SBIR Grants 2026: Phase I, Phase II, STTR, NIH, NSF

The SBIR program has funded small-business R&D since it was created in 1982. It's responsible for early-stage funding behind companies like Qualcomm, Symantec, and iRobot. But the program is also poorly understood, even by founders who've heard of it. Most online guides give you the Wikipedia version: three phases, 11 agencies, small business eligibility. This guide goes deeper, covering how review actually works, which agencies are realistic targets for your technology, and the practical tradeoffs that determine whether SBIR is worth your time.

Our analysis is that SBIR applicants lose most often when they choose the agency by award ceiling instead of by customer, topic owner, and review culture. For example, a dual-use software company may be a better fit for DOD if a program office owns the problem, while the same company may be a better fit for NSF if the first buyer is commercial and the technical risk is platform-level R&D.

Where official sources stop: agency SBIR pages explain topics, phases, eligibility, and award caps, but they don't tell a founder whether the idea is research risk or normal product work. A software service with no technical research question should skip SBIR and look at customers, loans, or state support first.

What Changed and What to Check (October 2026)

Grantsights checked SBIR.gov, NIH Guide notices, NSF 26-510, CRS Insight IN12705 and the USDA NIFA SBIR page on October 16, 2026. Four checks for planning an SBIR year in FY2027:

• Reauthorized through September 30, 2031. The Small Business Innovation and Economic Security Act (P.L. 119-83, S. 3971 (opens in a new tab)), signed April 13, 2026, ended a six-month lapse, per NOT-OD-26-090 (opens in a new tab). From FY2027, every agency must cap proposals per company, according to CRS Insight IN12705 (opens in a new tab).
• Two NIH caps apply at once. HHS takes at most nine Phase I, Fast-Track and Direct-to-Phase II applications per company per fiscal year (NOT-OD-26-090). NIH also takes at most six applications per PI per calendar year, R43 and R44 included (NOT-OD-25-132 (opens in a new tab)).
• Ceilings and next dates. Up to $335,937 for Phase I and $2,239,578 for Phase II without an SBA waiver as of October 16, 2026 (SBIR.gov (opens in a new tab)), up from $323,090 and $2,153,927 in April; NIH's SEED site lists the same $335,937 and $2,239,578 (checked November 4, 2026). NIH's next date is January 5, 2027, with no late applications (NOT-OD-26-064 (opens in a new tab)). NSF takes full proposals November 4, 2026 and March 4, 2027, after a Project Pitch (NSF 26-510 (opens in a new tab)).
• Repeat winners face a benchmark. Firms with 21 or more Phase I awards in the past 5 fiscal years must meet SBIR.gov's Phase I to Phase II transition benchmark.

For example, one company that sends three NIH Phase I applications in a year with the same PI uses three of its nine HHS slots and three of that PI's six NIH slots.

Grantsights' apply/skip rule: apply when one technical risk, one agency reason, one Phase I experiment, one buyer path, and one non-grant runway plan fit on a single page. Skip when the founder only needs runway. The better next step may be customer revenue, state technology grants, accelerators, venture capital, or federal contracting.

Next route: Place SBIR inside the small-business grant route, then compare the SBIR Phase I path and live SBIR records. Use this page for SBIR rules and the route pages for agency fit.

How SBIR Grants Work: The Three-Phase Structure

SBIR operates on a phased model that mirrors how technology development actually progresses: prove the idea works, build it out, then take it to market.

Phase I: Feasibility and Proof of Concept

Phase I SBIR grants fund early exploration of whether your technical approach is viable. You're not building a product. You're testing whether your core hypothesis holds up.

Phase I proposals are typically 15 to 25 pages depending on the agency. The work plan should demonstrate a clear technical question and a realistic methodology for answering it. You don't need preliminary data at most agencies (NSF explicitly says you don't), but having some early results dramatically strengthens your application.

Grantsights' take: Phase I is often the hardest award to win because you have no track record with the agency. Phase II odds are generally better because you've already shown you can execute. Treat Phase I like a job interview: the technical plan matters, but so does demonstrating that your team can actually do the work.

According to NSF's SBIR/STTR Phase I guidance, the Phase I award is meant to test technical feasibility, not to fund a full launch. According to NIH SEED, applicants requesting budgets above the standard SBIR limits need institute approval before submission.

2026 SBIR Phase I and Phase II Award Amounts by Agency

Award ceilings move with agency guidance, so applicants should treat the solicitation as the controlling source. The 2026 planning ranges below are enough for early go/no-go work:

The practical rule: use NIH for biomedical product development when the institute fit is clear, NSF for commercial deep-tech ideas with broad market use, and DOD when a named defense customer or topic owner has a direct need. The best award amount is not the largest ceiling. It is the agency path where the reviewer can see why the federal mission needs your technical result.

For a deep dive on Phase I applications, see our SBIR Phase 1 grant guide.

Phase II: Full Research and Development

Phase II is where the real development happens. You've proven feasibility in Phase I, and now you're building a prototype or advancing the technology toward something that can actually be used.

At most agencies, you must have completed a Phase I award to apply for Phase II. Some agencies, including NIH and DOD, offer Direct-to-Phase II for companies that already have feasibility evidence. NSF doesn't: under NSF 26-510 (opens in a new tab), its routes are Phase I, Phase II, Fast-Track (a combined Phase I of up to $400,000 and Phase II of up to $1,155,000), and Phase IIB supplements of $50,000 to $500,000 for active Phase II awardees. The solicitation gives two Fast-Track totals: its award information says up to $1,555,000, which matches the two components, while its program description says up to $1,555,555. Budget to $1,555,000 unless NSF confirms the higher figure.

Phase II proposals are longer and more detailed, typically 40 to 60 pages. Reviewers expect a clear development plan, specific technical milestones, and a credible commercialization strategy. The commercialization plan matters more than most applicants realize. Agencies don't just want good science; they want technology that will eventually generate revenue, create jobs, or transition into government use.

Our SBIR Phase 2 guide covers the Phase II proposal in detail, including how to structure your commercialization plan.

Phase III: Commercialization and Follow-On Funding

Phase III is the most misunderstood part of SBIR. It's not a grant. There's no SBIR set-aside money for Phase III. Instead, Phase III refers to the commercialization stage, where you bring your SBIR-funded technology to market through private sales, federal contracts, or additional non-SBIR government funding.

What makes Phase III special is the sole-source contracting authority. Federal agencies can award Phase III contracts to SBIR awardees without competitive bidding. This is a major advantage for defense contractors, where the normal procurement process can take years. A DOD program office can directly contract with your company to produce the technology you developed under SBIR. If the real need is industrial-base expansion or production capacity, search DPA Title III in the federal grant database alongside your SBIR plan.

We cover Phase III in detail in our SBIR Phase 3 commercialization guide, including how to position for follow-on contracts during your Phase II work.

---

SBIR Eligibility: Who Qualifies

SBIR eligibility comes down to five requirements. Miss any one of them and your proposal gets rejected before a reviewer opens it.

1. For-profit business. Nonprofits, universities, and individuals without a formal business entity don't qualify. You need to be organized as a corporation, LLC, or partnership.
2. 500 or fewer employees. This is based on headcount, not FTEs. Part-time employees count. The count includes employees of affiliates if your company is majority-owned by or affiliated with another entity.
3. U.S. ownership and control. At least 51% of the company must be owned by U.S. citizens or permanent residents. Alternatively, the company can be majority-owned by another U.S. for-profit business that itself meets the ownership test.
4. PI employment. The Principal Investigator must be primarily employed by the small business during the award period. "Primarily employed" means more than 50% of their professional time.
5. Work location. The funded research must be performed in the United States, including its territories and possessions.

The 2026 reauthorization added two rules that matter before you write. According to CRS Insight IN12705 (opens in a new tab), agencies must screen applicants for security risk through due diligence and required disclosures, and a risk finding can be grounds to deny an award. The law also created strategic breakthrough awards of up to $30 million for Phase II awardees, which must be matched dollar for dollar.

The Venture Capital Wrinkle

If your company has taken venture capital funding, you need to pay extra attention to the affiliation rules. The SBA counts employees of affiliated companies when determining whether you meet the 500-employee threshold. If a VC firm holds a majority stake in your company and also holds majority stakes in other companies, those other companies' employees may count toward your total.

The SBIR/STTR Reauthorization Act of 2011 opened a path for companies majority-owned by multiple VC firms, hedge funds, or private equity firms (15 U.S.C. 638(dd)). NIH, DOE, and NSF may use up to 25% of their SBIR funds for these companies, and other agencies up to 15%, but each agency decides whether to use the option. NSF has opted out: its Seed Fund (opens in a new tab) doesn't fund companies majority-owned by multiple VC, private equity, or hedge fund firms. Using this path also involves additional ownership disclosures.

---

The 11 SBIR Agencies: Where the Money Goes

Eleven federal agencies are required to run SBIR programs because they each have extramural R&D budgets exceeding $100 million. They set aside 3.2% of those budgets for SBIR awards.

According to SBA SBIR program rules, agencies with extramural R&D budgets above $100M must reserve 3.2% for SBIR, while agencies above $1B also run STTR set-asides. That statutory set-aside is why SBIR remains a recurring funding path rather than a one-off competition.

Each agency's dollar total changes yearly with its extramural R&D budget. The SBA annual report on SBIR.gov (opens in a new tab) lists current per-agency obligations; in FY2022, agencies obligated $4.4 billion in SBIR awards and $662.3 million in STTR awards, according to CRS.

DOD runs the largest SBIR program by dollars, which means defense-related technology has the most opportunities by raw volume. But "defense-related" is broader than most people think: DOD funds work in materials science, cybersecurity, biomedical devices for warfighter health, energy systems, manufacturing processes, and even agriculture technology.

NIH is the second largest and dominates in biomedical and health-related technology. Their SBIR program spans all 24 institutes and centers, each with different priorities.

Companies that need a formal research-institution partner should compare the SBIR path with NIH STTR records in the federal grant database before choosing the application route.

For a detailed breakdown of how these agencies differ in review process, award timelines, and strategic priorities, see our SBIR agencies comparison guide.

---

SBIR Success Rates: What Your Real Odds Are

There's no single cross-agency success rate on SBIR.gov, and rates swing by agency, component, topic, and year. Treat any one figure you see online with caution, and look for the agency's own data instead.

Where to Find Real Odds

NSF 26-510, for example, estimates about 180 SBIR Phase I and 80 SBIR Phase II awards per year, which gives a sense of scale but not a success rate.

What overall numbers hide: Companies that have won SBIR awards before have an advantage: they know how to write for the review process, they have prior relationships with program managers, and they have past performance data that strengthens their proposals. First-time applicants should expect a harder first round.

Phase II odds are generally better. Because only Phase I awardees (or Direct-to-Phase II applicants at agencies that allow it) can apply, the pool is smaller and more qualified.

Our guide for first-time SBIR applicants covers strategies specifically designed to improve your odds when you don't have prior SBIR experience.

---

How SBIR Review Works Across Agencies

Understanding how your proposal gets evaluated is half the battle. Each agency runs its own review process, but there are common patterns.

NIH Review Process

NIH's parent SBIR notice, PA-27-100 (opens in a new tab) (posted May 28, 2026), sends applications to a Scientific Review Group convened by the Center for Scientific Review. SBIR applications still get five separately scored criteria: Significance, Investigator(s), Innovation, Approach, and Environment. NIH's simplified review framework, which folds those into three factors from January 25, 2025, covers research project grant codes such as R01 and R21 but doesn't list the SBIR codes (NOT-OD-24-010 (opens in a new tab)). The SBIR criteria ask business questions too: Significance asks whether a Phase I has commercial potential and whether a Phase II Commercialization Plan shows a high probability of commercialization, and Environment asks whether the small business has the support and resources the project needs.

Scores use NIH's 9-point scale, where 1 is best (NIH scoring guidance (opens in a new tab)). Under PA-27-100, generally only the top half of applications are discussed and get an overall impact score, but every application gets a written critique. Funding then depends on that merit score, a second review by the institute's advisory council, available funds, program priorities, and a security-risk check under the HHS Due Diligence Program.

Use what's public before you write: NIH posts rosters of its Scientific Review Groups (opens in a new tab), and if you've applied before, your own summary statement shows each reviewer's critique and criterion scores.

NSF Review Process

NSF uses ad hoc panels of 3 to 5 reviewers per proposal. Review criteria are Intellectual Merit and Broader Impacts, plus NSF's specific SBIR criteria including Commercial Potential. NSF program directors have wide discretion. A proposal that scores well in panel can still be declined if the program director doesn't see commercial viability, and a borderline proposal can get funded if the PD believes in the team.

DOD Review Process

DOD review varies enormously by component. DARPA uses internal review by program managers. Army and Navy components often use mixed panels of government scientists and external reviewers. Air Force's AFWERX uses a different model that includes pitch events. DOD topics are very specific, and your proposal must address the stated topic. Unlike NIH and NSF, where you propose your own research question, DOD tells you what problem to solve.

What Review Panels Actually Look For

What tends to separate funded proposals from the rest:

• A clear technical question. Not "we'll develop X." Instead: "We'll test whether approach Y can achieve performance metric Z under conditions W." Specificity wins.
• A team that can execute. Reviewers read your team section carefully. If your PI doesn't have relevant technical experience, that's a red flag.
• Awareness of the state of the art. Show you know what's been tried, what works, and where the gaps are.
• A realistic timeline. Proposing to accomplish 3 years of work in 6 months makes reviewers question your judgment.
• Commercial potential that goes beyond hand-waving. "The market is $10 billion" means nothing. "We've talked to 15 potential customers and 8 expressed purchase intent at $X price point" means everything.

For detailed guidance on writing your proposal, see our guide on how to write an SBIR proposal.

---

SBIR vs STTR: Which Program Should You Choose?

SBIR and STTR are sister programs, but the eligibility rules create meaningful differences in how you structure your project.

Choose SBIR when: You can do the work in-house, your PI is on staff, and you want access to the full range of 11 agencies.

Choose STTR when: A university researcher has the domain expertise you need, you want a PI who can stay at their faculty position, or the research institution has facilities you can't replicate.

For a detailed comparison, our SBIR vs STTR guide covers every difference that matters for your application decision. And if you're specifically interested in STTR, our STTR research partner guide covers how to find and structure the required partnership.

If you are deciding between NIH small-business lanes, compare our NIH SBIR/STTR guide with NIH SBIR and STTR records in the federal grant database.

---

The SBIR Application Process: Step by Step

Here's the actual sequence from "I want to apply" to "I have an award."

Step 1: Register Your Business (2-6 weeks)

Before you can submit a single proposal, you need:

• SAM.gov registration. This is the System for Award Management. You need a Unique Entity Identifier (UEI), which replaced DUNS numbers. Allow 2 to 4 weeks for processing. Start this first.
• SBA Company Registry. Register at sbir.gov as a small business concern. This verifies your size and ownership.
• Agency-specific registration. NIH requires eRA Commons. NSF requires Research.gov. DOD requires a DSIP account. Register for these in parallel with SAM.gov.

Step 2: Find the Right Opportunity (1-4 weeks)

Each agency solicits proposals differently:

• DOD posts specific topics on DSIP (dodsbirsttr.mil), usually 2 to 3 times per year. You must respond to a listed topic.
• NIH has standing omnibus solicitations for most institutes, with standard due dates of September 5, January 5, and April 5. You propose your own research topic. NIH doesn't accept late SBIR or STTR applications under NOT-OD-26-064, so the due date is final.
• NSF requires a Project Pitch first; invited full proposals go in by set deadlines (November 4, 2026 and March 4, 2027 under NSF 26-510). A company can have only one Project Pitch pending at a time and can send at most two in a 12-month period. You propose your own commercial technology.
• DOE, NASA, USDA post topics on set schedules, usually annually.

Search open solicitations at sbir.gov (opens in a new tab), which aggregates opportunities across all 11 agencies.

Step 3: Write Your Proposal (4-8 weeks)

A competitive Phase I proposal takes 4 to 8 weeks of focused work. This isn't something you can knock out over a weekend. Your proposal needs:

• Technical approach and methodology
• Statement of work with milestones
• Budget and budget justification
• Team qualifications and roles
• Facilities and equipment description
• Commercialization plan (especially for NSF)
• Letters of support (if applicable)

Step 4: Submit and Wait (several months)

After submission, review timelines vary by agency:

• NSF: check the notification timeline in NSF 26-510
• DOD: depends on the component and topic
• NIH: runs through peer review and an advisory council round, so it's usually among the slower paths
• DOE and NASA: each solicitation lists its expected selection date

Step 5: Negotiate the Award (2-8 weeks)

If selected, you'll negotiate the award with the agency's contracting office. This involves finalizing the budget, agreeing on milestones, and executing the contract or grant agreement. DOD awards are typically contracts (requiring FAR compliance). NIH and NSF awards are typically grants (under 2 CFR 200).

---

Which Agency Should I Apply To? A Decision Tree

This is a common question. The answer depends on your technology area, your team's strengths, and your commercialization goals.

Start with technology fit:

• Biomedical, health, pharmaceutical, medical devices → NIH
• Software, algorithms, platform technology, AI/ML → NSF
• Defense, national security, dual-use → DOD
• Energy, materials, climate → DOE
• Aerospace, space technology → NASA
• Agricultural technology, food safety → USDA
• Homeland security, border, cybersecurity → DHS

If your technology spans categories, pick the agency whose mission aligns most closely with your primary customer. A medical device for military medics? That's DOD, not NIH. A cybersecurity tool for financial institutions? That could be NSF (commercial) or DHS (government).

If you've never applied before, NSF and NIH are the most founder-friendly. NSF in particular has extensive resources for new applicants, including the I-Corps program and the Beat the Odds bootcamp. DOD requires responding to specific topics, which can be harder for newcomers to match.

See our SBIR agencies comparison for a detailed breakdown of each agency's process, culture, and what makes them different in practice.

---

Commercialization Expectations: What Agencies Actually Want

SBIR isn't a research grant. It's a program designed to produce commercial outcomes. Every agency tracks commercialization metrics, and your performance affects future applications.

The SBA Commercialization Benchmark

SBA established a minimum commercialization benchmark that companies must meet to remain eligible for Phase I awards. According to SBIR.gov performance benchmarks (opens in a new tab), the Commercialization Rate Benchmark applies when a company has received 16 or more Phase II awards during the past ten fiscal years, excluding the two most recently completed fiscal years. The company must average at least $100,000 of sales and/or investment per Phase II award, or hold patents equal to at least 15% of its Phase II awards. A separate transition benchmark applies at 21 or more Phase I awards in the past five fiscal years. Companies that don't meet a benchmark can be barred from new Phase I awards for a period.

What "Commercialization" Means by Agency

• DOD: Production contracts, technology transition to a program of record, licensing to defense primes
• NIH: FDA clearance/approval, licensing deals, venture funding for clinical development, sales revenue
• NSF: Commercial product sales, licensing revenue, follow-on private investment
• DOE: Technology licensing, joint ventures, deployment at scale

Positioning for Commercialization from Day One

Don't wait until Phase II ends to think about commercialization. Agencies evaluate your commercial readiness at every stage. During Phase I, you should be identifying customers. During Phase II, you should be developing those relationships into letters of intent, pilot programs, or beta testing agreements.

For DOD specifically, positioning for Phase III follow-on contracts should start during your Phase II performance period.

---

What Top-Scoring SBIR Proposals Do

Strong proposals tend to do three things that weaker proposals don't.

First, they specify a customer. Not "the defense market" or "the healthcare industry" but a named contracting entity, program office, or identified buyer. For example, a DOE applicant could name the national laboratory it intends to work with in Phase III and tie its roadmap to that laboratory's published research agenda, which gives reviewers something concrete to score.

Second, they present preliminary data. The SBIR program is designed to fund feasibility work, but proposals that show some evidence the concept isn't purely theoretical, even a simulation, a prior related experiment, or a literature base, consistently outscore proposals that are purely theoretical.

Third, they match the tone of the agency. NIH reviewers respond to clinical language and patient outcomes. NSF reviewers respond to scientific rigor and broader impact framing. DOD reviewers respond to operational use cases and technology readiness levels. Using the wrong register for the agency is one of the most common reasons technically strong proposals score lower than they should.

What sbir.gov Doesn't Tell You

The official SBIR website is a useful starting point, but it leaves out several things that matter for your success.

Agency culture shapes your experience more than published rules. NSF program directors will talk to you before you submit. NIH program officers will too, but they're less likely to give you direct feedback on your idea. DOD topic authors are often willing to discuss the topic but may be restricted in what they can share. These informal conversations are some of the most useful things you can do before writing a proposal, and the official site barely mentions them.

Resubmission is normal, not a failure. Most successful SBIR companies didn't win their first proposal. NIH explicitly allows resubmissions with reviewer feedback incorporated. NSF lets you reapply in a later window. The official site tracks award data but doesn't show you the resubmission history behind those awards. Starting in FY2027, per-company proposal caps under P.L. 119-83 mean each resubmission also uses one of your limited slots, so plan them.

The "small business" requirement creates a cliff. If your company grows past 500 employees during an active award, you can finish the award but can't apply for new ones. Companies approaching this threshold sometimes spin off SBIR-focused subsidiaries, which is allowed as long as they meet all eligibility requirements. The official guidance doesn't address this common growth scenario.

Award amounts are negotiable within limits. Published maximums aren't fixed ceilings. NIH allows budget exceptions above the standard cap with program officer approval. DOD topics sometimes specify different amounts. Don't assume the standard amounts are your only option, especially if your project has a strong justification for additional funding.

---

Related Guides

• SBIR Phase I winners and award history
• NSF SBIR Phase I winners and award history
• DARPA Grants
• SBIR Success Stories
• SBIR Phase 2 Guide
• SBIR Commercialization Plan Guide
• Guide for First-Time SBIR Applicants
• SBIR Agencies Comparison

---

Translational Routes to Compare Before SBIR

Some NHLBI therapeutics projects may be better served by Catalyze than SBIR/STTR. Compare the Catalyze R61/R33 target-identification program record when the work still needs target validation and screening. Use the Catalyze direct R33 lead-series program record when the project already has validated compound-series or prototype evidence.

---

Grantsights Trust Check: Apply or Skip This Path

Run these 4 checks before a Phase I draft. Apply when the company's size and ownership, the agency topic, the budget ceiling, and the submission date all match the current solicitation, and the company and PI still have proposal slots left under the FY2027 caps. Skip or pause when the fit is only thematic, the company can't document eligibility, registrations won't finish before the date, or the agency's topic points to a different route.

Sources

• SBA SBIR/STTR Program Overview: https://www.sbir.gov/about (opens in a new tab)
• SBIR/STTR Annual Reports: https://www.sbir.gov/impact/impact-reports (opens in a new tab)
• NIH SBIR/STTR Funding: https://seed.nih.gov/ (opens in a new tab)
• PA-27-100, NIH, CDC and FDA parent SBIR (R43/R44), Section V review criteria: https://simpler.grants.gov/opportunity/d1ba49e5-3684-4420-849a-ab2330ec493e (opens in a new tab)
• NOT-OD-24-010, Simplified Review Framework activity codes: https://grants.nih.gov/grants/guide/notice-files/NOT-OD-24-010.html (opens in a new tab)
• NSF SBIR/STTR Program: https://seedfund.nsf.gov/ (opens in a new tab)
• DOD SBIR/STTR Innovation Portal: https://www.dodsbirsttr.mil/ (opens in a new tab)
• Small Business Innovation and Economic Security Act, P.L. 119-83 (S. 3971): https://www.congress.gov/bill/119th-congress/senate-bill/3971 (opens in a new tab)
• CRS Insight IN12705, Small Business Research Programs Reauthorized After Six-Month Lapse (July 13, 2026): https://www.congress.gov/crs-product/IN12705 (opens in a new tab)
• NOT-OD-26-090, HHS annual SBIR and STTR application limit: https://grants.nih.gov/grants/guide/notice-files/NOT-OD-26-090.html (opens in a new tab)
• NOT-OD-25-132, NIH six-application limit per PI: https://grants.nih.gov/grants/guide/notice-files/NOT-OD-25-132.html (opens in a new tab)
• NOT-OD-26-064, NIH late application policy: https://grants.nih.gov/grants/guide/notice-files/NOT-OD-26-064.html (opens in a new tab)
• USDA NIFA SBIR/STTR program information: https://www.nifa.usda.gov/grants/programs/sbir-sttr/program-information (opens in a new tab)
• SBA Size Standards and Affiliation Rules: https://www.sba.gov/size-standards (opens in a new tab)