Palmer Luckey Got Fired From Facebook at 24. His $30 Billion Revenge Is Rewriting Modern Warfare.

Here's a sentence I want you to read carefully, because every word of it is true: the most important weapons company in America was founded by a twenty-four-year-old homeschooled kid who got fired from Facebook over a $10,000 political donation, then recruited five veterans of a CIA-linked data company, named his startup after a fictional sword from Lord of the Rings, and built it into a $30.5 billion autonomous weapons manufacturer that the Pentagon now considers essential to national security.

If you read that and thought, "that can't possibly be real," congratulations. You have a functioning brain. It sounds like the plot of a thriller written by someone who couldn't decide between a tech startup story and a military espionage novel, so they just mashed both together and figured no one would notice.

But that's the Palmer Luckey story. And the company — Anduril Industries — isn't a startup anymore. It's the company that shipped autonomous drones to Ukraine within weeks of Russia's invasion. The company that built a reusable interceptor missile that lands vertically like a SpaceX rocket. The company that designed an unmanned fighter jet in under three years while the F-35 program is still over a decade behind schedule with $183 billion in cost overruns. The company that's building a five-million-square-foot factory in Ohio to mass-produce autonomous weapons at a scale the defense industry has never attempted.

And Palmer Luckey wants to take it public. Which means you and I might soon be able to buy shares in the company that is, for better or worse, building the machines that will fight the next war.

Let me walk you through how a kid who built VR headsets in his parents' garage became the most consequential figure in American defense since the people who split the atom. And why the answer to whether you should invest isn't nearly as simple as the stock pitch would suggest.

Palmer Luckey's childhood was not normal. Homeschooled in Long Beach, California, he spent his teenage years the way most of us spent ours — except instead of playing Xbox, he was building virtual reality headsets from scratch in his parents' garage. Not modifying existing ones. Building them. From components he scavenged, bought, and in some cases fabricated himself. By 17, he had built a prototype that was better than anything the VR industry — such as it was in 2010 — had produced in decades.

Here's the thing about VR in 2010: nobody cared. The technology had been "the next big thing" since the early 1990s, and it had delivered approximately nothing. Every major tech company that had tried VR had produced expensive, nausea-inducing hardware that looked like a refrigerator strapped to your face. The industry was dead. The dream was dead. Virtual reality was a punchline.

And then a teenager from Long Beach built something in his garage that actually worked.

The breakout moment came when John Carmack — the legendary game developer behind Doom and Quake, a man whose opinion carries roughly the same weight in gaming as Warren Buffett's does in investing — got his hands on one of Palmer's headsets and used it for a demo at E3, the gaming industry's biggest annual event. Suddenly, Palmer's garage project was front-page news.

In April 2012, Palmer founded Oculus VR. He launched a Kickstarter campaign with a goal of $240,000. He raised $2.4 million. Ten times the target. The tech world went berserk. This wasn't a incremental improvement — it was a paradigm shift built by a kid who wasn't old enough to rent a car.

Mark Zuckerberg had a problem. Facebook was the most powerful social network on Earth, but it didn't own any hardware. Apple had the iPhone. Google had Android. Facebook had … a website. And Zuckerberg was furious about it, because in the tech industry, not owning a hardware platform means you're always one policy change away from irrelevance. Ask any developer who's ever had their app pulled from the App Store.

So when Zuckerberg saw Oculus, he didn't see a gaming gadget. He saw a platform. A future where people would interact through VR headsets the way they interacted through phones. A future where Facebook controlled the hardware, the software, and the entire experience. In March 2014, Facebook paid $2 billion to acquire Oculus.

Palmer Luckey was twenty-one years old. He was, briefly, the youngest self-made billionaire in tech. He was building the future of computing inside the world's most well-resourced social media company. Everything was perfect.

And then he made a $10,000 mistake.

In September 2016, a political action group called Nimble America surfaced online. Its stated purpose was to create pro-Trump memes and anti-Hillary Clinton advertising. It was exactly the kind of organization that exists in every election cycle — small, loud, largely irrelevant. Except for one detail: Palmer Luckey had donated $10,000 to it.

Ten thousand dollars. That's what it cost to end Palmer Luckey's career at Facebook. For context, $10,000 to a man worth over a billion dollars is the financial equivalent of you or me dropping a quarter in a Salvation Army bucket. It's not even a rounding error. It's a rounding error of a rounding error.

But Silicon Valley in 2016 was not a place that tolerated political dissent, and "political dissent" in this context meant any public support for Donald Trump. When the donation became public, Facebook employees revolted. Developers threatened to stop building for Oculus. Internal Slack channels turned into organized resistance. The company that built its entire brand on connecting people suddenly discovered that some connections were unacceptable.

Zuckerberg's team reportedly pressured Palmer into issuing a public apology. Palmer complied, posting a statement saying he was sorry for the impact his actions had. It didn't matter. In March 2017, Palmer Luckey was fired from Facebook.

He later described the experience with the kind of controlled anger that suggests the wound never fully healed: "I was fired at the height of my career. My gears were ground and I really wanted to prove that I was somebody, that I was not a one-hit wonder, and that I still had it in me to do big things."

Palmer Luckey was 24 years old. He'd built one of the most important tech products of his generation, sold it for $2 billion, and been thrown out of the company over a donation that wouldn't cover a month's rent in San Francisco. He was humiliated. He was furious. And he was about to channel that fury into building something far more dangerous than a virtual reality headset.

He disappeared from public view. The tech press wrote obituaries for his career. The consensus was clear: Palmer Luckey was done. A cautionary tale about mixing politics and technology. A one-hit wonder who'd gotten lucky and then gotten stupid.

The consensus, as it turned out, was spectacularly wrong.

To understand why Anduril exists, you need to understand something about the defense industry that most people — including most investors — don't fully grasp. It's not just big. It's grotesquely big. Global defense spending is approximately $2.4 trillion per year. The United States alone accounts for roughly 40 cents of every dollar spent on defense anywhere on Earth. That's nearly a trillion dollars a year flowing through the Pentagon and into the pockets of defense contractors.

And virtually all of that money goes to five companies.

Lockheed Martin. Raytheon. Northrop Grumman. Boeing. General Dynamics. The "Big Five" — or as anyone who's worked in defense procurement calls them, the "primes." Together, these five companies absorb the overwhelming majority of U.S. defense spending. They build the fighters, the missiles, the ships, the satellites, and the systems that project American military power around the globe.

They also, to put it politely, operate with the innovative spirit of a post office.

The reason is structural, and it's called the cost-plus model. Here's how it works: the government signs a contract with a defense company to build something — let's say a fighter jet. The company estimates the cost, adds a percentage on top as profit, and the government agrees to pay cost-plus-profit. If the project goes over budget? The government pays the overrun plus profit. If it takes longer than expected? The government pays for the delays plus profit.

Read that again. Under cost-plus, a defense contractor literally gets paid more when projects are late and over budget. The incentive structure isn't just misaligned — it's inverted. Speed and efficiency are punished. Delays and bloat are rewarded.

The F-35 is the poster child. Lockheed Martin's Joint Strike Fighter program began in 2001. It was supposed to produce a cutting-edge fighter jet at reasonable cost. As of 2024, the program is over a decade behind schedule with $183 billion in cost overruns. The total lifetime cost is estimated at $1.7 trillion. For a single aircraft program. That's not a defense contract — it's a wealth transfer mechanism disguised as a fighter jet.

Meanwhile, the technological revolution that transformed every other industry on Earth — software, AI, autonomous systems, cloud computing — had barely touched defense. The best engineers in the world were in Silicon Valley, optimizing ad auctions and food delivery routes, while the most important industry for national survival was operating with technology and processes designed in the Cold War.

Palmer Luckey and Trae Stephens looked at this situation and saw something everyone else had missed: the defense industry wasn't just overdue for disruption. It was begging for it. And the gap between what Silicon Valley could build and what the Pentagon was actually buying wasn't just a market inefficiency. It was a national security crisis.

Palmer Luckey didn't start Anduril alone. And the team he assembled tells you everything about what he was planning to build.

First was Trae Stephens. Stephens was a partner at Founders Fund — Peter Thiel's venture capital firm — and before that, he'd worked at Palantir, the data analytics company co-founded by Thiel that had built software used by the CIA, NSA, and virtually every intelligence agency in the Western world. Trae didn't just understand defense tech. He understood how to sell to governments. He understood the procurement process, the bureaucratic maze, the seventeen layers of approval required to get anything through the Pentagon. That knowledge would prove to be worth billions.

Then came Brian Schimpf, former Director of Engineering at Palantir, who became Anduril's CEO. Matt Grimm, also ex-Palantir, became COO. And Joe Chen, who'd worked with Palmer at Oculus and was a former Army paratrooper, bridged the gap between Silicon Valley software culture and the reality of military operations.

Notice the pattern: four of the five co-founders came from Palantir. That's not a coincidence. Palantir was the only company that had successfully built enterprise software for the intelligence community, and its alumni understood something crucial: the defense industry's problem wasn't a lack of good hardware. It was a lack of good software. The fighter jets and missiles and ships worked fine. What didn't work was connecting them, coordinating them, and making them intelligent.

They named the company Anduril, after the fictional sword from The Lord of the Rings. It translates to "flame of the West." The symbolism was deliberate: they were positioning themselves as defenders of Western civilization against emerging threats. Whether you find that inspiring or slightly unhinged probably says more about you than about them.

Their first meeting was a lunch at Palmer's house. A basic pitch deck. By the end of the meal, everyone was all in. Thanks to Trae's connections to Peter Thiel's Founders Fund, they secured initial funding in 2017. But even with a billionaire founder who'd had a $2 billion exit and a team of Palantir veterans, they struggled to raise capital. VCs had verticals for software, healthcare, fintech — nobody had a "defense tech" vertical. The conventional wisdom was clear: defense companies were government contractors, not venture-scale businesses.

And then they built their first product. And it nearly killed the company.

Anduril's first invention wasn't a weapon. It was a firefighting tank. The Sentry Tank was a tracked vehicle that could carry tons of firefighting foam and coordinate with overhead drones to battle wildfires. They even brought in Jamie Hyneman from MythBusters to help design it. The vehicle worked perfectly.

Nobody wanted to buy it.

Not a single government agency. Zero customers. They'd spent months and significant capital building technology that solved a real problem and found exactly zero demand for it. It was a near-death experience for a company that was less than a year old, and it taught Palmer and his team the most important lesson in defense tech: don't build cool things and then look for customers. Find out what the customer needs, then build the thing.

There's a certain cosmic irony in the fact that Palmer Luckey got fired from Facebook because of his political leanings, and then his next company was saved by a policy pushed by the politician he'd supported. History doesn't always rhyme. Sometimes it just laughs at you.

In 2018, the federal government allocated $641 million for border security technology. The physical wall that Trump had campaigned on was politically contentious and logistically complicated. But a "virtual wall" — a network of AI-powered surveillance towers that could detect and track border crossings in real time — was something both parties could quietly support.

Anduril pivoted. Hard. The Sentry Tower system they developed could scan 360 degrees across 7.5 miles, using artificial intelligence to distinguish between animals, vehicles, and humans crossing illegally. It didn't build a wall. It built something arguably more effective: an invisible net of sensors, cameras, and AI that could see everything and report it instantly.

Of course, the media had a field day. Headlines screamed about "the Oculus founder building virtual reality borders for Trump." Editorials were written. Twitter was furious. The narrative wrote itself: fired tech bro builds surveillance state for authoritarian president. It was the kind of story that generates clicks and outrage in roughly equal proportions.

The reality, as usual, was more nuanced. Understanding what crosses your border — drugs, humans, vehicles — isn't a Republican issue or a Democratic issue. It's a national security issue. And the Sentry Tower deployments didn't just survive the transition from Trump to Biden. They grew under Biden. By 2024, Anduril had hundreds of towers installed along the U.S. border, and no one in Washington was seriously talking about removing them.

But the border was never the point. The border was the beachhead. The revenue that kept the lights on while Palmer and his team built what they'd actually set out to build: autonomous weapons systems for the most powerful military on Earth.

And then, in February 2022, the world changed. And suddenly, everything Anduril had been building wasn't theoretical anymore.

When Russia invaded Ukraine on February 24, 2022, every military analyst on Earth was forced to rethink everything they thought they knew about modern warfare. The conventional wisdom held that wars between nation-states would be decided by the traditional hierarchy of military power: fighter jets, tanks, aircraft carriers, cruise missiles. Big, expensive platforms operated by highly trained personnel.

Ukraine obliterated that assumption in about six weeks.

What the world watched, in real time, via satellite imagery and social media, was the emergence of a new kind of warfare. Cheap, autonomous drones — some of them barely more sophisticated than the quadcopters you can buy at Best Buy — were destroying Russian tanks that cost millions of dollars each. A $500 drone with a repurposed grenade was killing $3 million armored vehicles. The math of warfare had inverted overnight. Suddenly, the expensive thing was the vulnerable thing, and the cheap thing was the weapon.

Anduril was ready. Disturbingly ready.

Within weeks of Russia's invasion, Anduril shipped Ghost X reconnaissance drones to Ukrainian forces. Not months. Not "after a lengthy procurement review." Weeks. In an industry where getting a new weapon system from concept to deployment typically takes a decade, Anduril had combat-capable drones on the front lines before most defense contractors had scheduled their first meeting to discuss the situation.

The Ghost drones didn't just work. They evolved. Facing some of the most advanced electronic warfare systems Russia had deployed, Anduril's engineers developed flying mesh networks that allowed drone swarms to relay data across multiple units, maintaining communication even when individual drones were being jammed. They were iterating in weeks, not years. Software updates pushed to drones in the field the way Apple pushes iOS updates to your phone.

Then came the Altius loitering munitions. These are the modern version of the kamikaze — autonomous drones that fly to an area, circle overhead waiting, identify targets using computer vision, and then strike. The UK supplied Altius 600M units to Ukraine through contracts worth £30 million, designed specifically for destroying Russian armor. A drone that costs a fraction of a tank, deployed in weeks, updated in days, and operated by soldiers who trained on it in hours rather than months.

Ukraine didn't just validate Anduril's thesis. It proved it in the most visceral way possible: on an actual battlefield, against an actual adversary, in an actual war. Autonomous weapons weren't the future. They were the present. And the Pentagon was watching.

And there was one more data point that made the generals in Arlington sweat. In 2023 and 2024, Houthi rebels in Yemen began attacking commercial shipping in the Red Sea with cheap drones. The U.S. Navy responded by shooting them down — using interceptor missiles that cost $3 to $4 million each. In 15 months, the Navy spent nearly $1 billion shooting down drones that cost a few thousand dollars apiece. The economics were suicidal. You cannot sustain a military posture where the defense costs a thousand times more than the attack.

The Pentagon needed something new. And Anduril had spent the previous five years building exactly that.

What Anduril has built in seven years would be impressive for a defense company that had been operating for seventy. The product portfolio reads like someone gave a twenty-four-year-old a blank check and said, "Build everything the Pentagon is too slow to build itself."

It starts with Anvil. Revealed in 2019, the Anvil is an unmanned quadcopter designed to kill other drones. It flies at 200 miles per hour, autonomously identifies hostile drones, calculates an intercept course, and physically destroys them. The Anvil M variant carries an explosive payload for targets that need more persuasion than a high-speed collision. The Marine Corps ordered $200 million worth of Anvil systems for base defense. Why? Because existing air defense systems were designed for aircraft and missiles, not swarms of cheap commercial drones. Using a $100,000 Patriot missile to shoot down a $500 drone isn't defense — it's fiscal suicide.

Then came the Ghost series. Ghost drones aren't hobbyist quadcopters with cameras strapped to them. They're military-grade reconnaissance platforms designed for contested environments — areas where the enemy is actively trying to jam, hack, or shoot down anything that flies. The U.S. Army selected Ghost for their company-level small UAS program, which means thousands of units deployed across infantry companies.

The Altius family followed. Anduril acquired Area-I in 2021 specifically to add loitering munition capability to their portfolio. Altius drones patrol an area, identify targets using computer vision, coordinate strikes with other Altius units, and then execute. They're the spear to Ghost's eyes.

But the two products that really make defense industry veterans lose sleep are Roadrunner and Fury.

Roadrunner is an interceptor missile that lands vertically and can be reused. Read that sentence again. Traditional interceptor missiles are single-use: you fire one, it destroys the target, and you need a new missile. At $3-4 million per shot, that adds up catastrophically fast when you're defending against swarms of cheap drones. Roadrunner launches vertically with twin turbo jets, intercepts the threat at high speed, and then — if it doesn't need to make contact — comes back and lands vertically, like a SpaceX Falcon 9 booster. One unit can conduct hundreds of missions over its operational lifetime. The Navy is already evaluating it for deployment on ships.

And then there's Fury. Think Top Gun, but without a pilot. Fury is an unmanned combat aircraft designed to fly alongside manned fighters as an autonomous wingman. The specs are ridiculous: Mach 0.95 top speed, 50,000-foot ceiling, 9g maneuverability. It's about half the size of an F-16, but it can pull maneuvers that would kill a human pilot. It can be risked in ways that you would never risk a human being.

The strategic concept is called "affordable mass." Instead of spending $80 million on a single F-35 and putting a human pilot in it, you deploy swarms of unmanned Fury aircraft that cost a fraction of the price and overwhelm enemy defenses through sheer numbers. Each Fury that gets shot down is a financial loss. Each F-35 that gets shot down is a financial loss, a political crisis, and a dead pilot.

Here's the number that should make every legacy defense contractor's board of directors nervous: Anduril designed and built Fury in under three years. The F-35 program began in 2001 and is still having problems. That's not a comparison. That's an indictment.

Anduril has also expanded underwater with Dive-LD, a 3-ton autonomous submarine for deep-sea missions including mine countermeasures, anti-submarine warfare, and infrastructure inspection. Australia awarded Anduril a joint development program for their Ghost Shark autonomous submarines based on this platform. The world's most critical infrastructure — oil pipelines, fiber optic cables carrying your internet traffic — sits on the ocean floor, and autonomous underwater vehicles provide persistent monitoring at a fraction of the cost of traditional submarines.

Seven products. Seven years. A surveillance tower, an anti-drone system, a reconnaissance drone, a loitering munition, a reusable interceptor missile, an unmanned fighter jet, and an autonomous submarine. Any one of these would be the entire product line of a traditional defense startup. Anduril built all of them — and they're not even the most important thing the company has created.

Palmer Luckey's original insight — the one that separates Anduril from every defense contractor that came before it — was deceptively simple: the future of warfare would be defined by software, not steel.

Every product I just described — Ghost, Anvil, Altius, Roadrunner, Fury, Dive — is impressive hardware. But hardware, in Palmer's view, is a commodity. Anyone can build a drone. Anyone can build a missile. The hard part — the part that creates an actual, defensible competitive advantage — is building the software that connects all of it together and makes it intelligent.

That software is called Lattice OS. And it might be the most important thing Anduril has built.

The operational implications are staggering. When a Sentry Tower detects an incoming threat, Lattice can automatically task a Ghost drone for visual confirmation, calculate intercept trajectories for Roadrunner missiles, coordinate a swarm of Altius loitering munitions, and present the human operator with recommended actions. All in seconds. Not minutes. Not hours. Seconds.

A single operator, sitting in front of a screen, simultaneously controlling drones, submarines, anti-drone systems, and surveillance towers. A human brain augmented by AI that handles the complexity, the sensor fusion, the coordination, and the split-second calculations that no human could perform alone. The operator makes the decision. Lattice executes it across every platform in the network.

But here's why Lattice matters even more from a business perspective than from a military one: hardware margins are thin and contracts are inconsistent. The military might order 10,000 Ghost drones one year and zero the next. Hardware revenue is lumpy, unpredictable, and subject to the whims of the procurement cycle.

Software subscriptions are not.

Lattice creates sustainable, sticky, long-term, high-margin SaaS revenue in the defense industry. Once the military integrates Lattice into its operations, switching costs become enormous. You don't rip out the operating system that connects your entire autonomous arsenal because a competitor offers a 5% discount. Lattice is the lock-in mechanism. It's the moat. The hardware is impressive, but Lattice is what turns Anduril from a defense contractor into a defense platform — and platforms, as we've learned from Apple, Google, and Microsoft, are worth trillions.

There's a point in every hardware company's life when it has to answer an uncomfortable question: can you actually build this at scale? A prototype is a proof of concept. A few hundred units is a pilot program. But tens of thousands of autonomous weapons systems per year? That's a manufacturing challenge of an entirely different order.

Anduril's answer is Arsenal 1: a five-million-square-foot hyperscale manufacturing facility in Columbus, Ohio. Billion-dollar investment. Four thousand workers when fully operational. It's not a factory — it's a statement. A declaration that Anduril isn't a startup that builds clever prototypes and waits for Lockheed Martin to license the technology. It's a full-scale defense manufacturer that intends to build its own products, at its own facilities, at a scale that competes directly with the primes.

The manufacturing philosophy is as innovative as the products. Traditional defense production requires specialized tooling for each platform. If the military orders 5,000 Ghost drones this year but pivots to Roadrunner interceptors next year, a traditional factory would need entirely new assembly lines. Arsenal 1 uses common tooling across all product lines. The same machinery that builds Ghost drones can be reconfigured to produce Roadrunner interceptors or Dive underwater vehicles. It's composable manufacturing — the same materials and production lines can be repurposed to build whatever the military needs most urgently.

There's a strategic dimension here, too. America's advantage in World War II came from industrial capacity — the ability to out-produce the Axis powers in ships, planes, tanks, and ammunition. Today, China holds that manufacturing edge. Arsenal 1 represents an attempt to reclaim industrial superiority, at least in the domain of autonomous systems, through software-driven manufacturing that can adapt faster than China's scale advantage can respond.

The capital to build Arsenal 1 is one of the primary reasons Anduril has raised over $4 billion across six funding rounds. The trajectory tells the story: Series C in 2020 at a $1.9 billion valuation. Series G in June 2024 at $30.5 billion, raising $2.5 billion from investors including Morgan Stanley and Fidelity. These aren't seed-stage VCs taking fliers. These are institutional heavyweights setting the stage for an IPO.

The revenue growth supports the valuation. Anduril approximately doubled revenue to $1 billion in 2024. They've won contracts worth billions — $967 million from Special Operations Command for counter-drone systems, $642 million from the Marine Corps, plus the $200 million Anvil order. The backlog is substantial and growing.

Palmer has been explicit about the IPO: "We are definitely going to be a publicly traded company. We are running this company to be the shape of a publicly traded company."

The case for why is straightforward: massive defense contracts require the transparency and trust that comes with public status. Palmer has also said he wants American citizens to be able to own shares in a company funded by their tax dollars — a populist argument that plays well regardless of which side of the political aisle you sit on.

When Anduril does go public, it will likely be valued not as a defense contractor but as an AI and software company that happens to make weapons. That distinction matters enormously. Traditional defense contractors trade at 15-20x earnings. Software companies trade at 30-50x or more. If Wall Street decides Anduril is a software platform — and Lattice gives them every reason to — the IPO valuation could be staggering.

And there's an even bigger story lurking behind the defense numbers. Palantir started as a defense contractor. Today it's one of the most valuable technology companies in the world because it proved its defense software worked in commercial applications. Could Anduril follow the same path? The same Dive autonomous submarines that monitor naval threats could inspect underwater oil pipelines. The same drone technology that surveys battlefields could manage agricultural operations. The same Lattice platform that coordinates military assets could coordinate any organization that manages autonomous systems.

Defense might be the beginning. Not the ceiling.

I've spent the last nine sections telling you why Anduril is extraordinary. Now let me spend one telling you why it might be a terrible investment. Because if you're going to put money into a company that builds autonomous weapons, you should do it with your eyes wide open.

Risk 1: The One-Customer Problem. Anduril's revenue comes overwhelmingly from one customer: the United States Department of Defense. And the DoD is not a normal customer. It's a customer that operates on political cycles, is subject to congressional budgeting fights, and can cancel or restructure programs based on which party controls the White House and which generals are in favor this year. A change in administration, a shift in strategic priorities, a budget sequestration — any of these could dramatically impact Anduril's revenue pipeline. Concentration risk in traditional business is dangerous. Concentration risk when your customer is the federal government is existential.

Risk 2: Palmer Luckey Is a Political Lightning Rod. Palmer's political views are well-documented and, in certain circles, deeply controversial. He's a vocal Trump supporter in an industry where that position alienates significant portions of the talent pool. For a company that needs to recruit the best AI engineers in the world — engineers who overwhelmingly work in the politically left-leaning tech industry — having a founder with Palmer's profile is a challenge. Every news cycle that involves Palmer's politics is a news cycle that isn't about Anduril's technology.

Risk 3: The Ethics of Autonomous Weapons Are Unresolved. The United Nations has been debating autonomous weapons systems for years. The fundamental question is simple and terrifying: should machines be allowed to make kill decisions without a human in the loop? Anduril maintains that Lattice always keeps a human operator in the decision chain. But the technology itself doesn't require that. The same software that presents a human operator with recommended actions could, with minor modifications, execute those actions autonomously. The ethical Rubicon hasn't been crossed yet, but the bridge has been built and the troops are standing on it.

Risk 4: Pentagon Procurement Is Glacial. Yes, Anduril shipped drones to Ukraine in weeks. But standard DoD procurement cycles measure in years, not weeks. The process of getting a new system from evaluation through testing through certification through full-scale deployment is deliberately slow, bureaucratic, and designed to prevent exactly the kind of rapid disruption that Anduril represents. The irony of defense innovation is that the customer most needs the product is also the customer least equipped to buy it quickly.

Risk 5: China. The strategic argument for Anduril rests partly on the premise that the U.S. needs to out-innovate China in autonomous weapons. But China isn't standing still. China's defense budget is growing at double digits annually. Chinese companies are developing their own autonomous drones, their own AI warfare systems, their own manufacturing infrastructure — and they're doing it with the backing of a state that doesn't need congressional approval for funding rounds. The race Anduril is running isn't against Lockheed Martin. It's against the industrial capacity of 1.4 billion people run by a government that has declared technological supremacy a matter of national survival.

Risk 6: The Competition Is Coming. Anduril had the defense tech space largely to itself for several years. That's no longer true. Shield AI is building autonomous fighter pilots. Palantir — the company that supplied four of Anduril's five co-founders — is expanding aggressively into defense hardware coordination. L3Harris, Kratos, and even some of the primes themselves are investing in autonomous systems. First-mover advantage is real, but it's not permanent. And in a market where the customer is the government, contracts go to whoever wins the bid, not whoever got there first.

So here's where I leave you.

Anduril is building machines that will fight wars. Not might. Will. The technology works. Ukraine proved it. The Red Sea proved it. The shift from human-operated platforms to autonomous systems isn't a prediction — it's a transition already underway, funded by hundreds of billions of dollars in military spending from every major power on Earth.

Palmer Luckey built a VR headset in his parents' garage at seventeen, sold it for $2 billion at twenty-one, got fired over a $10,000 donation at twenty-four, and built a $30.5 billion defense company by thirty-two. The arc is extraordinary. The ambition is extraordinary. The risks are extraordinary.

The question isn't whether autonomous weapons will define the future of warfare. That debate is over. The drones are already flying, the submarines are already diving, and the interceptor missiles are already landing vertically and getting ready to fly again.

The question is whether Palmer Luckey — the homeschooled kid from Long Beach who turned rage into a weapons empire — is the right person to build the machines that will decide who wins the next war. And whether that question even matters anymore. Because the machines are being built regardless. The only question left is who builds them.

And right now, that answer is Anduril.