DOD SBIR vs. BAA vs. CSO: Defense Funding Vehicles Explained for Startup Founders

If you are a founder building dual-use technology, you have probably searched "defense grants for startups" and learned that the Department of Defense funds early-stage companies. That part is true. What nobody tells you is that "DOD funding" is not one thing. It is at least five different vehicles, and picking the wrong one can cost you 40 to 120 hours on an application you were never positioned to win.

The most common confusion is the DOD SBIR vs BAA difference. Here is the short version: a DOD SBIR is a structured, topic-based contract for small businesses doing early feasibility work. A Broad Agency Announcement (BAA) is an open research solicitation, run mostly by DARPA, that anyone can answer and that funds bigger, riskier R&D. A Commercial Solutions Opening (CSO) is a fast track for adapting a commercial product to a defense need.

Same agency. Three completely different games.

This guide breaks down all five DOD contract vehicles startups actually use -- SBIR, STTR, BAA, CSO, and the Other Transaction (OT) agreements CSOs produce. You get the comparison table, a map of which DOD component uses which vehicle, and a decision framework to pick the right entry point for your stage, your technology, and your timeline.

The one distinction founders miss: grant vs. contract vs. other transaction

Before you compare programs, understand the legal instrument underneath them. This is the layer that determines your intellectual property rights, your audit burden, and how fast you get paid.

Most founders ask "is this a grant?" That is the wrong first question for DOD. At the Department of Defense, an SBIR award is a contract, not a grant. BAAs can produce contracts, grants, cooperative agreements, or other transactions depending on the solicitation. CSOs almost always produce an Other Transaction agreement. The instrument, not the program name, drives your obligations.

Why it matters in practice:

• Contracts (most DOD SBIRs) come with federal acquisition rules, defined deliverables, and accounting expectations. You are selling the government a result.
• Grants and cooperative agreements (some BAAs, most civilian agencies) fund research with lighter deliverable structure and different cost rules.
• Other Transactions (CSOs, some DARPA programs, ARPA-H) are milestone-based agreements with negotiable IP and data rights. They sit outside the standard acquisition system, which is exactly why they move faster.

If you remember one thing: the same dollar amount feels very different depending on whether it arrives as a contract, a grant, or an OT.

DOD SBIR vs BAA vs CSO: the five vehicles at a glance

Here is the structured comparison. Award figures are typical ranges, not guarantees, and they vary by component and topic.

Read the table top to bottom and you can see the progression: SBIR is the structured on-ramp, BAAs fund the hard science, and CSOs buy commercial speed.

DOD SBIR and STTR: the structured on-ramp

A DOD SBIR is a federal contract that pays a small business to prove out a specific technology the military wants. It runs in phases, it is tied to published topics, and it is the most common first federal dollar for a defense-tech startup.

Eligibility is strict but simple. You must be a for-profit small business with fewer than 500 employees and more than 50% ownership by US citizens or permanent residents. (Source: SBIR.gov program policies.)

The phases:

• Phase I (feasibility). Across DOD, Phase I awards typically run $150,000 to $295,000 over 6 months. The Air Force's AFWERX Open Topic is the low-end outlier: its Phase I caps at $75,000 over a 3-month window focused on customer discovery, not deep R&D. (Source: AFWERX program pages, 2026.)
• Phase II (prototype). DOD Phase II awards typically run $1.25 million to $2 million over 18 to 24 months. AFWERX Open Topic Phase II caps at $1.25 million over up to 21 months. (Source: AFWERX, 2026.)
• Phase III (commercialization). There is no SBIR set-aside money in Phase III. The point is to win follow-on production or services dollars from a defense customer, often through sole-source authority your Phase I and II earned you.

If you already have feasibility data, you can skip ahead. Direct-to-Phase-II (D2P2) lets qualified companies enter at Phase II, with AFWERX D2P2 awards up to $1.25 million.

One rule that surprises first-timers: you cannot outsource the work. For SBIR, the small business must perform at least two-thirds of the work in Phase I and at least half in Phase II. (Source: SBIR.gov policies, 15 USC 638.)

STTR is the SBIR sibling for university spinouts. It requires a formal partnership with a research institution, and the work split is fixed by statute: at least 40% by the small business and at least 30% by the research institution. If your core IP sits with a faculty founder or a university lab, STTR is usually the cleaner path than SBIR.

DOD SBIR and STTR topics are posted and submitted through the Defense SBIR/STTR Innovation Portal (DSIP) at dodsbirsttr.mil, in numbered cycles like 26.1 and 26.2. You will run into two topic styles: Open Topics (sector-agnostic, you bring the idea, used heavily by AFWERX) and Specific Topics (a named component need you must match precisely).

DARPA BAAs: funding for breakthrough R&D

A Broad Agency Announcement is an open solicitation for research, authorized under Federal Acquisition Regulation 35.016. DARPA is the agency most associated with BAAs, and unlike SBIR, a BAA is open to all proposers, not just small businesses. Universities, large primes, and startups compete in the same pool.

There are two flavors, and the difference matters for your timing:

• Program-specific BAAs target one technical challenge. A DARPA program manager writes the BAA around a new program's vision, technical areas, milestones, and funding. These have defined deadlines and often a proposers' day event.
• Office-wide BAAs are standing invitations that stay open for 12 months or more, covering a DARPA office's broad interests, with rolling submission.

Funding is bigger and the instrument is flexible. BAA awards commonly range from $500,000 to $5 million or more per performer, and the award can be a contract, a grant, a cooperative agreement, or an OT. DARPA has also introduced an accelerated pathway for proposals under $2 million, with awards made within about 30 days of selection. (Source: DARPA BAA program pages, 2026.)

Here is the part that trips up SBIR veterans. A DARPA BAA is not scored against a fixed topic by a panel chasing a checklist. You are convincing a program manager that your idea advances their program. The evaluation weighs technical merit, relevance to agency priorities, and available funding. Most DARPA offices want an abstract or white paper first, and they will tell you whether a full proposal is worth your time.

Practical takeaway: if your technology is genuinely high-risk and high-payoff, and you can write a research narrative aimed at one expert rather than a committee, a BAA can fund more, faster, than a comparable SBIR.

DIU CSOs and OT agreements: the commercial-speed path

A Commercial Solutions Opening is a solicitation method the Defense Innovation Unit (DIU) pioneered to buy commercial technology at commercial speed. A CSO is how DIU finds you. The resulting agreement is almost always an Other Transaction (OT).

This vehicle exists for a specific kind of company: one with a working commercial product that can be adapted to a defense use case. It is built for nontraditional defense contractors, the companies that would never survive a traditional acquisition cycle.

The speed is the headline. DIU often awards prototype OT agreements within 60 to 90 days of selection. (Source: DIU program documentation.) Compare that to the 6-plus months a typical SBIR cycle takes from close to award.

The terms are why founders like OTs:

• Intellectual property, data rights, and payment milestones are negotiable, not dictated.
• There is no requirement for an approved government cost accounting system.
• DIU OT agreements operate outside the FAR acquisition framework, so standard GAO bid protest procedures do not apply, which removes a common source of delay.

If you already sell a real product and your problem is "how do I get a defense customer," a CSO is usually a better use of your time than starting at SBIR Phase I feasibility.

Which DOD component uses which vehicle?

"DOD" is a dozen organizations with different front doors. Matching your technology to the right component is half the battle. Here is the practical map.

If you take nothing else from this section: AFWERX is the widest startup door, DARPA is the deepest research door, and DIU is the fastest commercial door.

DOD SBIR vs BAA vs CSO: which vehicle is right for your startup?

Use this decision framework. It keys on three things: your company stage, your technology maturity, and your timeline.

1. Pre-prototype, small business, want a structured entry point. Start with an SBIR Phase I that matches an open topic. If you already have feasibility data, pursue Direct-to-Phase-II instead and skip the small first check.
2. University spinout with faculty-held or lab-licensed IP. Use STTR, where the 40/30 work split is built for a small-business-plus-institution team.
3. High-risk, breakthrough R&D you can frame as a research narrative. Target a DARPA BAA, especially an office-wide BAA you can answer on a rolling basis. Expect to write for a program manager, not a topic checklist.
4. Mature commercial product to adapt for a defense user, and you want speed plus IP retention. Go for a DIU CSO that produces an OT agreement.

Here is the honest part. Most defense-tech startups should not pick just one. A common strong move is running an AFWERX SBIR for structured non-dilutive cash while simultaneously pursuing a DIU CSO for a faster customer relationship. The vehicles are not mutually exclusive, and the right portfolio depends on your specific technology and runway.

Consider a fictional example. A four-person computer-vision startup with a deployed commercial product and 9 months of runway is usually a poor fit for an SBIR Phase I feasibility study, because feasibility is the thing they have already proven. The same company is a strong fit for a DIU CSO, where their commercial traction is the whole point. Pick the vehicle that matches what you have, not the one you heard about first.

The SBIR lapse that proved why vehicle choice matters

Here is a real strategic lesson from the last year, and it is the reason vehicle diversification is not just optimization.

The statutory authority for the SBIR and STTR programs expired on September 30, 2025. For six months, agencies could not issue new SBIR awards. The programs were finally reauthorized on April 13, 2026, when the Small Business Innovation and Economic Security Act of 2026 extended them through September 30, 2031. (Source: Congress.gov; Crowell & Moring client alert, 2026.) If you are timing a DOD SBIR application around the restart, our DOD SBIR 2026 reopening guide maps which components reopened first.

Now look at what kept running during that six-month lapse. BAAs, CSOs, and OT agreements operate under separate legal authorities from the SBIR set-aside. They did not stop. A founder who had bet their entire federal strategy on SBIR was frozen for half a year. A founder with a DIU CSO in motion kept moving.

The lesson is portfolio resilience. Relying on a single vehicle is a single point of failure. Understanding all five gives you options when one path stalls, which is exactly what a multi-program strategy is supposed to buy you.

Common mistakes founders make with DOD vehicles

• Treating a DOD SBIR like a grant. It is a contract. The accounting expectations and deliverable structure surprise founders who expected grant-style flexibility.
• Forcing a fit with a Specific Topic. If your technology does not cleanly match a named component need, an Open Topic (where available) is usually the better path than contorting your pitch.
• Writing a DARPA BAA like an SBIR topic response. A BAA wants a research narrative aimed at a program manager, not a checklist answer to a fixed topic.
• Pitching a CSO with a TRL-3 prototype. CSOs reward commercial-ready technology. If you are still proving feasibility, you want SBIR, not a CSO.
• Ignoring component fit. Pitching maritime technology to AFWERX or pitching an early lab result to DIU wastes everyone's time. Match the component to your mission and maturity first.

Pick the vehicle before you write the proposal

The DOD SBIR vs BAA difference, and the CSO and OT paths alongside them, comes down to a simple idea: match the vehicle to what your company actually is right now. A pre-prototype team and a commercial-ready product should not chase the same dollar.

Getting this wrong is expensive, because the wrong vehicle is 40 to 120 hours you do not get back. Getting it right means your first federal application is one you were positioned to win.

Cada's grant roadmaps include DOD program matching across all vehicle types -- SBIR, STTR, BAA, CSO, and OT -- alongside civilian and state programs. If you are not sure which DOD funding path fits your technology, that is the first question to answer, before you invest 80 hours in an application. Book a free consultation and we will map your technology to the right vehicle. No pitch, no obligation.

Cada has written 100+ grant proposals across 30+ agencies, including the DOD components above. Award amounts, timelines, and eligibility rules change between cycles -- always verify the specifics against the current solicitation on DSIP or the issuing component before you apply.