A decade of SBIR data reveals a program dominated by repeat winners, concentrated in two states, with only six percent of recipients reaching real commercialization.
The Small Business Innovation Research program has been celebrated for decades as America's premier innovation engine. Over $17.4 billion in federal defense R&D has flowed through it in the past ten years. Congressional authorizers talk about it as "America's Seed Fund." Program managers point to it as evidence that the Pentagon is serious about small business participation in defense innovation.
A decade of the data tells a more complicated story.
A point-by-point analysis of every SBIR contract awarded from August 2015 through July 2025 reveals a program that has drifted meaningfully from its statutory purpose. The concentration of awards among repeat winners is severe. The geographic distribution is narrow. And the commercialization outcomes, the metric that ultimately determines whether the program is doing what Congress asked it to do, are far weaker than the aggregate spending figure suggests.
The first issue is how tightly SBIR awards cluster among companies that have already won them before.
Across the full ten-year window, 84.1% of Phase 1 awards went to companies with two or more previous Phase 1 awards. For Phase 2, the figure was 87.7%. The stated purpose of the program is to foster participation by "emerging and undercapitalized small business concerns." The data shows a program dominated by established players.
The distribution gets starker at the top.
Phase 1 awards:
Phase 2 awards:
The overlap between Phase 1 and Phase 2 top performers is the sharpest signal in the dataset. 71.2% of Phase 1 Top 10% recipients also appear in the Phase 2 Top 10%. 77.2% of Phase 1 Top 1% recipients also appear in the Phase 2 Top 1%. The same forty-four companies dominate both phases.
When a multilinear regression is run across more than 300 variables for each SBIR recipient, the strongest single predictor of receiving a future SBIR award is having received a previous SBIR award. The more prior awards a company holds, the higher the likelihood of future awards. That's a predictive finding, not an intuition.
The concentration comes into sharp focus when named. Physical Sciences, Inc., based in Andover, Massachusetts, leads Phase 2 awardees with $183.0 million across 162 contracts. CREARE LLC of Hanover, New Hampshire, has captured $150.6 million across 163 Phase 2 contracts. Triton Systems, based in Chelmsford, Massachusetts, holds $144.4 million across 123 Phase 2 awards.
The Phase 1 leaders tell the same story. Triton Systems again, with $46.3 million across 244 contracts. Physical Sciences, with $37.9 million across 194 contracts. Lynntech Incorporated of College Station, Texas, with $29.7 million across 148 contracts.
Seven of the top ten Phase 1 recipients also appear among the top ten Phase 2 recipients. These are capable companies. That isn't the issue. The issue is that a program explicitly designed to seed new entrants has evolved into one that reliably returns capital to a well-defined group of established incumbents.
The distribution across the country is similarly narrow.
Together, two states account for nearly half of the program's total spend. That pattern would make sense if innovation capacity were genuinely concentrated in two regions. It isn't. There are capable technology companies building relevant capability in every part of the country, and a program designed to stimulate nationwide innovation should show a materially more distributed footprint than this.
The most serious finding isn't about who gets SBIR awards. It's about what happens after the awards are made.
Examining the full contract portfolio of every SBIR recipient produces three clear segments.
SBIR-Dependent companies (17.7% of recipients). These firms capture 35% of all SBIR obligations, $6.1 billion over the ten-year window. They derive 74% of their prime contract revenue from SBIR awards and show minimal transition into non-SBIR work. The program is their business.
Significantly SBIR-Dependent companies (36.9% of recipients). These firms capture another 40.2% of SBIR obligations, $9 billion in additional funding. They derive 43.5% of their revenue from SBIRs on average, with limited diversification beyond them.
Launchpad to Graduation (6.4% of recipients). These are the companies doing what the program is designed to produce. They begin highly SBIR-dependent (50% or more of revenue from SBIRs in their first two years) and successfully transition to primarily non-SBIR revenue over time (under 20% SBIR in their most recent two years). This is the cohort that represents the program's actual success stories.
Six percent.
The top 25% of SBIR awardees receive approximately 83.7% of all SBIR obligations, and most of them never convert SBIR-funded innovation into non-SBIR revenue. The program isn't acting as a launchpad. It's acting as a self-perpetuating funding stream for companies that have become very effective at writing SBIR proposals.
The SBIR program has four statutory purposes defined by the Small Business Administration. Two of them ("stimulate technological innovation" and "use small business to meet Federal R&D needs") are difficult to evaluate quantitatively. The other two are not.
Fostering emerging and undercapitalized small businesses. The heavy concentration of awards among repeat winners is a direct contradiction of this objective. New entrants face substantial barriers to entry. Established "SBIR mills" have developed refined proposal systems and long-standing relationships with program managers. The data doesn't show a program that's opening doors for new innovators.
Increasing private sector commercialization. A 6% graduation rate from SBIR-dependent to commercialized operations is the clearest evidence that the program is underperforming against its stated purpose. Innovation for its own sake is not the goal Congress wrote into statute. Technology transfer and commercialization is. The data suggests the program is delivering the former and largely failing to deliver the latter.
The SBIR program represents a serious federal investment with real potential to drive technological advancement. It is not working as designed. Five reforms would align it more closely with the purpose Congress wrote into law.
Enhance traceability and accountability. Establish rigorous tracking between SBIR contracts and subsequent non-SBIR contracts that result from SBIR-funded technology. Capture lessons learned from efforts that don't commercialize. Failed transitions are as informative as successful ones, and the program currently does not learn from them systematically.
Revisit the core program purpose. The primary mandate to "stimulate technological innovation" lacks a defined end state or measurable return on investment. As "America's Seed Fund," the program should carry explicit commercialization metrics and graduation pathways. Repeat SBIRs to the same companies with no follow-on contracts represent innovation without mission connection.
Refocus on problem owners and rapid prototyping. Necessity drives the sharpest innovation. A new class of SBIR that required solution development and testing within 30, 60, or 90 days, directly tied to operational end-users, would align the program more closely with how commercial technology actually gets built and adopted.
Address geographic disparities. Innovation capacity exists throughout the country. A program genuinely committed to nationwide participation would actively identify and develop capability in underserved regions rather than concentrating nearly half of its funding in two states.
Implement award diversity measures. Set-asides for first-time awardees, limits on consecutive awards to the same companies without demonstrated commercialization outcomes, and similar mechanisms would ensure that new entrants have realistic paths into the program.
When the top 10% of recipients capture more than 60% of total funding, and roughly 85% of awards flow to companies with multiple prior wins, the program has drifted from its mission. When only 6% of participants demonstrate actual commercialization, a program built to drive technology transfer is not doing what its founding legislation requires it to do.
Reform is necessary and achievable. Better tracking, sharper success metrics tied to commercialization rather than award volume, and active broadening of participation geographically and among new entrants would restore the program to the engine of American innovation it was designed to be.
The data is available. The patterns are clear. What remains is the decision to act on them.
No cadence. No fluff. Original analysis on government demand, defense procurement, and federal capital movement, sent when we have something worth reading.
For PE, VC, sell-side, BD, capture, and strategy teams covering defense.
