Palmer Luckey says Anduril's early investors weren't worried that his engineers would fail. They were worried that the customer would. Luckey founded the virtual-reality company Oculus VR and later the defense company Anduril Industries, which builds AI-powered defense systems. He says potential investors told him: "we believe that you'll be able to build things that work. We believe you'll be able to build products that are great, but we don't believe the government is capable of actually purchasing them at scale."
Luckey told the story on Moonshots with Peter Diamandis, in a session recorded at the Moonshots Live 2026 event on September 25 and published on September 28. He was answering a question from Salim Ismail, founder of Open ExO. Ismail called the Pentagon and defense ministries around the world "the most centralized entity in the history of humanity". He argued that a newcomer trying to decentralize defense must face a "massive immune system problem": insiders either try to absorb the outsider or push it out. He asked Luckey how Anduril handles that relationship.
A market dominated by five companies
Luckey said the investors' doubt came from how concentrated defense buying had become. Until "just a few years ago", he said, about 80% of major defense acquisition programs went to five companies. These programs, known as MDAPs, are the Pentagon's largest weapons-buying programs, or "basically anything that matters" in Luckey's words.
A 2024 Congressional Research Service report puts his figure in the right range, but only when you count programs. Of 78 major weapon systems in the Pentagon's fiscal 2024 budget materials, 58, or just over 74%, had at least one of five firms as a prime contractor: Lockheed Martin, Boeing, RTX, General Dynamics and Northrop Grumman. Measured by money, the share is smaller. Those five companies received about 29% to 35% of the Defense Department's domestic contract obligations each year from fiscal 2019 to 2022. The report traces how 51 major prime contractors consolidated into five. It also describes a market that often has a single government buyer and competitions where the winner takes all. Analysts disagree about the effects of that consolidation, the report notes. They differ over how much production capacity was lost, whether procurement costs rose and what efficiencies were gained.
In a market like that, a startup has to prove more than that it can build a good product. It also has to show that a buyer used to a handful of large suppliers will actually place large orders with someone new.
Build first, then sell
Luckey's answer was a different business model. "We're less of a contractor and more of a product company," he said. "We use our own money to build products. Then we sell them." Traditional defense contractors are usually paid by the government to develop a system. As Luckey describes it, Anduril takes on that development risk itself and offers the government something that already exists.
He says the approach pays off because of how the officials who buy are rewarded. In his account, people in government have "done very, very well" by betting on Anduril, and the company finishes on schedule. Then, he said, the next program manager thinks, "I know how I can make my job easy," and simply gives the contract to Anduril. A program manager is the official responsible for delivering a particular purchase. In Luckey's telling, each delivery that makes one of them look good makes the next one more willing to buy.
"The insanity was the reason"
Peter Diamandis, founder of XPRIZE and host of the show, said it must take guts to enter what he called "the most incestuous" industry on the planet. He asked what insanity had driven Luckey to do it.
"The insanity was the reason that I got into it," Luckey replied. He said the first page of the company's pitch deck promised it would "save taxpayers hundreds of billions of dollars a year by making tens of billions of dollars a year." The idea, he explained, was not only to sell products but to replace systems that cost "needlessly 10 times what they should."
Carmack's 1% argument
To explain why he took on such long odds, Luckey turned to John Carmack, whom he called one of his mentors and idols growing up. Luckey described Carmack as basically the inventor of the modern 3D game engine and of first-person shooters. Carmack also ran a rocket company, Armadillo Aerospace, which Diamandis recalled competing in an XPRIZE. Luckey said it was doing vertical takeoff and landing long before SpaceX. Luckey later hired Carmack away from the game studio id Software, and Carmack became chief technology officer of Oculus VR.
Luckey said that after Carmack left, he started his own company to go "all in" on artificial general intelligence (AGI), meaning AI that is broadly as capable as a person. Carmack wasn't trying to get there step by step. The company is Keen Technologies, which Carmack founded in 2022, according to a 2023 Alberta Machine Intelligence Institute announcement. The announcement says Keen raised an initial $20 million and that Carmack left his consulting CTO role at Meta in December 2022 to focus on it. It also announced a partnership between Carmack and AI researcher Richard Sutton and set a goal of building an AI prototype by 2030.
Luckey said he asked Carmack why he was doing it, given the odds. Luckey added that he himself still believed AI was going somewhere. He recalled Carmack's reply: even with a 1% chance of success, "which maybe is an overestimation", the potential payoff made it worth trying. As Luckey remembers it, Carmack told him that "the risk-adjusted outcome kind of demands that I do it."
Luckey presented the exchange as a parallel to how he felt about defense, not as the moment that led him to found Anduril. "If there was a 1% chance that I could help solve this problem, then it was one that would be massively impactful to work on," he said. He compared that with starting another gaming, space-launch or car company, and complained that "you don't have enough people getting into defense."