SBIR Phase I vs Phase II: Which One to Target First

The short answer

If you've never received an SBIR before and your preliminary data is thin, start with Phase I. If you already have compelling feasibility data from non-SBIR sources and a clear commercialization path, apply Direct-to-Phase II to skip 9–12 months and roughly $314K in ceiling.

Most biotech founders get this wrong in one of two directions: they pursue a Phase I they don't need and burn a year of calendar time, or they swing at a Phase II they aren't ready for and get triaged. The goal of this article is to make the decision obvious.

Phase I vs Phase II at a glance

The caps above are NIH's hard limits for FY 2025. They're the most common ceilings across agencies, but BARDA, DoD, and NSF use their own Phase I/II caps — don't assume the NIH numbers transfer.

What Phase I actually funds

Phase I is designed to answer a single technical question: does your approach work well enough to warrant scaled investment? It's not meant to produce a finished product, run a clinical study, or generate manufacturing-grade data. It's meant to kill risky assumptions cheaply.

Good Phase I aims look like this:

• Demonstrate that your assay detects the target analyte with ≥90% sensitivity in a pilot panel of 50 samples.
• Show that your lead compound reduces biomarker X by ≥40% in a validated mouse model.
• Establish that your device prototype meets the key performance spec in bench testing.

If your aim is "prove our therapy cures disease Y," reviewers will flag it as out of scope. Phase I is about technical feasibility, not clinical validation.

Budget reality

$314K sounds like a lot until you price in indirect costs (often 40–65%), PI salary, a postdoc, reagents, and a CRO contract for a key assay. By the time you account for all of those, most Phase I budgets have room for one core experimental aim and one supporting aim — not three, not four.

Pad your budget and you'll look naïve. Under-budget and you'll look underpowered. The sweet spot is a tight scope with clear go/no-go criteria.

What Phase II actually funds

Phase II is where the real money is — and where the real scrutiny starts. $2M over 24 months is enough to:

• Run a meaningful preclinical package (efficacy, PK/PD, tox at a GLP-like tier)
• Execute a small first-in-human feasibility study for certain device categories
• Build out manufacturing processes and release assays
• Complete IP filings and pre-IND interactions

Reviewers expect three things that Phase I doesn't demand:

1. Preliminary data that de-risks your central hypothesis. Either from your own Phase I or from prior funded work. A Phase II built on speculation gets triaged.
2. A commercialization plan that doesn't read like a pitch deck. Who buys this, at what price point, through what channel, against what incumbents. Letters of intent from customers/partners help.
3. A regulatory strategy. For therapeutics: pre-IND plan and target indication. For devices: 510(k)/De Novo path. For diagnostics: LDT vs. IVD strategy. Vague regulatory language is a red flag.

Direct-to-Phase II (D2P2): when to skip Phase I entirely

NIH allows Direct-to-Phase II if you can demonstrate that the work equivalent to a Phase I has already been completed — through non-SBIR funding. This is the right move when:

• You've already done the feasibility work under a foundation grant, academic lab funding, or founder-funded research
• You have publications, preprints, or internal data that unambiguously answers the feasibility question
• You're willing to assemble a Phase I–equivalent data package as part of the D2P2 submission

D2P2 buys you approximately 9–12 months of calendar time and preserves a Phase I ceiling of $314K you can use elsewhere. Done well, it's the highest-leverage path for technically mature companies. Done poorly — without strong prior data — it fails hard because reviewers have no excuse to forgive weaknesses.

Not every institute accepts D2P2, and the list of participating ICs changes annually. Verify in the current NIH Omnibus SBIR/STTR solicitation before committing.

Fast-Track: the hybrid

Fast-Track is a single application that includes both Phase I and Phase II in one submission, reviewed together. If the reviewers score it well, Phase II funding starts immediately after Phase I ends — no gap, no resubmission cycle.

Fast-Track works well when:

• Your Phase I has clean, objective go/no-go criteria that a reviewer can evaluate on paper
• Your Phase II plan depends on Phase I's outcome but doesn't require its data to be designed
• You have the capacity to write both plans at proposal quality simultaneously

It fails when founders treat it as a shortcut. Writing a Fast-Track is harder than writing a Phase I, and a weak Phase II narrative can sink an otherwise fundable Phase I.

The five-question decision framework

Ask these in order. The first "no" tells you where to focus.

1. Do I have preliminary data that answers the feasibility question?
   • No → Phase I
   • Yes → continue
2. Is that data from non-SBIR funding and well-documented?
   • No → Phase I
   • Yes → continue
3. Do I have a concrete commercialization plan with at least one external validator (customer, partner, regulatory body)?
   • No → Phase I or D2P2 with a plan upgrade
   • Yes → continue
4. Can I deploy $2M over 24 months right now without hiring a function I don't have?
   • No → Phase I while you build capacity
   • Yes → continue
5. Does the institute I'd apply to participate in D2P2 this cycle?
   • No → Phase I (or pick a different IC)
   • Yes → D2P2 is your strongest path

If you answered yes to all five, D2P2 (or Fast-Track if you want Phase II funding guaranteed) is your move. If you hit "no" at question 1 or 2, Phase I is the right starting point — there's no shame in it, and a strong Phase I is the fastest on-ramp to a defensible Phase II.

Three common mistakes

Applying to Phase II when you should have done Phase I. Reviewers will read your "preliminary data" as speculative and give you a score that doesn't fund. You'll lose 12 months and have to resubmit anyway. If you're not sure whether your data is strong enough, it isn't.

Applying to Phase I when D2P2 would have worked. The reverse problem. You burn a year and $314K ceiling doing work you've already done. Worse: if the Phase I is too easy, reviewers score it lower because the aims aren't ambitious. Match the mechanism to the maturity of your science.

Treating Phase I as a mini Phase II. Founders cram three aims into a budget that fits one and a half. The proposal looks unfocused. The review panel sees a team that doesn't understand the scope of what they're applying for. Keep Phase I tight.

What reviewers actually score

Across all SBIR phases, the scoring criteria are the same: Significance, Investigators, Innovation, Approach, and Environment — plus Commercial Potential for SBIR/STTR. But the weight shifts between phases:

• Phase I review skews toward Innovation and Approach. Can this work? Is the path credible?
• Phase II review skews toward Approach and Commercial Potential. Will this become a product? Who will buy it? What's the path to revenue or licensing?

If you've read our earlier post on how reviewers think, the same five signals apply — but in Phase II, "significance framing" and "commercial credibility" carry more weight than in Phase I.

Sequencing against your venture timeline

For most biotechs raising capital, we recommend this sequence:

• Pre-seed → seed: SBIR Phase I on a tightly scoped feasibility question. Use it as external validation when pitching seed investors.
• Seed → Series A: SBIR Phase II (or D2P2) to fund the data package that makes your Series A pitch undeniable. A $2M non-dilutive award at this stage typically reduces dilution by 20–30% and signals third-party validation to VCs.
• Series A → Series B: Phase IIB, CRP supplements, or transition to larger federal programs (BARDA, ARPA-H, DoD) that fund at $5M–$20M+.

For more on how to weave non-dilutive and venture funding, see Non-Dilutive vs VC: How to Combine Both.

The decision matters more than the writing

Founders spend weeks agonizing over proposal language and almost no time on mechanism selection. That's backwards. A beautifully written Phase I proposal submitted by a company that should have gone D2P2 still burns a year. A mediocre D2P2 from a company with strong feasibility data will often still fund.

Pick the right mechanism. Then write it well. In that order.

If you're not sure which one fits your stage, that's exactly the conversation we have with founders on their first call with Nexa. We map your data, your timeline, and your capital plan against the current SBIR landscape and tell you which mechanism gives you the highest expected value — not the one that sounds the most ambitious.