SBA Just Made SBIR Revenue Share an Eligibility Test. If You Won More Than 25 Phase II Awards in Five Years, Your FY2027 Grade Is Already Locked.

Most federal eligibility rules tell you what to do differently next year. This one tells you what you already did.

On September 2, 2026, the Small Business Administration published Updated Minimum Performance Standards for Commercialization for Firms That Receive Funding Through the Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) Programs in the Federal Register. It replaces a benchmark that had stood since 2013 and does something the old one never did: it makes the share of your revenue that comes from outside the SBIR program a condition of remaining eligible to compete.

The standard takes effect November 15, 2026. The first determination is made June 1, 2027. And because the test measures the three most recently completed fiscal years, the numbers that decide it are already in your general ledger.

What the new benchmark actually says

Three numbers define it.

Who is subject. Any small business concern that has received more than 25 Phase II awards — SBIR and STTR combined, across all participating agencies — during the five most recently completed fiscal years, excluding the current fiscal year.

What you must show. For the FY2027 assessment, a non-SBIR revenue share of at least 33 percent during the three most recent fiscal years, excluding the current one. For the FY2028 assessment and every year thereafter, that threshold rises to 50 percent.

What happens if you miss. The firm is not eligible to submit a proposal for a new Phase I or Direct-to-Phase II award from any participating agency for the one-year period beginning on the date the determination was made.

SBA defines non-SBIR revenue as the percentage of the firm's total revenue that is not derived from Phase I or Phase II SBIR or STTR award funding. The numerator therefore includes Phase III awards, private commercial sales, and non-SBIR government contracts. The denominator is total company revenue from all sources across the three-year window.

The authority is 15 U.S.C. 638(qq)(2), the minimum-performance-standards provision that arrived with the Small Business Innovation and Economic Security Act of 2026, signed April 13, 2026, which extended both programs through September 30, 2031 after a six-month lapse. The notice explicitly supersedes the benchmark announced at 78 FR 48537 and 78 FR 59410 — the 2013 pair of notices that established the standards firms have been graded against for thirteen years.

The change is structural, not incremental

Compare the tests side by side and the redesign becomes obvious.

The outgoing commercialization benchmark applied to firms with 16 or more Phase II awards in the past ten fiscal years, excluding the two most recently completed. To pass, a firm had to show either an average of at least $100,000 in sales and/or investment per Phase II award, or patents equal to at least 15 percent of the Phase II awards it received.

The incoming benchmark applies to firms with more than 25 Phase II awards in five fiscal years, and asks for a ratio.

Four things changed at once:

1. The applicability window compressed from ten years to five, and the count rose from 16 to more than 25. That is a far steeper rate — roughly five Phase II awards a year sustained over five years. Fewer firms are caught.
2. The test switched from an absolute floor to a ratio. The old standard asked whether you had produced a minimum amount of outside money per award. The new one asks what fraction of your business the government's early-stage R&D program represents.
3. The patent alternative is gone. Intellectual property no longer substitutes for revenue. A firm with a strong portfolio and thin outside sales had a pass route; it does not now.
4. The threshold escalates on a one-year timer. Thirty-three percent, then fifty.

The consequence of the second change is worth sitting with. Under the old rule, a firm could grow its way into compliance simply by selling more. Under the new one, a firm can grow non-SBIR revenue every single year and still fail, if SBIR revenue grows faster. The benchmark is not a productivity test. It is a dependency test.

The arithmetic that decides the first two assessments

Here is the part most firms will miss until it is too late.

The FY2027 determination happens June 1, 2027. It measures the three most recent fiscal years, excluding the current fiscal year. On a federal fiscal-year calendar, that is FY2024, FY2025, and FY2026 — all three of which will be closed books by the time the rule even takes effect in November 2026.

There is no compliance action available for the 33 percent test. It is a grade on work already delivered and revenue already booked. Firms subject to it can compute their own result today with a spreadsheet, and the answer will not change.

The FY2028 determination — the first at 50 percent — happens June 1, 2028 and measures FY2025, FY2026, and FY2027. Two of those three years are also already history. Only FY2027 remains movable. A firm that discovers it is at 38 percent has one fiscal year in which to alter a three-year average, which means the required swing in that single year is roughly three times the gap.

One caveat matters here and SBA's notice does not resolve it: the text refers to the small business concern's fiscal years, not explicitly to the federal fiscal year. Firms on a calendar year or an off-cycle fiscal year should not assume the windows above map exactly onto their own books. If your fiscal year ends in June or December, ask the SBA program office in writing which calendar governs before you model anything. The difference is an entire year of revenue in a three-year average.

Why SBA built this, and the argument on the other side

The policy target has a name in the trade press: SBIR mills — firms whose business model is winning SBIR awards rather than commercializing what the awards fund.

The concentration data is not subtle. An SSTI analysis of roughly 63,000 awards made between 2009 and 2019 found that 95 companies — 0.7 percent of the 13,046 firms that won anything — took more than 21 percent of all awards, using a threshold of 40 or more Phase I awards and/or 30 or more Phase II awards. At the other end of the same distribution, 41.5 percent of awardees won exactly one award, and 56 percent won between two and 19 Phase I awards.

Other cuts point the same direction. The top 25 companies, 0.53 percent of 4,703 firms in one dataset, received 18 percent of all Phase I and Phase II funding — over $2.3 billion, averaging $92 million per company. From 2010 through 2023, 25 companies captured 9 percent of all SBIR award dollars out of 17,563 total recipients. And on the defense side, 16 of the top 25 DoD SBIR recipients drew 50 percent or more of their DoD contracting revenue from SBIR awards rather than conventional contracting between FY2010 and FY2024 — a statistic that reads like a first draft of the rule SBA just wrote.

The counter-argument deserves a hearing, because it is empirical rather than rhetorical. Analysts including Gaster and Wessner have argued that the mills framing overstates both the concentration and the performance problem. Across 1992–2005, the top 20 firms received 8.6 percent of Phase II awards and the top 50 received 13.6 percent — meaningful but not dominant. Only 28 firms received at least 28 Phase II awards across that fourteen-year span. More pointedly, DoD Company Commercialization Report data showed that firms winning more than 125 Phase II projects generated average sales of $1.38 million per project — higher, not lower, than firms with fewer awards. Their conclusion: firms with the largest numbers of awards generate more commercial results in aggregate, not fewer. They also note that commercialization is only one of Congress's stated objectives; meeting agency research needs is another, and a firm that reliably solves hard agency problems is doing the job the statute describes.

The new benchmark does not resolve that dispute. It takes a side. A firm with $1.38 million in average sales per project still fails if SBIR is 70 percent of its revenue base, because the rule measures the ratio, not the return.

The other half of the squeeze: proposal caps

The benchmark does not arrive alone. The same 2026 reauthorization requires participating agencies to set limits on the maximum number of proposals a single small business concern may submit in response to SBIR Phase I or Phase II solicitations. Agencies may set those caps on a fiscal year basis, on a solicitation basis, or by individual topic within a solicitation.

The two mechanisms interact in a way worth modeling. Proposal caps reduce the number of awards a high-volume firm can win, which lowers SBIR revenue in the denominator — which helps the ratio. But they also reduce the award count that determines whether the firm crosses the 25-award applicability threshold in the first place. A firm sitting just above 25 Phase II awards over five years may find that agency caps carry it below the line within two or three cycles, exiting the benchmark entirely. A firm at 40-plus will not be so lucky.

Note also what the benchmark does not restrict. Failing it blocks new Phase I and Direct-to-Phase II proposals. Under the structure SBA has long applied to benchmark failures, a firm may continue performing its existing awards and remains able to pursue Phase II work flowing from an existing Phase I, and Phase III opportunities. The penalty cuts off the front of the funnel, not the back of it.

What to do between now and November 15

1. Compute your own ratio this week. Pull total company revenue by fiscal year for the last three closed years, split out Phase I and Phase II SBIR and STTR revenue, and divide. You will know your FY2027 grade before SBA does. If you pass at 33 percent, immediately run the same calculation against 50 percent — that is the real test, and it arrives twelve months later.
2. Confirm your award count. More than 25 Phase II awards across five completed fiscal years, all agencies combined, is the trigger. Firms that treat their DoD and their NIH portfolios as separate businesses often have not added them together. Your award history is public on sbir.gov; reconcile it against your own records, because the government's count is the one that governs.
3. Ask SBA in writing which fiscal year applies. If your books do not run October to September, this single question moves your measurement window by a full year.
4. Treat Phase III conversion as the highest-leverage revenue move available. Phase III dollars count as non-SBIR revenue in the numerator while still being SBIR-derived work you already know how to perform — the only lever that improves the ratio without requiring a new market. The reauthorization also directs SBA to update its policy directive requiring procurement center representatives to advocate for maximum practicable use of and transition to Phase III. That advocacy is now a compliance asset, not just a business-development nicety.
5. Re-examine whether you want every award you could win. For a firm near either threshold, an additional Phase II award is no longer unambiguously good news. It adds to the count that triggers applicability and to the denominator that determines the grade. That is a genuinely new strategic calculus for the SBIR-heavy firm, and the arithmetic should run before the proposal does.
6. If you are going to fail, plan the ineligibility year. A one-year block on new Phase I and Direct-to-Phase II proposals is survivable if you see it coming twelve months out and load the pipeline with Phase II continuations, Phase III conversions, and non-SBIR contract vehicles before June 1. It is not survivable if you learn about it in a determination letter.

For most of the roughly 17,000 firms that have touched the program, none of this applies — the applicability threshold is set well above where the median awardee operates, and single-award and few-award firms are untouched. For the several dozen firms at the top of the distribution, the SBIR program has just been redefined from a funding source into a funding fraction, with a number attached and a clock running.

Related reading: our analysis of the 2026 SBIR/STTR reauthorization and Strategic Breakthrough Awards, and proposal caps and foreign screening under the new rules.