Twenty-five companies classified as "small" collected $4.1 billion — 15 percent — of the Defense Department's $27.5 billion in Small Business Innovation Research (SBIR) Phase I and Phase II awards between fiscal years 2005 and 2024, according to an independent analysis published on War on the Rocks.

The 25 firms, which author Amanda Bresler terms "SBIR mills," won awards while employing near or above the 500-person ceiling and, in some cases, generating hundreds of millions in annual revenue. Top-five mills alone absorbed 5 percent of total DoD Phase I and Phase II spending. Meanwhile, less than 4 percent of $4.8 billion in Phase I dollars went to vendors with no prior federal experience.

Bresler heads SHELDON, an analytics unit of PW Communications, Inc., and has run SBIR-related analytics through the Naval Postgraduate School since 2018.

How the Pipeline Is Supposed to Work

President Ronald Reagan signed the Small Business Innovation Development Act in 1982. The program sequences awards in two principal phases:

• Phase I. Firms submit a short proposal addressing a government-defined problem. Awards typically run $40,000 to $150,000 and fund proof-of-concept work to reduce technical uncertainty.
• Phase II. Selected firms receive larger sums to prototype, test, and gather performance data, advancing technology toward acquisition.

The pipeline was meant to seed non-dilutive R&D capital for genuinely small firms and then transition outputs to government users. Bresler's data argue that pipeline has been diverted.

What Counts as "Small"?

The Small Business Administration lets firms qualify as "small" if they employ fewer than 500 people, regardless of revenue. That threshold lets firms worth hundreds of millions compete on the same footing as five-person startups. Bresler calls the loophole the program's central design failure.

Her three structural fixes:

• A revenue cap of $40 million annually to qualify.
• A $75 million lifetime ceiling on Phase I and Phase II awards per firm.
• A set-aside share of Phase I money reserved for first-time federal vendors.

What the Transition Data Show

Cross-referencing Federal Procurement Data System and USASpending.gov records, Bresler's team compared DoD SBIR awards with subsequent non-SBIR procurement contracts. Three named mills illustrate the gap between award volume and downstream procurement:

• Lynntech: $158 million in DoD Phase I/II funding; $17.3 million in subsequent non-SBIR contracts.
• Scientific Systems: $115 million in awards; $4.1 million in follow-on contracts.
• Spectral Energies: $97 million in awards; $1.2 million in follow-on contracts.

By contrast, 15 firms with 12 or fewer Phase I/II contracts landed more non-SBIR procurement than those mills — evidence, Bresler writes, that "the fact that so few new entrants break into the program exposes the merit-based fallacy." The Government Accountability Office reached similar conclusions in a separate review.

Why Evaluators Pick the Incumbents

Former Air Force chief of the Digital Integration Office Michael Meyer told Bresler that he and fellow evaluators routinely received mill proposals that cited work by genuinely small firms to validate their technical approach — then watched those same small firms submit stronger, lower-scoring proposals. Program offices, Meyer said, gravitated toward familiar recipients and justified choices "based on the quality of the mills' proposals."

Where the Legislation Stands

The SBIR authorization lapsed on September 30, 2025. Sen. Joni Ernst's INNOVATE Act would impose Bresler's three reforms. Her analyses have also informed the 2019 National Defense Authorization Act and Congressional and Senate testimony. Mills, their lobbyists, and host-state lawmakers counter — via Federal News Network commentary — that exclusion would punish merit rather than reward it.

What an R&D Manager Should Take Away

Methodological caveat: Bresler's datasets cover DoD SBIR only; civilian-agency patterns may differ. She disclosed receiving one Phase I and two Phase II DoD awards from 2019 to 2022 and acknowledged competing interests through her firm's federal and commercial consulting work.

For R&D leaders at genuinely small firms, the practical read is that SBIR functions less as a merit-based prize than as a bureaucratic literacy test, justifying dedicated compliance overhead in any bid budget. For program offices, the data raise a portfolio-governance question: how heavily to weight evaluator familiarity when scoring technical merit — and whether to track post-award transition rates before issuing the next tranche.

Whether the INNOVATE Act's structural fixes survive conference will determine whether the next reauthorization opens the channel back to the firms it was chartered to fund.