From Lab to Program of Record: What the Valley of Death Actually Costs You

Everyone in defense tech knows the phrase. "Valley of Death" gets thrown around at every conference, in every pitch deck, in every congressional hearing about defense innovation. What gets discussed far less is what the crossing actually costs — in dollars, in time, and in organizational damage that doesn't show up on a balance sheet.

Photo by cottonbro studio on Pexels.

The standard story goes like this: a startup wins an SBIR, builds a prototype, impresses some program office, and then... nothing. Funding dries up before a program of record materializes. The gap between prototype dollars and procurement dollars kills the company. Tragic, preventable, someone should fix it.

That story is true. It's also incomplete.

The Gap Is Only Part of the Problem

Yes, there's a funding gap. But the more dangerous issue is a sequencing problem that most founders — and honestly, most investors — don't fully price in before writing the first check.

Consider what actually has to happen between a successful demo and a program of record. The technology has to be mature enough for a formal requirements document. A program office has to own it, budget for it, and defend that budget through the PPBE process — which runs on two-year cycles and was designed for a world where nothing moved fast. Independent Verification and Validation (IV&V) has to happen. Legal reviews. ITAR compliance checks. Cybersecurity accreditation under RMF. A contracting vehicle has to exist or get created.

Each of those steps takes time. Combined, they routinely add three to five years to what founders model as an eighteen-month runway problem.

Three to five years. For a startup burning $800K a month.

What This Means for Investors

If you're underwriting a defense tech deal assuming the company converts its prototype contract into a POR in two years, you are probably wrong. Not because the technology isn't good — it might be excellent. Because the procurement system is not optimized for speed, and no amount of Pentagon innovation theater changes that at the program level.

This is where deal structure matters more than people admit. The startups that survive the valley aren't always the ones with the best tech; they're the ones that were capitalized for the actual timeline, not the optimistic one. That means investors need to either plan for a longer hold with bridge capacity, or ensure the company has a credible commercial revenue stream that funds operations while the government grinds forward.

Dual-use isn't just a market thesis. For many defense tech companies, it's a survival mechanism.

graph TD
    A[SBIR / Prototype Contract] --> B(Technology Maturation)
    B --> C{Requirements Alignment?}
    C -- No --> D[Pivot or Die]
    C -- Yes --> E(IV&V + Accreditation)
    E --> F(PPBE Budget Cycle)
    F --> G[Program of Record]

The diagram looks clean. The reality is that C, E, and F each have their own internal loops, delays, and stakeholders with veto power you haven't met yet.

What Founders Usually Get Wrong

Most founders underestimate the cost of staying relevant during the crossing. It's not enough to build the thing and wait. You have to actively maintain relationships with the program office, track personnel rotations (program managers turn over constantly), brief new stakeholders who weren't in the room for the original demo, and keep the technology current as requirements inevitably drift.

That's a full-time job. For a small startup, it often means a dedicated person whose entire role is government relations and program navigation — someone who has held a clearance, knows how budget lines get defended, and can speak fluent acquisition. That person costs $250K–$350K annually and doesn't write a single line of code.

Few seed-stage models account for that hire. They should.

The Investors Who Get This Right

What separates the better defense tech investors isn't just pattern recognition on technology. It's an honest accounting of time. The firms doing this well are modeling five-to-seven year paths to meaningful government revenue — not because they're pessimistic, but because they've watched enough companies fail on optimistic timelines to know what the data actually shows.

They're also building portfolios where at least some companies have near-term commercial revenue, so the fund doesn't depend entirely on the government moving at a startup's pace. It won't.

The Valley of Death is real. But it's not an unsolvable problem — it's a known cost of doing business in this market. Price it in. Staff for it. Build the runway to cross it, not just to reach it.

That's the difference between a defense tech company and a defense tech cautionary tale.