A grant roadmap turns the overwhelming landscape of 11+ federal agencies and hundreds of SBIR topics into a clear, sequenced plan targeting 5-10 best-fit programs across 3-5 agencies. The portfolio approach nearly triples win probability (from 20% on a single application to 49% across three), and each additional proposal costs less because research and positioning transfer. That shift -- from one-shot gamble to portfolio strategy -- is the single biggest factor in whether a startup wins non-dilutive funding.
Most startups either apply to one program and hope, or get paralyzed by the 11+ agencies and hundreds of topics and do nothing. A roadmap fixes both problems.
What is a grant roadmap and what should it include?
A grant roadmap is a prioritized, company-specific list of funding opportunities ranked by fit, competitive position, and strategic value. It's not a generic list of "grants for startups." It's an assessment of which programs your specific technology is most likely to win, given your stage, team, IP, and competitive landscape.
A good roadmap includes:
- 5-10 programs across multiple agencies, ranked by recommendation tier
- Agency and program details -- award amounts, timelines, eligibility, deadlines
- Company-specific fit analysis -- why each program fits your technology, what framing to use, what concerns to address
- Competitive positioning -- how your proposal would be evaluated relative to the field
- Action items -- the specific steps you need to take for each program (register on SAM.gov, develop preliminary data, find a university partner, etc.)
- Sequencing -- which programs to apply for first, second, third based on timeline alignment and strategic value
What it's not: a database dump. A roadmap that lists 30 programs without telling you which 5 to focus on is just noise.
Why does a portfolio approach beat single-shot applications?
The math on single-grant applications is brutal. SBIR Phase I success rates range from 15-25% depending on agency. That means a 75-85% chance of rejection on any single application.
But grant applications across agencies are largely independent events. If you submit to 3 agencies with a 20% success rate each:
| Strategy | Probability of at least 1 win |
|---|---|
| 1 application (20% each) | 20% |
| 2 applications (20% each) | 36% |
| 3 applications (20% each) | 49% |
| 5 applications (20% each) | 67% |
Going from 1 application to 3 nearly triples your chances. Going to 5 gives you better-than-even odds. This is the portfolio effect -- and it's why the roadmap approach works. At Cada, every client engagement starts with a roadmap that identifies 5-10 programs precisely because the math is this clear.
How do you assess which programs fit your technology?
Not every program that accepts your technology category is worth applying to. Program fit depends on five factors:
1. Technology alignment. Does the program fund the type of R&D you're doing? NSF funds fundamental research. AFWERX funds defense applications. NIH funds health outcomes. Same technology, completely different framing.
2. Stage match. Are you at the right TRL for this program? SBIR Phase I funds feasibility (TRL 2-4). DIU funds commercial prototypes (TRL 5+). DARPA BAAs fund breakthrough concepts (TRL 1-4). Mismatched stage = wasted application.
3. Competitive position. Can you actually win? If a program historically funds university labs and you're a 3-person startup with no publications, your odds drop regardless of technology fit. Conversely, some programs explicitly favor commercial companies -- that's where you compete best.
4. Strategic value. What does winning this grant unlock beyond the money? An NSF SBIR signals scientific rigor to investors. A DoD Phase I opens the door to Phase III procurement. An ARPA-H award signals you're working on something transformative.
5. Practical feasibility. Can you actually submit a competitive application? Some programs require a university partner (STTR), preliminary data (NIH), or a customer letter of intent (AFWERX). If you can't meet these prerequisites in the application window, the program isn't ready for you yet. It might be in 6 months.
The agency matching framework (stage + sector + technology)
Use this decision matrix to narrow from 11+ agencies to your top 3-5:
| If your startup is... | Primary agencies | Secondary agencies |
|---|---|---|
| Health/biotech, pre-clinical | NIH, ARPA-H | NSF (if fundamental science), BARDA |
| Health/biotech, clinical stage | NIH, ARPA-H, BARDA | DoD (military health), VA |
| AI/ML, general purpose | NSF, DARPA | NIH (if health AI), DOE (if energy AI) |
| AI/ML, defense applications | DARPA, AFWERX, DIU | Navy, Army SBIR |
| Clean energy / climate | DOE, NSF | EPA, USDA (if agricultural) |
| Defense / dual-use hardware | AFWERX, Navy, Army, DARPA | DIU (if TRL 5+), SOCOM |
| Agricultural / food tech | USDA, NSF | EPA, DOE (bioenergy), FFAR |
| Cybersecurity | DHS, DARPA I2O | NSA, CISA, DoD cyber |
| Space technology | NASA, SpaceWERX | DARPA STO, NRO, SDA |
| Education technology | NSF, ED | IES |
| Cross-border (US + Canada) | NSF + IRAP, NIH + CIHR | SR&ED tax credits |
Sequencing submissions: what order and why
Order matters. The best sequencing strategy considers three factors:
1. Lead with your best fit. Your first application should target the agency where your technology and team are most competitive. A strong first win builds momentum, credibility, and reusable content for subsequent applications.
2. Align with deadlines. SBIR deadlines vary by agency. NSF has quarterly submission windows. NIH has three annual deadlines. DoD varies by component. Map your applications to the calendar so you're not writing three proposals simultaneously.
3. Use early applications to build later ones. An NSF Project Pitch (3,000 characters, 3-week turnaround) is the lowest-effort entry point. If you get invited, the full NSF proposal becomes your "anchor" document that you can adapt for NIH, DoD, and other agencies. Start with the program that gives you the fastest feedback.
Recommended sequencing for most startups:
| Order | Action | Timeline | Why |
|---|---|---|---|
| 1 | NSF Project Pitch | Week 1-2 | Fastest feedback (3 weeks). Low effort. Validates whether NSF is a fit. |
| 2 | DoD SBIR topic search | Week 2-4 | Identify matching topics across AFWERX, Navy, Army. Deadlines are fixed. |
| 3 | First full proposal (best-fit agency) | Week 4-10 | Your anchor document. Invest the most effort here. |
| 4 | Second proposal (adapted from first) | Week 8-14 | Reuse 40-60% of the anchor. Different agency framing. |
| 5 | Third proposal or DARPA/ARPA-H | Week 12-20 | If applicable. These have different formats but benefit from prior research. |
Timeline expectations by agency
| Agency | Time to decision | Submission frequency | Notes |
|---|---|---|---|
| NSF | 3 weeks (pitch), 6-9 months (full) | Quarterly windows | Pitch first, then proposal if invited |
| NIH | 9-12 months | 3 deadlines/year (Jan, Apr, Sep) | Longest timeline. Plan accordingly. |
| AFWERX | 90-120 days | Continuous open topic | Fastest DoD SBIR |
| Navy SBIR | 4-6 months | Annual + periodic topics | Specific topic matching required |
| Army SBIR | 4-6 months | Annual + periodic topics | Specific topic matching required |
| DARPA | 6-18 months | BAA-specific | White paper first |
| ARPA-H | 9-15 months | Rolling ISOs | Solution Summary first |
| DIU | 3-6 months | CSO-specific | 60-90 day award target |
| DOE | 6-12 months | Annual solicitation | LOI often required |
| USDA | 6-9 months | Annual solicitation | Narrower topic areas |
What a real grant roadmap looks like (anonymized example)
Here's a simplified version of an actual Cada Partners roadmap we built for a health AI startup (details changed):
| Program | Agency | Recommendation | Award | Why It Fits |
|---|---|---|---|---|
| SBIR Phase I | NSF | Strongly Recommend | $305K | Core AI platform fits NSF's "Intelligent Systems" topic. Team has PhD in ML. Low barrier -- start with Project Pitch. |
| SBIR Phase I | NIH (NIBIB) | Recommend | $314K | Clinical validation angle. Needs preliminary data from 2 sites. Apply after generating pilot results. |
| Open Topic | AFWERX | Recommend | $75K | Military health screening application. Small award but fast timeline and opens DoD door. |
| ISO (PHO) | ARPA-H | Consider | $1M-$5M | Fits "early detection" mission. High bar -- needs stronger clinical narrative. Worth exploring after NIH submission. |
| BAA | DARPA (BTO) | Watch | $500K-$2M | BTO occasionally funds health AI. Monitor for relevant BAA. Not ready to apply yet. |
The startup applied to NSF first (invited after pitch), then AFWERX (awarded Phase I), then NIH. Three shots on goal within 6 months, two of which resulted in awards.
The 3 tiers: Apply Now, Prepare, and Monitor
Every roadmap categorizes programs into three tiers:
Apply Now (Strongly Recommend / Recommend) Programs where your technology fits, you meet eligibility requirements, and you can submit a competitive application within the next 3-6 months. These are your primary focus.
Prepare (Consider) Programs that are a good fit but require preparation -- developing preliminary data, finding a university partner, getting SAM.gov registered, building a relationship with a program manager. These become "Apply Now" after you complete the preparation steps.
Monitor (Watch / Defer) Programs that are relevant but either not currently open, require a capability you don't have yet, or are lower priority given your current resources. Review these quarterly -- they may move to "Prepare" as your company evolves or new solicitations open.
What roadmap mistakes do founders make most often?
1. Targeting too many programs. A roadmap with 15 programs sounds impressive but leads to shallow applications. Focus on 5-7 where you can submit genuinely competitive proposals. Quality beats quantity.
2. Applying to the wrong agency. NSF funds science. NIH funds health outcomes. DoD funds defense applications. The same AI technology needs completely different framing for each. Submitting an NIH-framed proposal to NSF wastes everyone's time. See our agency decision guide for help matching.
3. Ignoring competitive positioning. "Good technology fit" is necessary but not sufficient. If a program historically funds academic labs and you're a startup with no publications, your positioning is weak regardless of technology fit. A good roadmap accounts for who you're competing against.
4. Treating all programs as equal. A $305K NSF Phase I and a $75K AFWERX Phase I require similar effort to apply for. But the NSF award is 4x larger and carries more investor-signal value. Prioritize by total strategic value, not just "can we apply?"
5. Not sequencing. Writing three proposals simultaneously means all three are mediocre. Sequence them so each builds on the last, with the highest-probability program first.
6. Forgetting prerequisites. SAM.gov registration takes 2-6 weeks. NIH eRA Commons setup takes 1-2 weeks. University partnership agreements take months. Start these early -- they're not optional.
7. Stale roadmaps. Programs open and close. Award amounts change. Agencies shift priorities. A roadmap from 6 months ago needs updating. Review quarterly.
When does a grant roadmap not make sense?
- Your company has no R&D component (pure services, marketplace, or distribution)
- You're a SaaS company with no novel technology (government grants fund R&D, not product development)
- You need funding in the next 30 days (even the fastest programs take 3+ months)
- You're not willing to invest 40-80 hours in proposal writing per application
- Your technology is commercially proven and you just need growth capital (try venture debt or revenue-based financing instead)
Grants work best for companies with genuine technical innovation, a clear R&D question to answer, and 3-12 months of runway to wait for funding decisions. If that describes you, a roadmap is the most efficient way to navigate the landscape.
How to maintain and update your roadmap over time
A roadmap is a living document. After the initial build:
Monthly: Check for new solicitations from your target agencies. SBIR topics are released on rolling schedules. DARPA BAAs appear unpredictably. Set up alerts on SAM.gov and agency-specific portals.
Quarterly: Review your full roadmap. Update program statuses (open/closed/paused). Reassess your competitive position -- new publications, new data, new team members can change which programs you're competitive for. Move programs between tiers as your company evolves.
After each submission: Win or lose, update the roadmap. A rejection with reviewer feedback tells you where to improve. A win opens new follow-on pathways (Phase II, Phase III, related programs). Both events change your roadmap.
After major company milestones: Raising a round, publishing a paper, getting a patent, signing a customer -- all of these change your competitive positioning and may make previously marginal programs worth pursuing.
Ready to build your grant roadmap?
We build personalized grant roadmaps for tech startups across every sector and stage. Our Strategy Review is the first step -- a 15-minute conversation where we assess your technology, identify the 3-5 strongest programs, and tell you honestly whether grants make sense for your company.
For a deeper dive into which agency is the best starting point, read our agency decision guide. For an overview of all non-dilutive funding options beyond SBIR, see our non-dilutive funding guide.