Defense Startups Face 2-3 Year 'Valley of Death' Before First Revenue

A defense technology startup that wins a Small Business Innovation Research (SBIR) Phase I contract receives $50,000-$250,000 and a proof of concept. Phase II provides $500,000-$1.5 million for a prototype. But between Phase II completion and an actual production contract (Phase III), there is a 2-3 year gap with zero government funding. The startup must survive on venture capital while navigating the byzantine process of getting a program of record to adopt their technology. Most do not survive. This valley of death means the DoD systematically funds innovation and then watches it die before fielding it. Taxpayers pay for R&D that never reaches warfighters. The startups that do survive are those with enough VC funding to burn cash for years — selecting not for the best technology but for the best-funded. Anduril survived the valley of death because Palmer Luckey invested $200 million of personal wealth. Most founders cannot do that. The result is a defense industrial base that remains dominated by the same five primes (Lockheed, Raytheon, Northrop, Boeing, General Dynamics) who know how to navigate procurement but have weak incentives to innovate. The problem persists because no single organization in the DoD owns the transition from prototype to production. SBIR program managers fund R&D. Acquisition program managers buy production systems. These are different people in different organizations with different budgets and different incentive structures. The acquisition PM has zero incentive to take a risk on a startup's unproven technology when they can buy a proven (if outdated) system from a prime and not get blamed if it works poorly. Innovation adoption is a career risk with no career reward. Congress has attempted fixes — the SBIR Phase III mandate, the Rapid Innovation Fund, and various OTA (Other Transaction Authority) pathways — but they treat symptoms rather than the structural cause. The fundamental issue is that DoD budgets are allocated to programs of record years in advance, and inserting a new technology into an existing program requires re-baselining the program, which triggers oversight reviews, which no PM wants to initiate.