Defense tech is flooded with money, but who’s built to last?
Anduril just hit a $14 billion valuation. Mach Industries, which most of the public hasn’t heard of, is now worth $3 billion. This isn’t a bubble; it’s a declaration of a new industrial order, one where Silicon Valley’s playbook of moonshots and software margins is being violently rewritten for the messy, lethal reality of state power. The U.S. government, in a rare moment of bipartisan clarity, is signaling a 40% budget increase for defense. A gold rush is underway, and the prospectors are floo
Analysis
Anduril just hit a $14 billion valuation. Mach Industries, which most of the public hasn’t heard of, is now worth $3 billion. This isn’t a bubble; it’s a declaration of a new industrial order, one where Silicon Valley’s playbook of moonshots and software margins is being violently rewritten for the messy, lethal reality of state power. The U.S. government, in a rare moment of bipartisan clarity, is signaling a 40% budget increase for defense. A gold rush is underway, and the prospectors are flooding in with pitch decks about autonomous systems and AI-driven targeting.
But here’s the unvarnished truth that the press releases and funding announcements obscure: this isn’t a technology race. It’s a systems race, and most of these new entrants are already dead. They just don’t know it yet. Ross Fubini, the VC who placed the first bet on Anduril, put it plainly when he warned about the “Valley of Death.” That’s not just a cute industry phrase. It’s a mass grave for startups.
The core delusion is believing that building a superior prototype—the thing Valley founders are brilliant at—is even 20% of the battle. The real war is waged in windowless Pentagon offices, in the grinding gears of the Federal Acquisition Regulation (FAR), and in the slow, paranoid digestion of institutional risk. You don’t win a $10 billion JADC2 contract with a slick demo. You win it by embedding your engineers into the Air Force’s development cycle for three years, by learning to speak the arcane language of requirements documents, and by accepting that your elegant, agile software will be bolted onto a decade-old legacy platform because that’s what the budget line item covers.
Anduril and Mach didn’t succeed just because they had good tech. They succeeded because they had founders who understood that the product wasn’t just hardware or code; the product was trust with a very specific, very risk-averse customer. They mastered the art of “dual-use” not as a marketing buzzword, but as a survival strategy—proving value to the DoD while maintaining a commercial foothold that keeps the lights on and the tech moving. They built cultures that could tolerate the glacial pace of government alongside the frantic sprint of startup iteration. That’s a cultural mutation most pure-play defense startups are completely unequipped for.
So what will happen to the new wave? Most will fizzle out in the funding gap between their Series A and their first real program-of-record contract. They’ll burn through cash trying to “sell” to the government, not realizing that the government doesn’t buy things; it awards contracts through processes that reward past performance and risk mitigation above all else. The few who make it through will do so not by being the most technically innovative, but by being the most operationally and politically adept. They’ll become, in essence, defense contractors—a role that Valley types often disdain.
This isn’t a criticism; it’s a diagnosis. The real disruption isn’t coming from a new AI model that can identify a tank faster. It’s coming from companies like Anduril that are fundamentally redesigning the business of being a defense prime. They’re applying software business models—subscription services for autonomous drone fleets, for instance—to hardware, creating recurring revenue streams that Wall Street loves and the Pentagon is slowly learning to tolerate. They’re creating a new category, one that sits uncomfortably between Lockheed Martin and a Silicon Valley SaaS startup.
The exciting frontier isn’t just more startups. It’s the pressure this places on the entire ecosystem. Can the DoD’s procurement office actually adapt to buy software like a service? Can the big primes, drowning in their own bureaucratic inertia, acquire and integrate these fast-moving entities without killing them? The real competition isn’t between Anduril and the next upstart. It’s between Anduril’s model of rapid, software-defined iteration and the decades-old defense acquisition model that has defined the industry since the Cold War.
I’m bullish on the mission. A modernized, more agile defense capability is a genuine national security necessity. But I’m deeply skeptical of the crowd chasing the money. The valley they need to cross isn’t filled with prototypes; it’s filled with paperwork, legacy systems, and cultural friction. The next great defense giant will look less like a startup that got lucky and more like a strange, new hybrid organism—part software company, part government integrator, part political entity. The frenzy is understandable, but the graveyard is filling up. The winners will be those who realize this isn’t a software problem to be solved, but a political and institutional labyrinth to be survived.
Disclaimer: The above content is generated by AI and is for reference only.