NSF's SBIR/STTR Program Came Back On With a New Solicitation, a $210M Pool, and a July 27 Deadline. Here's How the Phase I Math Actually Works.

For a deep-tech founder, the National Science Foundation's SBIR/STTR program is the closest thing the federal government offers to non-dilutive seed capital. It writes checks to companies, not universities; it does not take equity; and it funds the unglamorous middle of the technology-readiness curve — the part where an idea is too proven for a research grant and too unproven for a Series A. After a tense stretch in which the program's authorization lapsed, NSF turned the funnel back on with a new solicitation, NSF 26-510, and set the first full-proposal deadline at July 27, 2026.

That date is the headline, but it is not the number that should drive a founder's planning. The number that matters is the one most first-time applicants discover too late: you cannot submit a Phase I proposal on July 27 unless NSF has already invited you to. This is the deep dive on how the new solicitation is structured, what the reauthorization changed underneath it, and how to sequence the next several weeks so the deadline is reachable at all. (For the breaking-news version of the program's restart, see Granted News.)

The Phase I math: $305,000, six to eighteen months, no equity

Under NSF 26-510, a Phase I award runs up to $305,000 for six to eighteen months of research and development. That is a meaningful bump from the program's long-standing $275,000 ceiling, and it reflects the higher cost basis of the technical work NSF is trying to de-risk. NSF expects to make roughly 180 SBIR and 50 STTR Phase I awards annually under the new solicitation.

Phase I is a proof-of-concept stage. NSF is not asking for a finished product; it is asking whether the central technical risk in your idea can be retired. A strong Phase I proposal names the one thing that, if it works, makes the rest of the company plausible — and proposes to test exactly that.

Phase II is where the program becomes serious money: up to $1,250,000 over 24 months, available only to companies that have already completed an NSF Phase I. NSF makes roughly 80 SBIR and 10 STTR Phase II awards a year. There is also a Fast-Track option — up to $1,555,555 combined across a $400,000 Phase I component and a $1,155,000 Phase II component — for companies confident enough to commit to the full arc up front. Total program funding across all of this is approximately $210 million.

The structural point for a founder is that this is a ladder, not a lottery. The Phase I award is small relative to the Phase II that follows it, which means the real prize is option value: winning Phase I is what buys you the right to compete for the $1.25M. Treat the $305,000 as the entry fee to a much larger game.

The Project Pitch gate — and why July 27 is the wrong date to plan around

Here is the rule that derails first-time applicants. You cannot submit a Phase I full proposal without first submitting a Project Pitch and receiving an official invitation by email from NSF program staff. The Project Pitch is a short, three-page-equivalent summary of the technology, its innovation, its commercial potential, and the team. NSF program directors review it and respond — usually within a few weeks — with either an invitation to submit a full proposal or a decline.

That means the operative deadline for a company that has not yet pitched is not July 27. It is whenever the pitch needs to go in so that an invitation arrives with enough runway to write the full proposal before July 27. In practice, a founder aiming at the July 27 deadline should have a Project Pitch submitted no later than the first days of July, and ideally sooner. Leave the pitch until mid-July and the math simply does not close.

The solicitation also imposes traffic rules on the funnel: a company may submit a maximum of two Project Pitches per year, and a maximum of three submissions for the same underlying project across its lifetime. That makes the pitch a resource to spend deliberately, not a form to fire off. A declined pitch is not free — it consumes one of your two annual slots.

What the 2026 reauthorization changed underneath the solicitation

NSF 26-510 is not just a refreshed version of last year's documents. It is the first NSF SBIR/STTR solicitation written under the Small Business Innovation and Economic Security Act of 2026, which President Trump signed on April 13, 2026, extending the SBIR and STTR programs through September 30, 2031 after a roughly six-month lapse. Several provisions of that law shape how a founder should read the new rules.

National-security screening is now mandatory and heavier. The reauthorization requires agencies to assess whether applicants pose national-security risks and prohibits awards to companies with certain foreign ties. In practice this means more scrutiny of ownership structure, foreign investors, key-personnel ties to countries of concern, and cybersecurity practices. Founders with international cap tables or co-founders on certain visas should not treat this as a checkbox — it is a substantive review that can sink an otherwise fundable proposal. Get the ownership and affiliation disclosures clean before you pitch, not after you are invited.

Per-company proposal caps are coming. Beginning in fiscal year 2027, agencies will set their own limits on how many proposals a single company can submit per solicitation or per topic. For the July 27, 2026 window this is not yet binding, but it signals the program's direction: NSF wants depth over volume, and the era of carpet-bombing the agency with proposals is ending.

A new Phase II "strategic breakthrough" lane exists — but not really at NSF yet. The Act created a Phase II "strategic breakthrough" vehicle with a ceiling as high as $30 million, reserved for agencies whose annual SBIR expenditures exceed $100 million. That is a story to watch at the Department of War and other large-spend agencies more than at NSF, but it tells you where Congress wants the program's ambition to go.

Award decisions are supposed to come faster. The law pushes agencies toward a 90-day decision timeline. For a cash-strapped startup, the time-to-money has always been the program's worst feature; the reauthorization is an attempt to fix it.

The instrumentation pilot: a new lane worth knowing about

Alongside the core solicitation, NSF is running a pilot emphasis on next-generation instrumentation, novel experimental platforms, and scientific equipment. This matters because the SBIR program historically tilted toward software and devices with obvious consumer or enterprise markets, leaving hard-science tooling — the microscopes, sensors, and lab platforms that other researchers depend on — under-served. If your company builds the picks-and-shovels of science rather than a consumer-facing product, this pilot signals that NSF is explicitly inviting you. Frame the commercial-potential section around the researchers and labs who would buy the instrument, and the addressable market becomes legible to reviewers who might otherwise struggle to see it.

How NSF actually scores it: three criteria, not one

NSF reviews on Intellectual Merit, Broader Impacts, and Commercial Potential. First-time applicants from academia over-index on the first and neglect the third; first-time applicants from the startup world do the reverse. The proposals that win treat all three as load-bearing.

• Intellectual Merit is the technical risk and the credibility of the plan to retire it. Name the hard problem; do not bury it.
• Commercial Potential is whether a real customer will pay real money. Reviewers want evidence of customer discovery — conversations, letters of interest, a defensible wedge — not a top-down market-size slide.
• Broader Impacts is NSF's mission language: jobs, U.S. competitiveness, workforce, downstream scientific benefit. It is not boilerplate; it is part of the score.

Eligibility, in one paragraph

You must be a U.S. small business with no more than 500 employees (including affiliates), majority U.S.-owned in the ways the new security rules require, and the principal investigator must be employed at least 51% by the company at the time of award — a rule that trips up founders still holding a university appointment. For STTR specifically, the work must be done in partnership with a research institution, with defined minimum work splits. The PI commitment is at least one calendar month of effort per six months of the project for standard Phase I/II.

What to do before July 27

If you are serious about the July 27 window, the sequence is unforgiving and the first step is the urgent one:

1. Submit a Project Pitch this week. Without an invitation, July 27 is irrelevant. This is the single highest-leverage action.
2. Clean up your ownership and affiliation disclosures now, ahead of the national-security review — foreign investors, co-founder visa status, and cybersecurity posture.
3. Do customer discovery in parallel so the Commercial Potential section rests on real conversations, not assertions.
4. If you build scientific tooling, target the instrumentation pilot and frame the market as the research community.
5. Decide Phase I vs. Fast-Track deliberately — Fast-Track is faster to real money but demands a more developed commercialization case up front.

The reauthorization bought the SBIR program five more years of life, a larger Phase I check, and a faster clock. What it did not change is the gate: the founders who win in July are the ones who pitched in the first days of the month. Everyone else is planning for November.