Last Updated on July 7, 2026 by Taya Ziv
In 2016, Palmer Luckey donated $9,000 to a political group supporting Donald Trump. He was 24 years old, already worth hundreds of millions from selling Oculus to Facebook for $2 billion, and about to learn what might be the most expensive lesson in Silicon Valley history: in the Bay Area, your politics can cost you more than your product ever earned you.
Facebook fired him in 2017. The VR community turned on him overnight. Game developers publicly pulled Oculus support. Conference invitations disappeared. A kid who had single-handedly revived virtual reality and sold the company for $2 billion at 21 couldn’t get a meeting in San Francisco.
So he did something nobody in tech expected. He started building weapons.
Yesterday, Palmer Luckey’s defense technology company Anduril Industries raised $5 billion in a Series H round at a $61 billion valuation. That’s double the $30.5 billion it was worth less than a year ago. Revenue hit $2.2 billion in 2025 and is projected to reach $4.3 billion this year. The Pentagon just handed it a $20 billion contract. The round was led by Thrive Capital and Andreessen Horowitz, and the company has now raised more than $11 billion in total.
For context, $61 billion is higher than Palantir was at its IPO. It’s higher than SpaceX was in 2020. It makes Anduril one of the most valuable private companies in American history, and it got there by doing the one thing Silicon Valley has spent a decade telling founders not to do: work with the military.
The Market Nobody Wanted
There’s a reason Luckey had the space to himself when he started Anduril in 2017. Silicon Valley had spent years building a culture where defense contracts were toxic. Google employees revolted over Project Maven in 2018, forcing the company to abandon an AI contract with the Pentagon. Microsoft faced internal protests over its HoloLens deal with the Army. The message from the talent pool was clear: build weapons, lose engineers.
So most founders didn’t. The few defense tech startups that existed were small, underfunded, and treated like the weird kids at the VC mixer. The prevailing wisdom was that government procurement was too slow, the sales cycles too long, and the moral baggage too heavy for a startup to carry.
Luckey and his co-founders, several of whom came from Palantir, saw something different. They saw a $900 billion annual US defense budget being spent on companies that hadn’t innovated since the Cold War. Lockheed Martin, Raytheon, Northrop Grumman, and Boeing were delivering hardware that was over budget, behind schedule, and running software that belonged in a museum. The gap between what Silicon Valley could build and what the Pentagon was buying was enormous.
And because nobody in the startup world wanted the business, there was almost no competition.
What Anduril Actually Built
This is where the story gets interesting, because Anduril didn’t build weapons the way defense primes do. It built a software platform and then wrapped hardware around it.
The company’s core product is Lattice, an AI-powered operating system for the battlefield. Think of it as the central nervous system that connects every sensor, drone, camera, and weapon system into a single operational picture. A commander doesn’t look at 15 different screens from 15 different vendors. They look at Lattice.
On top of Lattice, Anduril builds autonomous systems: counter-drone interceptors that can identify and destroy enemy drones without a human pilot, the Ghost Shark autonomous submarine under contract with the Royal Australian Navy, surveillance towers that patrol borders using computer vision, and precision strike munitions.
The key insight is that Lattice can integrate with hardware from other vendors too. So once the military adopts Lattice as its command layer, Anduril becomes the default software backbone for an entire theater of operations. The hardware is valuable. The software is the moat.
If that sounds familiar, it should. It’s the same playbook every successful enterprise SaaS company runs, except the customer is the Department of Defense and the “users” are carrying rifles.
The Anthropic Mirror
Here is the part that should make founders think.
Three weeks ago, we covered how Anthropic walked away from the Pentagon’s classified AI program because it conflicted with the company’s safety commitments. That was a principled decision and I respect it. But it was also a business decision. Anthropic chose a market position: safe, commercial, enterprise AI. No classified work. No weapons.
Anduril chose the opposite. And the market is telling you which choice pays better, at least right now.
Anthropic’s last reported valuation is around $61 billion (with a $900 billion round in discussions). But Anthropic has burned through billions in compute costs, faces commoditization pressure from open-source models, and competes head-to-head with OpenAI, Google, and Meta. Anduril hit $61 billion while selling to a customer that has a $900 billion annual budget, faces almost no startup competition, and has contracts that lock in revenue for a decade.
I’m not saying one path is morally better. But if you’re a founder evaluating markets, look at the competitive dynamics, not the dinner-party vibes. Anthropic is fighting a five-front war. Anduril is one of the only companies in its arena that can actually deliver modern technology to the Pentagon at startup speed.
The Uncomfortable Market Thesis
Here’s the founder lesson, and it goes beyond defense.
The most valuable markets in AI right now are the ones that make polite Silicon Valley conversation awkward. Defense. Energy infrastructure. Industrial automation. Surveillance. Insurance underwriting. Government services. These aren’t sexy. They don’t trend on Twitter. They don’t get you invited to the right parties in San Francisco.
But they share three characteristics that should matter to any founder thinking clearly.
First, they have massive incumbent spending. The US defense budget alone is $900 billion. Energy infrastructure runs in the trillions. These are real dollars being spent on real problems, not speculative TAM slides.
Second, the incumbents are terrible at technology. Lockheed Martin’s software is decades old. Utility companies run SCADA systems from the 1990s. Insurance still processes claims with humans reading PDFs. The gap between what AI can do and what these industries currently use is wider than anything in consumer tech.
And third, the competition is thin because the talent avoids it. Every Stanford CS grad wants to build the next ChatGPT wrapper. Almost nobody wants to build autonomous counter-drone systems or industrial AI infrastructure that replaces physical labor. When everyone crowds into the same market, margins compress and CAC skyrockets. When nobody shows up, you set the terms.
Palmer Luckey didn’t choose defense tech because he loved the military. He chose it because he was locked out of consumer tech, looked at the market with fresh eyes, and saw that the biggest opportunity was the one everybody else was avoiding.
What This Means For Your Cap Table
The defense tech funding wave isn’t just Anduril. Shield AI raised $300 million last quarter. Rebellion Defense secured a $1 billion Pentagon contract. Palantir, the elder statesman of defense tech, is trading at $175 billion in public markets. The money that VCs are pouring into defense tech in 2026 would have been unthinkable in 2020.
And it’s not just government-friendly VCs. When Founders Fund deployed $4.6 billion into just seven companies in under a year, defense and hard-tech were dominant themes. Thrive Capital, which led Anduril’s latest round, is the same firm that led rounds for companies across the AI spectrum. Andreessen Horowitz, which participated, published an entire thesis on “American Dynamism” built around the idea that building for government and defense is the next venture-scale opportunity.
The investor signal is clear: the money has decided that defense tech is investable. The question is whether founders have caught up.
The Real Lesson From a $9,000 Donation
I keep coming back to the $9,000 number. Not because the political angle matters for this article, but because of what it reveals about how Silicon Valley allocates talent.
A founder got pushed out of the industry over an amount of money that wouldn’t cover one month’s rent in Palo Alto. He landed in a market that the tech world considered untouchable. Nine years later, that “untouchable” market made him worth more than most of the people who pushed him out.
There’s a lesson in there, and it’s not about politics. It’s about the cost of consensus thinking. Every founder knows they should look for markets with low competition. But when they find one, their first instinct is to ask why nobody else is there, and then talk themselves out of it.
Palmer Luckey didn’t have that luxury. He couldn’t go back to consumer tech, so he was forced to look at what was actually available. And what was available turned out to be worth $61 billion.
The best market for your startup might be the one that makes your investors slightly uncomfortable. The question is whether you’ll need to get fired first to see it.
