SBIR (Small Business Innovation Research) is a $4B+ annual federal program that awards non-dilutive grants -- funding that doesn't require giving up equity -- to small businesses for R&D. Eleven federal agencies set aside a portion of their research budgets to fund early-stage innovation at companies with fewer than 500 employees. The money doesn't dilute your ownership, doesn't need to be repaid, and the award itself becomes a credibility signal when you're raising your next round.
If you're a startup founder with a novel technology and real R&D to do, SBIR is probably the single best source of non-dilutive capital available to you. This guide explains how the program works, what each agency pays, who qualifies, and how to decide whether it's worth your time.
How SBIR works
Congress requires 11 federal agencies to set aside a portion of their R&D budgets for small businesses. That's SBIR. The agencies publish solicitations, you submit a proposal, reviewers evaluate it, and if you win, you get a grant to do the R&D you proposed.
Here's the four-step cycle:
- Agencies publish solicitations describing technology areas they want to fund
- You submit a proposal explaining what you'd build, how, and why it matters
- Reviewers evaluate on technical merit, commercial potential, and team qualifications
- If you win, you get a non-dilutive grant to execute the R&D
The program has three phases:
| Phase | Purpose | Typical Funding | Duration |
|---|---|---|---|
| Phase I | Feasibility study | $250K-$315K | 6-12 months |
| Phase II | Full R&D | $1M-$1.7M | 24 months |
| Phase III | Commercialization | No ceiling (sole-source contract) | Varies |
Phase III is the part most founders miss. It's not a grant -- it's a sole-source government contract. The agency can buy directly from you without a competitive bid. No funding ceiling. This is where SBIR companies go from research grants to actual government revenue.
The 11 SBIR agencies and what each pays
Not all agencies are created equal. They have different budgets, review processes, timelines, and technology interests.
| Agency | Phase I Award | Phase II Award | What They Fund | Typical Timeline |
|---|---|---|---|---|
| NSF | $305,000 | $1,000,000 | Broad science and tech (AI, biotech, materials, climate) | 4-6 months |
| NIH | Up to $314,363 | Up to $2,095,748 | Biomedical, health tech, diagnostics, therapeutics | 6-9 months |
| DoD (AFWERX) | $50K-$250K | $750K-$1.7M | Air Force dual-use technology | 2-4 months |
| DoD (Navy) | Up to $250,000 | Up to $1,700,000 | Naval technology, undersea, cybersecurity | 4-6 months |
| DoD (Army) | Up to $250,000 | Up to $1,700,000 | Ground systems, communications, soldier tech | 4-6 months |
| DOE | $250,000 | $1,600,000 | Energy, nuclear, grid, climate tech | 6-9 months |
| NASA | $150,000 | $750,000 | Space, aeronautics, earth science | 4-6 months |
| USDA | $200,000 | $650,000 | Agriculture, food safety, rural technology | 6-9 months |
| DHS | $250,000 | $1,000,000 | Homeland security, cybersecurity, border tech | 4-6 months |
| ED | $250,000 | $1,000,000 | Education technology | 6-9 months |
DARPA operates differently from the standard SBIR program. Instead of fixed Phase I awards, DARPA uses Broad Agency Announcements (BAAs) with open-ended solicitations and variable funding. If your technology is genuinely breakthrough, DARPA's worth exploring -- but the application process is distinct. See our DARPA BAA guide for the full breakdown.
A note on amounts: The SBA (Small Business Administration) sets maximum award guidelines ($314,363 for Phase I, $2,095,748 for Phase II as of 2025), but each agency interprets them differently. NSF has a fixed $305K Phase I. NIH goes up to the full SBA cap. AFWERX starts with smaller Phase I awards but has a unique Phase I Option structure that adds more funding.
The three biggest SBIR agencies by volume are DoD ($2.3B annually), NIH ($1.2B), and NSF (~$174M). If you're a tech startup, one of those three is almost certainly where you should start.
A note on timing (March 2026): SBIR's authorization lapsed in September 2025. Congress has passed reauthorization legislation (extending the program through 2031 with new "Strategic Breakthrough Awards" and stricter foreign-affiliation screening), but it's awaiting presidential signature. Some agencies, including NSF, have temporarily paused new solicitations during this period. Check sbir.gov for the latest before starting an application.
Are you eligible?
SBIR eligibility comes down to four requirements:
- For-profit US small business -- incorporated in the US with a physical US presence. LLCs, C-corps, and S-corps all qualify.
- Fewer than 500 employees -- including all affiliates. If a parent company or majority investor has other portfolio companies, their employees may count toward your total.
- 51%+ US ownership -- US citizens or permanent residents must own a majority. If VCs or foreign entities own more than 49%, you may not qualify (though exceptions exist for SBA-registered SBIR participating investors).
- PI primarily employed by your company -- the Principal Investigator (PI) is the person leading the R&D work. At most startups, that's the technical co-founder or CTO. The PI must spend more than 50% of their time employed by your company during the project. For STTR (the university-partnered variant of SBIR), this requirement is relaxed.
Common eligibility questions
"We have foreign co-founders." Fine, as long as US citizens or permanent residents own 51%+ of the company. Your foreign co-founder can even be the PI if they have a green card.
"Our VC owns 52% of the company." Problem for standard SBIR. But certain VC firms are registered as "SBIR participating investors," which provides exceptions. Check the SBA's list.
"We're pre-revenue with no employees." You still qualify. Many SBIR applicants are very early. You just need to be an incorporated entity with a PI who will be employed by the company during the project.
"We're a university spinout." Eligible, as long as you're a separate for-profit entity and the university doesn't own more than 49%. Watch out for IP assignment -- reviewers will check that your company owns rights to the underlying technology.
SBIR vs. venture capital
| Factor | SBIR Grant | Venture Capital |
|---|---|---|
| Dilution | None -- you keep 100% | 15-25% per round |
| Amount | $250K-$315K (Phase I) | $500K-$5M (seed/A) |
| Timeline to money | 3-9 months from submission | 2-6 months from first meeting |
| Obligations | R&D milestones, quarterly reporting | Board seats, growth expectations |
| What you give up | Time to apply and report | Equity and some control |
| Best for | De-risking R&D, extending runway | Scaling, hiring, go-to-market |
| Credibility signal | "Federal reviewers validated our tech" | "Investors validated our market" |
They're not competing. The smartest founders use both. An SBIR award extends your runway (so you're not desperate when you negotiate term sheets), validates your technology (so investors see third-party proof), and de-risks the R&D (so investors' money goes toward growth, not basic research).
Qualcomm, iRobot, Symantec, and Genzyme all used SBIR funding early in their histories. Qualcomm's CDMA technology was developed with DoD SBIR money. iRobot's bomb-disposal robots (which led to the Roomba) were funded through military SBIR contracts. Google's PageRank research was funded by a related NSF grant (the Digital Libraries Initiative, not SBIR specifically), but it illustrates the same principle: federal R&D funding can seed transformative companies.
How to decide if SBIR is right for you
SBIR is a strong fit if you have genuine R&D to do, your technology needs 6-18 months of technical work, and you can identify a federal agency whose mission aligns with what you're building. Here's the detailed framework:
Apply if:
- You have genuine R&D to do (not just product development or feature shipping)
- Your technology has a 6-18 month runway of technical work ahead
- You have (or can get) preliminary data, a prototype, or published research
- You can identify at least one federal agency whose mission aligns with your technology
- You can afford to wait 3-9 months for a funding decision
Skip if:
- You need cash in the next 30 days
- Your product is a pure consumer app with no novel technology
- You have no R&D component (just integrating existing tools)
- You can't identify a federal use case or societal benefit
- Your company has more than 500 employees or is majority foreign-owned
The gray zone
"We're building AI, but is our model architecture novel enough?" or "We have a medtech device, but no clinical data yet." These are real judgment calls.
The best test: search sbir.gov for recent awards in your technology area. If companies similar to yours are winning, that's a strong signal. If the awards all go to universities or large defense contractors, the program might not be structured for companies at your stage.
How to apply (by agency)
The application process differs by agency. Here's the comparison:
| Agency | Steps | Deadlines | Time Investment | Best For |
|---|---|---|---|---|
| NSF | 2-step: Project Pitch, then Full Proposal (if invited) | Rolling pitches, ~3 proposal windows/year | ~4 hours for pitch, ~80 hours for proposal | First-time applicants, broad tech |
| NIH | 1-step: Full proposal to specific Institute | 3 fixed dates/year (Jan, Apr, Sep) | ~120-160 hours for full proposal | Biomedical, health tech with pilot data |
| DoD (AFWERX) | Pitch + proposal to open or specific topics | Rolling | ~40-80 hours | Dual-use military tech, fast timelines |
| DoD (Navy/Army) | Full proposal to specific topics via DSIP portal | Annual cycles (varies by branch) | ~80-120 hours | Military-specific technology |
NSF (National Science Foundation)
NSF uses a two-step process that's designed to save you time:
- Project Pitch -- a ~2-page summary of your proposed innovation. NSF reviews it in ~3 weeks and tells you whether to submit a full proposal. This is the screening step.
- Full Proposal -- if invited, you submit a detailed proposal (~15 pages). Review takes 3-4 months.
NSF is the friendliest agency for first-time applicants. They fund broadly (any science or engineering innovation), the pitch process lets you test fit without a huge time investment, and their program directors are genuinely helpful. If you're not sure where to start, start here.
NIH (National Institutes of Health)
NIH uses a single-step process with three annual deadlines (January, April, September):
- Submit your proposal to a specific NIH Institute (there are 27, each with different focus areas)
- Study section review -- a panel of scientists scores your proposal on five criteria (significance, innovation, approach, investigators, environment) on a 1-9 scale where 1 is best. Anything scored 1-3 is competitive; 4-5 is borderline; 6-9 is unlikely to fund.
- Council review -- the Institute's advisory council makes final funding decisions
NIH is the largest funder but also the most competitive. Preliminary data is critical. If you don't have pilot results -- benchmarks, a small clinical study, published findings -- you're at a serious disadvantage. NIH reviewers expect evidence, not just a good idea.
DoD (Department of Defense)
DoD SBIR runs through each military branch separately:
- AFWERX (Air Force) -- rolling deadlines, open topics, commercial-first evaluation. The most startup-friendly DoD program.
- Navy and Army -- annual solicitation cycles through DSIP (Defense SBIR/STTR Innovation Portal), the centralized DoD submission platform
- DARPA -- separate from standard SBIR, uses Broad Agency Announcements (BAAs)
DoD is where Phase III matters most. If a military branch likes your Phase I/II work, they can award you a sole-source contract to buy the product. No competition, no ceiling.
What makes a winning SBIR proposal
The five factors that most influence SBIR award decisions: (1) a clear problem with a gap in current solutions, (2) a specific, testable technical approach, (3) a credible path to commercialization, (4) a qualified team with relevant experience, and (5) alignment with the funding agency's mission. That's based on patterns across 500+ proposals at Cada Partners.
What wins:
- Clear articulation of the problem and why current solutions fall short
- Specific, testable technical approach (not hand-waving about "AI" or "machine learning")
- Credible commercialization path with named potential customers
- Team with relevant experience (not just impressive resumes -- domain experience matters)
- Preliminary data or proof of concept that de-risks the approach
What loses:
- Proposals that read like marketing materials instead of technical plans
- Vague innovation claims without evidence
- Teams with no relevant domain experience
- Proposals that don't address the specific topic or agency mission
- Poor writing, unclear structure, or missing required sections
The single biggest mistake: founders describe their product instead of the research. SBIR funds R&D. Your proposal needs to articulate what you don't know yet and how you'll figure it out, not just what you've already built.
How to get started
If SBIR sounds like a fit, here's your first 30 days:
- Identify 2-3 agencies that align with your technology. Use the table above to narrow down.
- Search recent awards on sbir.gov for companies similar to yours. This tells you whether your tech area gets funded and at what level.
- Read one solicitation from your top agency. Don't try to understand everything -- just get familiar with the structure.
- Talk to the program officer (at NSF, they're accessible and will tell you straight) or review the topic list (at DoD, via DSIP).
- Decide whether to go it alone or work with a consultant. First-time applicants face a steep learning curve, and 75-80% of first submissions get rejected. A good consultant compresses that curve. But you'll always need to bring the core ingredients: your technology, your data, your market insight.
The worst thing you can do is spend six months agonizing. The best thing is to submit a Project Pitch to NSF -- it takes a few hours -- and let the agency tell you whether you're on track.
Want a second opinion on fit?
We've written 500+ proposals across 30+ agencies and helped secure $1.6B+ in non-dilutive funding. If you want to know whether SBIR (or another program) is realistic for your company before you invest weeks in an application, that's what our Strategy Review is for. Fifteen minutes, no obligation, specific to your technology.