obtain prior approval to move to a different institute. The Phase II Research Strategy section requires a 6-page summary of Phase I results -- plan that documentation throughout Phase I, not just at the end.

DoD Phase III: procurement as the real exit

DoD Phase III does not require a new competitive award. Once you are a Phase II awardee, the DoD agency component can issue Phase III sole-source contracts under existing SBIR authority without a new competition. This is the most powerful feature of DoD SBIR: if you build technology an operational unit needs, the contracting path is substantially shorter than for non-SBIR companies. AFWERX's STRATFI and TACFI programs are Phase III mechanisms that have produced some of the largest single awards in SBIR history. Founders who enter DoD SBIR without a specific acquisition program in mind often win Phase I and II but find Phase III never materializes.

Choosing the right agency is the first decision. Choosing the right program office or institute within that agency is equally important and is frequently overlooked by first-time applicants.

At NIH: picking your institute matters

NIH has 27 institutes and centers (ICs), each running its own SBIR competition under the shared PHS omnibus solicitation. NCI funds cancer-related technology. NIMH funds mental health technology. NIBIB funds biomedical imaging and bioengineering. NHLBI funds cardiovascular and lung technology. Each IC has its own success rates, priority areas, and program officer contacts. Submitting to the wrong IC is a common and costly mistake. A project on an AI-powered ophthalmology diagnostic should go to NEI (National Eye Institute), not NCI -- even if the applicant's oncology advisor wrote the proposal. Email your target IC's program officer with a one-paragraph summary before applying. The PO response (or lack of one) tells you everything about fit.

At DoD: army vs. AFWERX vs. DARPA

Army SBIR has the highest topic volume, the broadest eligibility interpretation, and historically higher acceptance rates for well-written proposals. It is the most accessible entry point for defense-tech startups without existing DoD relationships. AFWERX (the Air Force innovation unit) offers Open Topics that do not require a defense-specific technology focus -- any commercially significant dual-use technology can apply. AFWERX also runs the largest Direct-to-Phase-II program. DARPA topics are fundamentally different: they are research bets on breakthrough technology concepts, not near-term product development. DARPA Phase I awards are often larger than typical, with less structured deliverables, but the acceptance rate is extremely low (5 to 10% by practitioner consensus) and successful DARPA performers often already have relationships with the program manager before the solicitation is published.

Program Officer contact: the highest-ROI pre-submission action

Every SBIR program office has accessible points of contact. At NIH, program officers are explicitly listed on each IC's website and routinely respond to pre-submission inquiries. At NSF, Program Directors take Zoom calls before Project Pitch submission and this is encouraged in the official program materials. At DoD, communication must go through the official Q&A system in DSIP (not direct email or calls), but the Q&A window is designed for this purpose. At DOE, the Science Officer contacts are published per topic. Founders who contact the right program officer before committing 150 hours to a proposal prevent at least 20% of rejection outcomes by verifying fit, scope, and framing before writing.

At NIH, your application is scored on a 1 to 9 scale (1 = best) across five criteria: Significance (does this matter?), Innovation (is this genuinely new?), Approach (is the science rigorous?), Investigators (can this team execute?), and Environment (do you have the facilities and resources?). The overall impact score is a weighted combination, with Approach typically dominating. The funding threshold varies by institute: you generally need to land in roughly the top 10 to 15 percentile to be funded, but the exact percentile cutoff changes each review cycle based on available funds.

Significance scores: the most common first-round killer

Significance is scored on whether the project addresses an important problem and whether the proposed research would advance the field. Applications that score poorly on Significance typically make one of two errors: they are too narrow (they solve a problem that few people have) or too broad (they describe the problem at a population level without explaining why existing solutions fail). The strongest Significance sections identify a specific gap in current clinical practice or scientific knowledge, quantify its scale, and explain precisely why existing approaches cannot close that gap.

NSF Broader Impacts: not a checkbox

NSF scores Broader Impacts as a separate, equal criterion alongside Intellectual Merit. Many applicants treat it as a compliance checkbox ("we will train graduate students and present at conferences"). Reviewers who read 30 proposals in a weekend recognize this pattern immediately. Strong Broader Impacts sections at NSF SBIR link commercialization outcomes directly to societal benefit: how many jobs does this technology create, what industries does it enable, which underserved populations gain access to capability they currently lack? Specificity beats ambition. "This technology will create 15 direct jobs in a high-unemployment county over 3 years" scores better than "this technology will transform the global market."

DoD Phase II transition risk: the most underweighted criterion

DoD evaluators look for Phase III transition potential in Phase I applications because they have seen too many Phase II awardees produce a prototype that no DoD program office ever adopts. The highest-performing DoD SBIR applicants have a named program office contact, a relevant program of record or technology area, and a realistic theory of how Phase III contracting would work before they submit Phase I. This is not required -- you can win Phase I without it -- but applications that demonstrate Phase III visibility score materially better on transition potential, which is an explicit evaluation criterion across most DoD components.

SBIR stands for Small Business Innovation Research. It is a federal program requiring 11 agencies to set aside a percentage of their R&D budgets for awards to small businesses. The program deployed $4.4 billion through SBIR plus $662 million through STTR in FY2022. Source: SBA FY2022 SBIR/STTR Annual Report (sbir.gov) It was reauthorized on April 13, 2026 through September 30, 2031 (S. 3971, Small Business Innovation and Economic Security Act of 2026). Source: Congress.gov, S. 3971

Phase I awards range from $200,000 (DOE) to $323,090 (NIH) depending on the agency. Phase II ranges from $1,250,000 (NSF) to $2,153,927 (NIH). No equity is taken. No cost-sharing is required. The program is designed for for-profit small businesses with 500 or fewer employees, at least 51% owned by US citizens or permanent residents, with a Principal Investigator primarily employed by the company.

Phase I is the feasibility award: 6 to 12 months, up to $323,090 at NIH or $305,000 at NSF, to prove your concept works. Phase II is the development award: 24 months, up to $2,153,927 at NIH or $1,250,000 at NSF, to build a working prototype or product. Phase II is a competitive re-application, not an automatic continuation. About 40 to 50% of Phase I awardees win Phase II.

Phase III is commercialization: no statutory cap, no SBIR funds. Financed by private investment, federal procurement contracts, or non-SBIR government grants. At DoD, Phase III sole-source contracts can be issued to Phase II awardees without a new competition, which is one of the most powerful features of the DoD SBIR program.

Choose the agency whose mission most directly aligns with your technology's end use: NIH for biomedical and health technology, NSF for deep technology in any domain without a clear agency mission home, DoD for defense or dual-use technology, DOE for energy and environmental technology, and NASA for space and aerospace systems.

Do not choose an agency based on Phase II cap size. Choose based on which program officers will believe your technology is directly relevant to their mission. A mismatch between your technology and the agency's mission is the most common rejection reason across all SBIR programs.

SBIR and STTR have identical funding amounts and the same eligibility rules for the company. The difference is STTR requires a formal research institution partner: at least 30% of the work must go to a university, federal lab, or nonprofit research organization, and at least 40% must stay with the small business.

STTR relaxes the PI employment rule: the PI can be primarily employed at the university. SBIR requires the PI to be primarily employed by the company. If your co-founder is a faculty member who cannot leave their university appointment, STTR is the right path. STTR also requires a formal IP Rights Agreement between the company and the research institution before submission -- allow 8 to 10 weeks for university technology transfer offices to process this.

The realistic timeline from starting your application to receiving your first payment is 9 to 14 months. At NIH, applications are due September 5, January 5, or April 5; awards typically start 6 to 8 months later. At NSF, the Project Pitch adds 1 to 2 months before the full proposal; total cycle is about 8 months. At DoD, the proposal-to-award timeline is 6 to 9 months from close, plus 1 to 3 months of contract negotiations.

Plan for at least one failed attempt before success. Most SBIR awardees apply 2 to 3 times before winning their first award. The realistic "idea to money in the bank" timeline for a first-time applicant is 18 to 30 months from decision to apply.

It depends on the agency and your cap table. NSF is the most restrictive: any company with majority ownership by a venture capital operating company, hedge fund, or PE firm is ineligible. DoD and NIH are more permissive under a 2022 SBA policy change that allows majority VC-owned firms to apply to select agencies.

The ownership test is at the individual level: 51% of the company must be owned by US citizens or permanent residents as individuals, not just as entities. If your cap table has significant foreign institutional investor concentration or a single investor above 50%, confirm eligibility with a lawyer before investing time in an application. The NSF restriction is particularly relevant for VC-backed deep-tech companies that would otherwise be strong NSF candidates.

Every SBIR applicant needs three registrations before any award can be made. First, SAM.gov: register to get a Unique Entity Identifier. Allow 2 to 4 weeks for new registrations, and renew annually (a lapsed registration delays awards). Second, the SBA SBIR Company Registry at sbir.gov: a separate profile capturing ownership, employee count, and prior award history. Allow 1 to 3 business days. Third, an agency-specific portal account: eRA Commons for NIH, Research.gov for NSF, DSIP at dodsbirsttr.mil for DoD, and Grants.gov for DOE.

Start all three registrations at least 4 weeks before your target deadline. A lapsed SAM registration is the single most common last-minute crisis. SAM.gov must be current (not just registered -- actively renewed) at the time of award. Check your expiration date today.

Yes. SBIR grants and the federal Section 41 R&D tax credit stack cleanly because they operate in different parts of your business. SBIR is a grant or contract covering eligible project costs. The R&D credit is a tax incentive on qualifying research wages -- including wages paid from SBIR funds.

Qualified small businesses with less than $5 million in gross receipts can apply up to $500,000 per year of the R&D credit directly against employer payroll taxes. File Form 6765 with your federal return to claim it. DoD contract awards have a different treatment under "funded research" rules (IRC 41(d)(4)) -- confirm with a CPA who has worked with both SBIR and Section 41 before claiming both for a DoD contract award.

The most common SBIR rejection reasons, in approximate order of frequency: (1) poor topic or institute alignment -- the technology does not clearly serve the agency's mission or the specific topic's stated need; (2) weak commercialization plan -- especially at NSF and NIH, where it is an explicit scored criterion; (3) insufficient proof of innovation -- the application does not demonstrate clearly why existing approaches fail; (4) PI employment issues -- the PI is not primarily employed by the company at the time of award; (5) missing or lapsed registrations discovered after the deadline; (6) budget errors or unreasonable cost estimates; and (7) failure to contact the program officer or topic author before submitting.

At NIH specifically, poor Significance scores are the most common first-round failure mode. At NSF, "the project is too applied" (meaning it is engineering or product development rather than unproven research) is the most common Project Pitch rejection. At DoD, proposal-topic misalignment is endemic. The fix for all three: contact the program officer before you write.