Ideas for Improving the Small Business Innovation Research Program

As Lars Erik Schönander makes clear in “Reauthorizing the Small Business Innovation Research Program Isn’t Enough” (Issues, Spring 2026), SBIR is massively improved. But it still fails to adequately address the decades-old problem posed by so-called SBIR mills, nor does it materially encourage first-time innovators to engage.

SBIR mills, a small number of companies that benefit from the program but never commercialize any products, were born of a 40-year schism in the perceived objectives of SBIR. On the one hand, evaluators praise the success of companies that received SBIR funding in their early days, such as AMGEN, Qualcomm, and Ginkgo Bioworks, and correctly conclude that the program performed superbly. The key measure, commercial sales divided by SBIR investment, or ROI, yields more than $25,000 of non-SBIR revenue, on average, for every SBIR dollar awarded to the 20 top performers. Graduation out of the program is the intended ideal outcome.

On the other hand, some companies acted as though performing SBIRs alone was in fact a business. Enabled by lax stewardship from program managers inexperienced in, or ambivalent to, commercialization, these companies were lulled into a false sense that commercialization was just a nicety, not the primary objective. The top winners of SBIR funding produced less than $1 in non-SBIR revenue for every SBIR dollar they consumed. This act of putting intellectual property “on the shelf” was even worse than not performing the research at all. Locked-up IP actually inhibits other innovators from pursuing the same technology and releasing a product.

There are several reasons why newcomers choose not to engage with SBIR. The program appears rigged for insiders, as evidenced by the ongoing presence of SBIR mills. Commercial innovators are not in the labor business, so they don’t want to get sucked into defense cost accounting. They determine that developing a 25-page proposal that takes months to evaluate, with a 15% win probability, is not worth the effort. While a great deal of progress has been made on the last two points, on balance it’s still not an overwhelming case.

Several improvements would help. The original draft legislation carried a $75 million lifetime funding cap per company. This should be restored. This limit had several benefits. It conveys the unambiguous message that SBIR is seed funding to get a company into business and not a business unto itself; $75 million is sufficient to bootstrap a product if a company pursues capital from additional sources; and it tells newcomers that they have a fair shot as all companies eventually have to graduate.

In addition, the program’s original limit of 500 employees per company should be updated to a more contemporary 200; commercial viability should be a full 50% of award criteria (evaluated by competent professionals); and the Department of Defense should provide a standing budget on the procurement side to fund end-user matches for the new “Strategic Breakthrough” production phase.

These changes would help curb historical abuse of the SBIR program, even in its reauthorized form. This would help the program reach its intended goal of launching innovative new companies over the infamous valley of death, graduate them, bring needed new products and services to market, and make room for the next class.

Warren Katz