How Much Does a Grant Writer Cost for an SBIR? The Real Numbers, and When to Write It Yourself

Cada has written 100+ grant proposals across 30+ agencies, including NIH, NSF, NASA, and DoD.

A grant writer costs $5,000 to $50,000 for a full SBIR Phase I proposal, depending on scope and pricing model. Project-pitch-only services run $349 to $600. Hourly consultants charge $100 to $300 an hour. The spread is wide because "proposal help" covers everything from a 3-hour pitch review to a 300-hour full application.

That is the honest answer to the pricing question. The more useful question is whether you should pay any of it.

This piece gives you three things: the going rates by pricing model, the expected-value math that tells you whether a fee is worth paying, and four specific cases where the right answer is to write it yourself or wait a cycle.

We publish it because in Cada's own pipeline, 7 of 57 closed-lost deals were lost on price or financial qualification, the single largest interpretable loss category we have. Founders are not confused about whether they need help. They are confused about what it should cost and whether the math works.

How much does a grant writer cost for an SBIR proposal?

Here are the going rates by pricing model.

Treat those bands as directional. They are compiled from the handful of providers who publish prices on their own sites and from 2026 third-party rate surveys, and the surveys have a problem worth naming out loud: several of the widely circulated price comparisons are published by firms that appear in the comparison. Most firms in this market publish nothing at all. When I checked provider sites directly, the common answer was "request a quote."

For contrast, general nonprofit grant writers run $40 to $150 per hour. That is why quotes on general grant writing sites read as suspiciously low next to SBIR work. They are not pricing the same job.

One exception is worth naming because it is rare. SBIR Grant Writers publishes fixed prices on its own site: as of August 2026, $495 for an NSF Project Pitch, $5,995 for an NSF SBIR Phase I, and $5,495 for an NIH SBIR Phase I. Almost nobody else does that.

Two things to notice about this table.

The top and the bottom are not the same product. $349 buys a Project Pitch of a few hundred words. $25,000 buys a full Phase I application with agency-specific strategy behind it. Comparing the two numbers tells you nothing.

And essentially nobody publishes a win rate next to the price. You can sometimes find out what a proposal costs. You cannot find out what it buys. If you want the payment structures compared side by side, we broke those down in flat fee vs success-based grant writing.

What do SBIR grant writers actually do?

An SBIR Phase I proposal takes 150 to 450 hours, according to published effort estimates from SBIR technical assistance providers. First-time applicants are commonly advised to budget 10 to 15 hours per week for three months, and to start 6 to 8 weeks before the deadline at minimum.

Those hours split into three buckets, and they are not worth the same money.

Bucket 1: only you can do it. The science. The commercialization story. Letters of support. Your PI's time. Roughly 40% to 50% of the hours, and no consultant can take them off your plate. If a firm says they can, they are planning to write filler, and you are the one who signs it.

Bucket 2: anyone competent can do it. Budget justification, biosketch formatting, SAM.gov and SBA registration, forms, compliance checklists, page-limit surgery. Real work, genuinely annoying, and the cheapest thing here to buy.

Bucket 3: judgment. Which agency. Which topic. Which framing. Whether the application is competitive enough to file at all. This is the expensive part, it is maybe 10 hours of actual work, and it is the part most low-cost services leave out.

Most founders overpay for bucket 2 and underbuy bucket 3.

Ask for quotes broken out by bucket. A firm that cannot say which hours are theirs and which are yours has not thought about your application yet.

Can you pay a grant writer with SBIR funds?

Not for writing the proposal. Two separate rules combine here, and almost nobody leads with either one.

Preparing the application is a pre-award cost. The hours spent writing the proposal happen before there is an award to charge them to, and they are not recoverable from the award the proposal wins. Consulting performed during the funded project is a different question. That can be allowable when it sits in your approved budget and is priced for the work rather than the outcome.

Contingent fees are unallowable either way. Which rule applies to you depends on how your award is issued, and that varies by agency. NIH and NSF SBIR awards are typically grants. Much of DoD SBIR runs as contracts.

If your award is a grant, 2 CFR 200.459 governs. Professional and consultant service costs are allowable when reasonable "and when not contingent upon recovery of the costs from the Federal Government."

If your award is a contract, FAR 31.205-33(b) uses almost the same language: consultant costs are allowable "when not contingent upon recovery of the costs from the Government." FAR 52.203-5, the Covenant Against Contingent Fees, goes further. It lets the government annul the contract without liability, or recover the full amount of the contingent fee.

The Grant Professionals Association and the Association of Fundraising Professionals both hold that commission-based grant fees are unethical, and both note that awards can be revoked over them.

One genuine uncertainty is worth flagging. FAR 52.203-5 carves out "bona fide" employees and selling agencies, so it is not a blanket ban even on the contract side. Whether a success fee paid entirely from your own non-federal cash is permissible is unsettled. What is not in dispute is the part that hits your bank account: the fee cannot be charged to the award. Ask your counsel, not your grant writer.

So every dollar you pay someone to write the proposal is your own cash, spent before you know the outcome, at Phase I success rates that hover around 17% across agencies. NIH runs roughly 15% to 18% on a three-year average through FY2025. NSF is the most favorable of the 11 SBIR agencies at around 20%.

That is not a fee. That is a bet.

One caveat if this is your first application: those are all-applicant rates. Published agency analyses consistently note that first-time proposers see materially lower rates than repeat proposers, though almost nobody publishes the split as a hard number. Assume your odds sit below the headline.

For scale on the other side of the bet, SBA guidelines as of April 2026 let agencies issue a Phase I award up to $323,090 without SBA approval. Most agencies set lower internal ceilings, so check the solicitation rather than the guideline.

Success fee vs upfront: what the pricing model tells you

The pricing model tells you what the firm optimizes for.

Pure upfront. The firm gets paid whether you win or lose, so volume is the business model. The tell is whether they ever turn anyone away. A firm that accepts every applicant who can pay is selling you a document, not a decision.

Pure success fee. Cleanest-sounding, most legally fraught, and structurally biased. A firm paid only on wins selects for applicants who were already likely to win. Good for you if you are strong, useless if you are not, because you will simply be declined without being told why.

Small upfront plus a success component. Closest to aligned incentives, and where this market is heading. It also forces the firm to filter honestly, because it now shares your downside.

One question cuts through all of it: what percentage of inbound applicants do you turn away, and for what reasons? A firm that turns away nobody is selling volume. A firm that cannot answer has never counted.

Is hiring a grant consultant worth it? Run this math first

The rule is one line:

Hiring is worth it when (your win probability with help - your win probability alone) x award size > the fee.

Work it through with a fictional example. Say a four-person startup building a ceramic filtration membrane for industrial water reuse. Pre-revenue, one provisional patent, no publications, no prior federal award. They are targeting a $300,000 Phase I.

Alone, call their odds 10%. That sits below the roughly 17% cross-agency average, which is fair for a first-time applicant with no publication record.

With genuinely good help, call it 18%. An 8-point improvement.

Eight points times $300,000 is $24,000 of expected value. Against a $20,000 fee, that is marginally positive. If their solo baseline was already 15% rather than 10%, the same $20,000 fee is a losing bet.

If you have not worked out whether you are competitive at all, do that first. Our 7-point grant readiness assessment is the faster version of that question, and it is free.

Here is the counterintuitive part, and it is the part worth internalizing.

Paid help is worth the most to applicants in the middle. Strong applicants would have won anyway, so the delta is small. Weak applicants stay unfunded either way, so the delta is small. The value curve peaks in the middle, where a competitive application is being undermined by fixable problems.

Two honest caveats. Nobody publishes credible data on that delta, including Cada, so the 10% and 18% are illustrative rather than measured. And the math is sensitive: move the solo baseline 5 points and the decision flips. Run it with your own estimate, and be pessimistic.

When is a DIY SBIR application the right call?

Four cases where the honest answer is to write it yourself, or not to apply at all.

1. You are below the readiness floor. No SAM.gov registration, no PI who meets the primary employment test, no prior art check. Money spent on writing does not fix a missing foundation. In Cada's own scoring methodology, missing publications costs an NIH applicant 15 points and a missing academic partner costs 10, and no amount of prose recovers those. Fix the foundation, apply next cycle.

2. You are applying to a pitch-first program. An NSF Project Pitch is a few hundred words, and roughly 30% to 40% of pitches get encouraged to submit a full proposal. Paying $349 to $600 to outsource it costs you the cheapest, fastest signal you will ever get about your own fit. Write it yourself and read the response carefully.

3. You have a strong technical writer on the founding team. If a co-founder has written funded federal proposals before, buy 4 to 6 hours of review, not 200 hours of drafting. At $100 to $300 an hour that is $400 to $1,800, and it catches most of what a full engagement would.

4. The fee is more than about 10% of your remaining runway. At these win rates, a bet that large is a solvency decision, not a funding strategy. If a $25,000 fee takes you from 9 months of runway to 7, the answer is no regardless of how competitive you are.

Three cases where hiring reliably pays for itself

Notice what they have in common.

In all three you are buying judgment, not typing. That is the only version of this purchase that reliably pays for itself.

Four questions to ask before you sign

1. What percentage of applicants who approach you do you turn away, and why?
2. Which of the three work buckets are you doing, and which am I still doing?
3. Is any part of this fee contingent on an award? If so, how do you handle 2 CFR 200.459 and FAR 31.205-33(b)?
4. What is your win rate with first-time applicants specifically, and how many applications is that based on?

Question 4 is the one that produces evasion. A firm with a real number gives it to you with the denominator attached. "We have a very high success rate" without a count is not an answer, and the evasion is your answer.

So should you hire a grant consultant?

The question is not really how much a grant writer costs for an SBIR. The published SBIR grant writer fees are in the table above. The question is whether the fee buys enough additional probability to be worth your own cash.

Run the three checks in order. Are you above the readiness floor? Is the fee under 10% of your remaining runway? Does the probability delta times the award beat the fee?

If any one is a no, write it yourself or wait a cycle. You will lose less by waiting than by filing an application that was never competitive.

If all three are yes, buy judgment rather than typing, and ask the four questions above before you sign.

The most valuable thing we do at Cada is tell founders which programs they are actually competitive for. If you want a straight answer on whether you sit in the middle band where paid help pays for itself, we run a free 15-minute assessment call.

Bring your technology in two paragraphs, your rough stage, and the deadline you are working toward. You leave with a read on your competitiveness and a shortlist of programs that fit. No pitch, no obligation, nothing to pay.

Sometimes the outcome of that call is that you should not apply this cycle, and we will say so. That answer is free too, and it is worth more than a proposal you were never going to win.

Sources

• 2 CFR 200.459 -- Uniform Guidance treatment of professional and consultant service costs on federal grants, including the prohibition on costs contingent upon recovery from the Federal Government
• FAR 31.205-33 -- professional and consultant service costs on federal contracts, and the parallel contingency language in paragraph (b)
• FAR 52.203-5 -- Covenant Against Contingent Fees, including the government's right to annul the contract or recover the fee
• SBIR.gov -- SBIR/STTR Policy Directive, the inflation-adjusted Phase I award guidelines, and the list of 11 participating agencies
• NSF SBIR/STTR (America's Seed Fund) -- Project Pitch mechanism, Phase I solicitation, and program success reporting
• NIH SEED -- NIH small business program structure and published success rate reporting
• Grant Professionals Association and the Association of Fundraising Professionals -- both codes of ethics prohibit percentage-based and commission-based compensation for grant work
• SBIR Grant Writers -- one of the few providers publishing fixed prices on its own site; the $495, $5,995, and $5,495 figures are taken from that page
• Rate bands, the three-bucket hour split, the applicant-filter conditions, and the closed-lost figures are Cada's own analysis and internal pipeline data (100+ proposals across 30+ agencies)

Rate bands move and most providers publish nothing, so treat every figure here as directional and get quotes in writing. Success rates and Phase I award ceilings are set per agency and per solicitation, so verify against the solicitation you are applying under. Nothing here is legal or tax advice, and contingent-fee arrangements in particular require counsel rather than a blog post. All company examples are fictional and used for illustration only.