The Pentagon has a new operating model, and it looks a lot like Sand Hill Road. Emil Michael, the former Uber executive who now serves as the Department of Defense’s undersecretary for research and engineering, is systematically rebuilding how the world’s largest military buyer finds, funds, and fields technology. The result is a procurement machine that runs more like a venture capital firm than a government agency — and the traditional defense primes that have fed on cost-plus contracts for decades are watching their moat evaporate in real time.
The VC Playbook Inside the Pentagon
Michael’s approach is deceptively simple: treat defense procurement like a portfolio. Instead of locking into decade-long contracts with a handful of legacy giants, he’s placing faster, smaller bets on startups that can deliver capability now — not in 2035. His signature move has been the expansion of the Accelerate the Procurement and Fielding of Innovative Technologies (APFIT) program, which recently made its largest awards ever: five firms splitting as much as $42 million. That’s pocket change by Pentagon standards, but the speed and intent behind it represent a philosophical earthquake.
The throughline of Michael’s tenure is unmistakable: reduce bureaucracy, shorten decision cycles, and get capability into operators’ hands faster. He’s not the first Pentagon official to say those words. He might be the first one who actually means them — because he’s the first one who’s operated inside the system that made “move fast and break things” a corporate religion.
Why Defense Contractors Should Lose Sleep
Here’s the quiet part that nobody at Lockheed Martin, Raytheon, or Northrop Grumman wants to say out loud: if the Pentagon starts buying like a VC, the incumbents lose their structural advantage. The traditional defense business model is built on relationships, clearances, compliance moats, and the sheer inertia of multi-year contracting cycles. A startup can’t compete with that — unless someone inside the Pentagon decides those barriers are bugs, not features.
That someone is Michael. And he has backup. Venture capitalists poured over $28 billion into defense tech startups in 2025 alone, funding companies building autonomous drones, underwater vehicles, AI-powered logistics, and cybersecurity platforms. Companies like Anduril, Shield AI, and Palantir have already proven that Silicon Valley-bred firms can win defense contracts. Michael’s job is to make sure the Pentagon’s buying infrastructure doesn’t throttle that pipeline.
The Anthropic Problem — And What It Reveals
You can’t understand Michael’s Pentagon without understanding the Anthropic fight. Earlier this year, the Pentagon effectively blacklisted Anthropic after the AI company refused to allow its Claude models to be used in fully autonomous weapons systems or mass surveillance of American citizens. Secretary of War Pete Hegseth invoked federal law to designate Anthropic as a supply chain risk. President Trump ordered the federal government to stop using the company’s products entirely.
Michael’s position was blunt: the military needs AI companies that will allow “all lawful use” of their technology. Anthropic’s competitors — Google, OpenAI, and Elon Musk’s xAI — agreed to those terms. Anthropic didn’t. CEO Dario Amodei has said the company will challenge the Pentagon’s designation in court, arguing that current AI systems simply aren’t reliable enough to power fully autonomous lethal weapons.
What’s revealing isn’t the disagreement itself. It’s what the disagreement tells you about Michael’s philosophy: he views ethical restrictions on technology as operational obstacles, not guardrails. That’s a VC mindset, not a bureaucrat’s mindset. VCs don’t ask whether a technology should exist — they ask whether it will win. Michael is applying the same lens to weapons systems, and the Pentagon is letting him.
The Talent Pipeline Is Already Shifting
The money tells one story. The talent tells another. Engineers are leaving Google, Meta, and Amazon to build battlefield technology at defense startups. That flow used to be unthinkable — five years ago, Google employees staged a walkout over Project Maven, a Pentagon AI contract. Now, defense tech is one of the hottest sectors in Silicon Valley, and the stigma has almost entirely evaporated.
Michael is both a cause and a beneficiary of this shift. His presence at the Pentagon sends a signal to the Valley: this isn’t your grandfather’s defense procurement. This is a place where operators move fast, where startups can compete, and where the person making the decisions understands cap tables, burn rates, and product-market fit. Whether that’s reassuring or terrifying depends on how you feel about the world’s most powerful military adopting the same decision-making framework as a Series B investment committee.
Follow the Money — And the Incentives
There’s a second-order effect here that deserves more scrutiny. When the Pentagon buys like a VC, it also starts creating winners like a VC. Fast-tracked contracts to small firms mean the Pentagon is essentially kingmaking — deciding which startups survive and which don’t, with national security as the justification for speed over process.
That’s a concentration of power that would make any antitrust regulator uncomfortable in the commercial world. In the defense world, it’s being celebrated as innovation. The question nobody’s asking yet: what happens when the Pentagon’s portfolio strategy fails? VCs expect most of their bets to fail. Can the Department of Defense afford the same hit rate when the portfolio includes autonomous weapons, space-based missile defense, and AI-powered command systems?
The Verdict
Emil Michael is doing exactly what he was hired to do: bring Silicon Valley’s speed, aggression, and risk tolerance into a bureaucracy designed to resist all three. The traditional defense primes are right to be nervous. The $28 billion flooding into defense startups is real, the talent pipeline is shifting, and the person controlling procurement decisions thinks like a venture capitalist, not a procurement officer.
But here’s the uncomfortable truth: venture capital works because most bets fail and a few succeed spectacularly. In business, failed bets mean lost money. In defense, failed bets mean something else entirely. Michael’s Pentagon is moving faster than any defense procurement apparatus in modern history. The only question left is whether speed and lethality are things you want optimized by the same people who gave us Uber Surge Pricing.
