How to Sell to DARPA

The Defense Advanced Research Projects Agency (DARPA) has a reputation for funding the impossible. A $4 billion annual budget, a hands-off management style, and a portfolio spanning AI, autonomous systems, biotechnology, and quantum computing. If you're building moonshot technology, DARPA might fund it. The catch: they don't buy finished products. They fund development of capabilities that don't yet exist.

DARPA's mandate is to prevent technological surprise. They're not interested in incremental improvements to existing systems. They want breakthroughs. This makes them fundamentally different from other federal buyers. They embrace risk, they expect failures, and they move fast—but only if you're solving a specific problem they've identified.

DARPA's Structure and Funding Model

DARPA has six technical offices: Adaptive Execution Office, Biological Technologies Office, Defense Sciences Office, Information Innovation Office, Microsystems Technology Office, and Tactical Technology Office. Each has its own multi-year research programs and solicitations.

Funding flows through research programs, not contracts to vendors. You propose a team (usually with academic partners, other contractors, and industry) to solve a multi-year research challenge. DARPA provides milestone-based funding—you hit technical gates, you get the next tranche. Miss a gate, the program stops.

Program budgets range from $10M to $100M+ depending on scope. Your company's piece might be $2M over three years, but that's still meaningful. And DARPA's emphasis on innovation over cost control means you have flexibility that traditional DoD procurement doesn't allow.

How DARPA Actually Buys

DARPA publishes "Broad Agency Announcements" (BAAs) on their website and on SAM.gov. A BAA might say something like: "We're seeking innovative approaches to autonomous swarming algorithms that can coordinate 1,000+ agents in contested environments." Then they let companies and teams propose solutions.

This is fundamentally collaborative. DARPA wants you to team with universities, national labs, and other contractors. Solo small businesses rarely win large DARPA contracts. You need partners who can handle subcontracted work.

The timeline is brutal but honest. DARPA evaluates proposals in 3-4 months and makes awards quickly. Once you're in, you need strong program management because DARPA program managers are technical experts who audit your progress quarterly. You can't fake progress.

The Key to DARPA Success: Technical Depth

DARPA reviewers are world-class researchers and engineers. If you're proposing an AI solution, your technical team needs to include people who can discuss algorithmic innovation at depth. Buzzwords and polished slides don't work here. Technical rigor does.

Proposals need to demonstrate:

• A clear technical approach with milestones and metrics
• Why existing approaches fail and why your innovation matters
• A team with proven track records solving hard problems
• Realistic timelines and budgets (they'll call you out on fantasy numbers)
• Integration paths for the military to actually use your result

The Common Mistakes

Treating DARPA like traditional DoD procurement is fatal. DARPA doesn't care about FAR compliance or NAICS codes in the way other agencies do. They care about innovation and execution.

Another mistake: over-promising and under-delivering. DARPA expects ambitious goals, but they expect you to hit them. If you miss technical milestones, even with good reasons, they'll kill the program. Plan conservatively, deliver aggressively.

DARPA doesn't want to hear why something is hard. They want to hear how you're going to do it anyway, and why no one else has figured it out yet.

From Research to Acquisition

DARPA's transition problem is real. They fund amazing research that rarely becomes operational capability. To increase your odds of transition, engage with the end user from day one. If you're solving an Air Force problem, get the Air Force involved. If it's a Navy problem, get Navy sponsorship early.

Some programs have explicit "transition partners" who committed to buying the capability if development succeeds. This is a huge advantage—you're de-risking commercialization while still in R&D phase.

What to Do This Week

Visit DARPA's website and read three BAAs in your technology area. Find the program manager for each and reach out informally—ask clarifying questions about their technical approach. This shows you're serious and gets you on their radar. Read recent DARPA papers on your topic (they publish heavily) to understand current thinking. Finally, identify potential academic or lab partners for future proposals. DARPA expects collaborative teams, so start building relationships now.