NSF vs NIH vs DoD SBIR: Which Program Fits Your Technology

The phrase “SBIR grant” covers four structurally different federal programs at NSF, NIH, the Department of Defense (DoD), and the Department of Energy (DOE). They share the name and the small-business eligibility rules. Almost everything else — the technical focus, award size, review style, timeline, and what makes a strong proposal — is different.

Picking the wrong agency is the single most expensive early-stage mistake SBIR applicants make. It costs you 4–6 months of calendar time and teaches you almost nothing generalizable, because the four programs reward different things. This article is the agency-fit decision framework we use internally at FundUnlocked before recommending which program to apply to.

The 60-second comparison

Dimension	NSF SBIR	NIH SBIR	DoD SBIR	DOE SBIR
Technical focus	Deep tech, broad commercial impact	Biomedical & health research	Defense, dual-use, national security	Energy, climate, advanced materials
Phase I award	Up to $305,000	Up to ~$314K	$50K–$250K (service-dependent)	Up to $250K
Phase II award	Up to ~$1.25M	Up to ~$2.1M	$1.25M–$2M	Up to $1.6M
Pre-submission step	Mandatory Project Pitch	None (full app direct)	Topic-specific response	Mandatory Letter of Intent
Review style	Intrinsic merit + commercial	Peer review (study sections)	Mission/operational fit	Technical + mission fit
Typical timeline to award	4–6 months	5–9 months	3–8 months	4–6 months
Deadline cadence	3–4 rolling windows/year	3 standing dates/year	Monthly topic releases	2 releases/year

How to pick the right agency in five questions

Instead of comparing programs feature-by-feature, work through these five questions in order. The first “yes” usually names your agency.

1. Does your technology have clinical or human-health endpoints?

If yes — drug discovery, medical device, diagnostic, therapy, digital health measuring clinical outcomes — you're almost certainly NIH SBIR. NIH has 20+ institutes, dedicated peer review, and the largest Phase II dollars of any agency. NSF will explicitly redirect health-application projects to NIH at the pitch stage.

2. Does your technology directly address a defense or national-security use case?

If yes — autonomous systems, defense communications, cyber, sensors with military application, dual-use hardware — DoD SBIR is usually the right door. DoD has the largest annual SBIR spend of any federal agency and the fastest review cycles, but award structures (Phase I as low as $50K) and topic-based solicitations mean you're answering a specific government-written need, not proposing open-ended research.

3. Is energy, climate, grid, or advanced materials the core innovation?

If yes — battery chemistry, fusion, solar, advanced manufacturing for energy, grid software, carbon-capture, fuel cells — DOE SBIR is the specialist. DOE's Office of Science runs the most technically-rigorous SBIR review of any civilian agency. Expect the LOI requirement and a heavier scientific rigor bar.

4. Is your innovation a broadly-commercializable deep-tech capability without a clinical, defense, or energy-specific lens?

If yes — novel algorithms, general-purpose robotics, materials without energy specificity, AI/ML infrastructure, advanced manufacturing, climate tech at the research stage — NSF SBIR is the best fit. NSF explicitly funds technology with strong commercial potential regardless of end-market, so long as the science is genuinely novel. Start with What is an NSF Project Pitch?.

5. Do you have a strong university co-investigator already?

If yes — and especially if the research is genuinely better done jointly with a non-profit research institution — STTR (at any of the four agencies) deserves a serious look. STTR allows the Principal Investigator to be employed by the research institution, which opens the door for professor-led or spinout-from-lab teams who can't meet SBIR's PI employment rule.

NSF SBIR in depth

NSF is what most people mean when they say “SBIR.” It's the open-topic program — you propose what you think is worth researching, not what the government has pre-specified.

What NSF funds

Deep-tech research with clear commercial potential. Strong fits: novel algorithms and architectures, advanced materials, robotics, cleantech R&D, climate-tech sensing, advanced manufacturing process innovations, AI/ML capabilities beyond application packaging.

What NSF does not fund

• Pure software products without a research question (“we'll build the SaaS”).
• Clinical trials or health-endpoint research (redirected to NIH).
• Defense-specific projects (redirected to DoD).
• Services businesses, business-model innovations, or market-facing features.

What makes NSF different

• Mandatory Project Pitch first. You can't apply for Phase I without pitch encouragement.
• Two-criteria review. NSF weighs intellectual merit and broader impacts equally — commercial potential is a first-order concern, not a second-order one.
• Rolling windows. No annual fixed calendar — see our 2026 NSF SBIR deadlines guide.
• Maximum Phase I award: $305,000. Higher than most agencies.

NIH SBIR in depth

What NIH funds

Biomedical and public health research — therapeutics, diagnostics, medical devices, digital health with measurable clinical endpoints, behavioral health interventions, clinical trial infrastructure, and health-related research tools. NIH is structured around 20+ Institutes and Centers (NCI for cancer, NIMH for mental health, NIA for aging, etc.); your proposal is routed to the right institute.

What makes NIH different

• No Project Pitch gate. You submit a full Phase I application directly.
• Peer review by study section. NIH's review is run by standing panels of academic and industry scientists, similar to R-series research grants. Review is rigorous and the written critiques are substantial.
• Three standing deadlines per year. January 5, April 5, September 5 — predictable calendar, unlike NSF.
• Application is long. A full NIH Phase I SBIR application is 6+ times the work of an NSF Project Pitch. Specific Aims, Research Strategy, biosketches, budget justification, data sharing plan, human-subjects protection (if applicable), etc.
• Strong commercialization expectations at Phase II. Many NIH institutes require a detailed commercialization plan and evidence of market engagement.

Who NIH fits best

Biotech startups, medical-device companies, digital-health platforms with clinical outcomes, diagnostic-tool developers. If there's a doctor, nurse, or patient in your customer segment, start here.

DoD SBIR in depth

What DoD funds

Technologies that solve defense or dual-use problems — autonomous systems, sensors, communications, cyber, materials, logistics, biotech for military medicine, and increasingly AI/ML for defense-adjacent use cases. DoD SBIR is run by service branches (Army, Navy, Air Force, Space Force, Missile Defense Agency, Defense Health Agency, DARPA, SOCOM).

What makes DoD different

• Topic-based solicitations. The government writes the topics; you propose a solution. You're answering a stated need, not proposing open research.
• Rolling monthly release cadence. Topics release on the first Wednesday of each month and typically close on the last Wednesday of the following month.
• Phase I awards run smaller. $50K–$250K depending on service branch and program.
• Shorter review cycles. Phase I decisions typically land in 3–6 months.
• Open-topic tracks exist. AFWERX and SOCOM run “Open Topic” solicitations that allow you to propose technology aligned with the branch's broader mission rather than a narrow topic description.

Who DoD fits best

Dual-use deep-tech companies (the technology also has commercial markets), companies with prior customer discovery inside the military, or founders with DoD-specific domain expertise.

DOE SBIR in depth

What DOE funds

Energy, climate, and DOE-mission research — battery and energy storage, fusion, fission, hydrogen, grid software and hardware, carbon capture, advanced manufacturing for energy, fuel cells, materials for extreme environments, high-performance computing for scientific applications.

What makes DOE different

• Mandatory Letter of Intent. No LOI on file, no application allowed.
• Fixed annual calendar. Two releases per fiscal year, each with fixed FOA / LOI / application dates — the most predictable SBIR calendar of the four major agencies.
• National-lab coupling. Many DOE proposals benefit from (or require) coordination with a National Laboratory.
• Heaviest scientific rigor. DOE's Office of Science reviewers are generally the most technically-demanding SBIR reviewers in the federal system.

Who DOE fits best

Cleantech, energy, and materials-science startups with deep technical teams and often university or National Lab ties. If you're an energy startup applying to NSF instead of DOE, you're probably giving up award dollars and leaving the wrong reviewer pool reading your work.

Want this done for you? See our $349 NSF Project Pitch service →

SBIR vs STTR: the 30-second decision

Both SBIR and STTR are run at all four agencies. The difference isn't the agency — it's the team structure.

Dimension	SBIR	STTR
Research institution partner	Optional, up to 33% (Phase I) / 50% (Phase II)	Required, minimum 30% of work
Principal Investigator	Must be primarily employed by small business	Can be employed by small business or research institution
Award sizes	Identical to STTR	Identical to SBIR
IP rights	Held by small business	Governed by allocation-of-rights agreement

Heuristic: if the research requires a faculty co-investigator whose expertise can't be moved in-house, apply STTR. If the research can live inside your company, apply SBIR — it's operationally simpler.

Common mistakes founders make choosing an agency

1. Applying to NSF for a medical device. NSF will not encourage a pitch with clinical endpoints. The redirect almost always costs you a cycle.
2. Applying to NIH with no NIH-standard research design. NIH reviewers are mostly academic scientists. A proposal without specific aims, hypotheses, and preliminary data structured the way R-grants are structured will score poorly regardless of the underlying technology.
3. Applying to DoD without reading the topic description. DoD topics are narrow. If the topic asks for a sensor that works at X temperature at Y cost, and you propose a sensor that works at Z temperature, you're out — regardless of technical brilliance.
4. Treating all four agencies as interchangeable. “Let's apply everywhere” without agency-specific framing of the same technology is a distributed waste of writing time.
5. Ignoring STTR when you have a faculty co-founder. A good STTR fit dramatically raises your odds at any agency by leveraging the institution's credibility.

Bottom line

NSF, NIH, DoD, and DOE SBIR programs serve fundamentally different missions and reward fundamentally different proposal shapes. Pick the agency that matches your technology's mission fit, not the one with the shortest calendar or highest headline number.

If your technology fits NSF's deep-tech-for-commercial-impact mission, the FundUnlocked done-for-you NSF Project Pitch service is built to ship a submission-ready draft in 48 hours, at a 43% first-round encouragement rate versus the 18% national rate. If your technology is better suited to NIH, DoD, or DOE, we can still help — reach out for a scoping conversation.

Continue with What is an NSF Project Pitch?, NSF SBIR eligibility, or NSF SBIR success rates.