The defense technology sector is experiencing a capital surge that would have been unthinkable a decade ago. Anduril Industries recently doubled its valuation to $14 billion, while hydrogen-powered drone startup Mach Industries quadrupled its own valuation to $1.6 billion in just over a year. The U.S. Department of Defense is proposing a 40% increase in its base budget for fiscal year 2025, pushing total national security spending toward $1 trillion.
Yet Ross Fubini, the managing partner at XYZ Venture Capital who wrote Anduril’s first check in 2016, warns that the current wave of enthusiasm is creating a dangerous bottleneck. In a recent interview with PitchBook, Fubini said the majority of today’s defense tech startups will get lost in what the industry calls the “Valley of Death” — the gap between winning a small prototype contract and scaling to full-rate production.
Also read: Cruise Giants Face $440M Liability After Supreme Court Rules on Seized Cuban Property
The Valley of Death is widening
The Valley of Death is not a new problem. For decades, the Pentagon has struggled to transition promising technologies from research labs into fielded systems. What has changed, according to Fubini, is the sheer volume of companies entering the funnel. The Defense Innovation Unit now processes thousands of proposals annually, but the number of contracts that lead to sustained production remains low.
Fubini argues that many founders underestimate the gap between a $2 million prototype award and the $100 million-plus production contracts required to build a real business. “The government is not a venture capital firm,” he said. “They buy products, not potential.”
Also read: Three Major Hospital Systems Accuse CVS of Secretly Siphoning Hundreds of Millions in Drug Savings
Capital is not a strategy
Anduril succeeded, Fubini explained, because it built its own manufacturing capacity early and focused on systems that could be fielded quickly. The company’s Lattice AI platform and autonomous drones are now in active use by U.S. Special Operations Command and the Marine Corps. But Anduril’s path is the exception, not the rule.
Many newer entrants are chasing the same narrow set of programs: counter-drone systems, loitering munitions, and battlefield AI. Fubini cautioned that the Pentagon has limited appetite for multiple competing solutions in the same category. “The government doesn’t need 50 drone companies,” he said. “They need three or four that can actually deliver at scale.”
What the government actually buys
Defense procurement data from the Government Accountability Office shows that the Department of Defense awarded $420 billion in contracts in fiscal 2023. Of that, roughly 70% went to the five largest prime contractors: Lockheed Martin, Raytheon, General Dynamics, Northrop Grumman, and Boeing. Small businesses captured only 23% of prime contracts, and most of those were for services, not hardware.
For startups, the path to meaningful revenue typically requires either a prime contractor partnership or a direct award from a special innovation office like the Defense Innovation Unit or the Strategic Capabilities Office. Those offices, while growing, still represent a fraction of total spending.
Survival depends on manufacturing, not fundraising
Fubini’s central thesis is that the current funding environment is distorting incentives. Startups that raise large rounds based on valuation momentum rather than production milestones risk running out of runway before they can demonstrate manufacturing capability. “Raising money is easy right now,” he said. “Building something that works at scale is the hard part.”
The coming years will separate the startups that are genuinely building for the long haul from those riding a wave of cheap capital. For investors, the question is no longer which companies have the best pitch deck, but which ones can actually produce hardware that survives combat testing and government audits.
