ED/IES SBIR's FY2026 Triple-Track Closes June 29 — Why Phase IA, Phase IB, and Direct-to-Phase-II Are Three Different Bets on Education R&D

The Department of Education's Institute of Education Sciences runs the smallest SBIR program in the federal government — roughly $7 million in annual obligations against $1.2 billion at NIH and $215 million at NSF — but it also runs one of the most consequential. Almost every commercially successful education-research product of the past fifteen years (Renaissance Learning's myIGDIs, BrainPOP's adaptive assessment work, Mursion's mixed-reality teacher training, the original Quill.org grammar engine) traces a substantial chunk of its R&D back to an ED/IES SBIR Phase I or Phase II award. The agency's reviewer panel — drawn from education researchers, classroom practitioners, and edtech entrepreneurs in roughly equal measure — has historically been better than its NIH or NSF counterparts at distinguishing R&D that will reach actual students from R&D that will publish a paper and disappear.

On April 30, 2026, ED/IES released its FY2026 solicitations, and the entire annual cycle now closes on Monday, June 29, 2026: Phase IA and Phase IB at 11:00 AM Eastern, Direct-to-Phase-II at 2:00 PM Eastern. There is no second window. The triple-track structure — three distinct solicitations published under separate procurement numbers (91990026R0003 for Phase IA, R0004 for Phase IB, R0005 for Direct-to-Phase-II) — codifies a theory of education R&D that the agency has been refining for the past several cycles: that the federal dollar should flow differently depending on whether you are inventing a new category, improving an existing one, or scaling proven research into a market product.

For founders, the choice of track is the single most strategic decision in the application. Picking the wrong one will sink an otherwise strong proposal.

Phase IA — the novelty track

Phase IA is the cleanest mental model. The agency funds $250,000 over nine months for the development of novel education technology where minimal prior technological work exists. The phrasing matters. "Novel" is interpreted strictly: the proposed technology must be independent from existing commercial products and must employ a different technical approach than any prior work by the applicant. A startup pitching a "better" AI tutor that uses standard transformer architectures is not a Phase IA project, even if the user experience is improved. A startup pitching a tutor that uses a fundamentally different feedback-loop architecture — say, a reinforcement-learning-from-classroom-interaction model trained on millions of authentic teacher-student exchanges — is.

The Phase IA narrative is, in practice, a research proposal disguised as a product proposal. Reviewers want to see a clear articulation of the technical hypothesis ("our approach will achieve X result that prior approaches cannot"), a falsifiable feasibility plan for the nine-month period, and a credible pathway to Phase II should the feasibility work succeed. Successful Phase IA awardees may apply in FY2027 for $1 million in Phase II funding across two years — but the Phase II competition will be judged on whether the Phase IA actually demonstrated the technical claim, not on the elegance of the original pitch.

The hidden requirement: Phase IA awardees must show that the technical novelty translates into something a teacher or student will actually use. ED/IES reviewers reject Phase IA applications that are technically impressive but lack a credible end-user story. Founders who can pair a strong technical hypothesis with a school partner who has already committed to pilot-test the resulting prototype are the standard winners.

Phase IB — the component-extension track

Phase IB is the more recently codified track and the one most founders misunderstand. It also funds $250,000 over nine months, but the project scope is fundamentally different: Phase IB funds new components added to existing research-based education technology products. The component must be distinct from the applicant's current offerings and developed using a different technological methodology than the parent product.

The classic Phase IB story looks like this: a small edtech company built a literacy product around early-reading fluency assessment, the product has been validated in published studies, and the company now wants to extend the product to capture vocabulary acquisition. The vocabulary module is not a separate product — it lives inside the existing literacy platform — but it uses different technical machinery (a different psychometric model, a different content-tagging approach, a different adaptive engine). Phase IB funds the development of that module.

The Phase IB track is, in effect, a structural answer to the criticism that SBIR funding incentivizes startups to invent new products rather than improve products that schools already use. By creating a dedicated track for extensions and components, ED/IES is signaling that the agency wants its portfolio companies to deepen the products that have already cleared market validation, not to fragment their attention across novel-but-untested concepts.

Phase IB applicants must be especially careful about novelty framing. A "new component" that is really just a feature update to an existing module will be rejected. The component must use a different technical methodology than the parent product — different ML architecture, different content model, different psychometric approach. Founders who pitch incremental improvements to an existing product will be told to fund the work themselves.

Direct-to-Phase-II — the scale-up track

The Direct-to-Phase-II solicitation is the largest individual award available ($1 million over two years) and the most narrowly scoped. The agency funds the scaling of existing, evidence-based innovations originally created by researchers at universities or non-profit education research organizations. The word "evidence-based" is load-bearing: the underlying innovation must have been studied in a rigorous evaluation that meets the What Works Clearinghouse standards or an equivalent threshold. Non-evidence-based projects are explicitly ineligible.

The intended applicant for Direct-to-Phase-II is a startup that has licensed (or whose founders developed) a research-validated approach at a university lab and now wants to convert it into a commercial product. The classic example is a small-team licensing relationship between a learning-sciences professor and a former-postdoc-turned-founder, in which the academic IP has demonstrated efficacy in a peer-reviewed RCT and the commercial work is now about productization: user interface, integration with school information systems, teacher-dashboard design, COPPA and FERPA compliance, and sales-readiness.

The two-year, $1 million budget is calibrated to the actual cost of commercializing a research product. It is not enough money to fund a sales team or a national marketing rollout — those are the kinds of activities a Series A would underwrite — but it is enough money to build a production-ready version of a research prototype, run a multi-district usability study, and arrive at a credible go-to-market plan that a private investor or strategic acquirer can evaluate.

The competitive bar is high. Reviewers want to see published efficacy evidence, a clear license or assignment of the underlying IP, a development plan that does not require additional fundamental research (the "direct-to-Phase-II" framing presumes the research is done), and a commercialization plan that articulates exactly how the product reaches end-users at scale.

The compliance traps

Three compliance failures account for a disproportionate share of ED/IES rejections, and all three are mechanical rather than substantive.

First, SAM.gov registration with a valid Unique Entity Identifier (UEI) listed in the proposal at the time of submission. Applications without a UEI are rejected without review. Founders who started SAM.gov registration the week before the deadline frequently discover that the registration process now takes substantially longer than the agency's documentation suggests — sometimes three weeks, occasionally more if any of the supporting documentation requires re-verification. The June 29 deadline implies an effective SAM.gov registration deadline of mid-May at the latest for first-time registrants.

Second, duplicate proposals across the three 2026 solicitations are rejected without review. A founder cannot hedge by submitting the same idea simultaneously as a Phase IA and a Phase IB application. The agency's submission system de-duplicates by content as well as by title. Submitting truly distinct projects with different objectives across multiple solicitations is permitted; submitting the same project with different framing is not.

Third, the technical narrative limit and the budget-justification format. ED/IES uses a specific page-count and section-structure convention that differs from NIH and NSF SBIR templates. Founders who reuse an NIH SBIR narrative without adapting it to the ED/IES format frequently find that critical sections are dropped to fit the page limit, weakening the proposal.

How this fits the broader SBIR moment

The ED/IES cycle closes against a backdrop of broader SBIR-program turbulence. As our coverage of the NIH SBIR/STTR omnibus hiatus documented, the NIH program is mid-reset with no active omnibus NOFOs until September 8. As our analysis of the NSF SBIR/STTR relaunch covered, NSF reopened its program in late May with a July 27 deadline. The Department of War's FY26 Release 2 closes June 24 with 42 topics. For edtech founders watching the federal SBIR calendar, ED/IES is the most direct mission match — and the only program in the FY26 cycle whose entire window for fiscal-year activity closes in a single eight-day stretch in late June.

For founders whose product strategy fits the ED/IES theory of change, the June 29 deadline is the most important date on the 2026 calendar. Applications are submitted through the federal procurement portal referenced in each solicitation. The deadline is hard. There is no second cycle.