Every so often a company arrives that forces a whole industry to reorganise itself around what it is doing, and in defence technology that company is Anduril. In roughly eight years it has gone from a Silicon Valley curiosity to a $61 billion behemoth selling autonomous weapons and battlefield software to the Pentagon, and it has dragged a wave of venture money into an industry that used to be the exclusive preserve of a handful of century-old contractors. This is the first entry in our Company Profiles library of ones to watch, and it is the obvious place to start, because Anduril is the company everyone else in this section is measured against.
A note before we begin: this is a business and investment profile of a company operating in a serious and ethically contested field. We cover the money, the strategy, the record and the debate, at the level already public. We do not cover, and will not, the operational detail of any weapon.
The founder: from virtual reality to real ordnance
You cannot understand Anduril without Palmer Luckey, who is one of the more improbable figures in modern tech. He built Oculus, the virtual-reality headset company, in his parents’ garage as a teenager, sold it to Facebook for billions, and was then pushed out of Facebook in 2017 in a row over a political donation. Rather than retire on the proceeds at 24, he founded Anduril, named after a sword from Tolkien, with the explicit aim of dragging the US defence industry into the software age.
Luckey is the public face, the showman and the ideology, but he does not run the company day to day. That job belongs to CEO Brian Schimpf, alongside co-founders including Trae Stephens and Matt Grimm. It is a deliberate split: Luckey provides the vision and the megaphone, Schimpf provides the operational grip needed to actually build hardware at scale and win government contracts. The company now employs around 7,000 people.
What Anduril actually builds
The heart of it is software. Lattice is Anduril’s AI platform, which fuses data from cameras, radar and other sensors into a single real-time picture of a battlefield or a monitored area, and coordinates autonomous systems within it. The pitch to the Pentagon is that Lattice is the operating system for modern warfare, and that owning that layer is more strategically valuable than any single piece of hardware.
Around that software sits a growing hardware catalogue: the Sentry surveillance tower, the Ghost-X reconnaissance drone, the Roadrunner-M interceptor, the Bolt loitering munition, the Barracuda family of cruise missiles, and, most ambitiously, the YFQ-44A Fury, an autonomous fighter aircraft designed to fly alongside crewed jets, now reportedly in production at Anduril’s Arsenal-1 factory in Ohio. A reported $20 billion-plus in Pentagon programmes has cemented Lattice’s place inside the US military’s institutional plumbing.
| Anduril at a glance | Detail |
|---|---|
| Founder / CEO | Palmer Luckey (founder), Brian Schimpf (CEO) |
| Founded | 2017 |
| Latest valuation | $61bn (Series H, roughly double 11 months earlier) |
| 2025 revenue | ~$2.2bn (roughly doubled) |
| Employees | ~7,000 |
| Core product | Lattice AI command software, plus autonomous hardware |
The record is not spotless
For all the momentum, the technology has had public stumbles, and it is only fair to put them next to the valuation. The Wall Street Journal reported a run of setbacks: more than a dozen Anduril drone boats failing during a Navy exercise, a mechanical issue damaging its Fury jet, and an August test of its Anvil counter-drone system that touched off a 22-acre fire in Oregon. Its combat record in Ukraine has been mixed too, with loitering drones that crashed or missed, leading some Ukrainian units to stop using them in 2024. None of this is disqualifying for a company building hard things fast, but it is a useful corrective to the idea that autonomous warfare is a solved problem you simply buy off Anduril’s shelf.
Watch: Anduril’s Fury and the autonomous-fighter push
For a sense of the scale of ambition, Anduril’s YFQ-44A Fury, the autonomous combat aircraft meant to fly ahead of crewed fighters, entered production in 2026. Coverage of the programme and the company’s own reveals are collected by outlets such as The Aviationist, which tracks the milestones as they land.
The ethics, said plainly
There is no honest profile of this company that skips the obvious. Anduril builds lethal autonomous systems, and the central objection to the whole field is that handing life-and-death decisions to software crosses a moral line, and that the technology will spread to actors with fewer scruples. Luckey’s own framing is that his company builds AI weapons “responsibly”, and that a strong, deterrent-capable West is the moral choice in a dangerous world. Critics counter that “responsible autonomous weapons” is close to a contradiction, and that market incentives will erode whatever restraint the early players promise. Both arguments are serious, and where you land on them is a values question this profile will not pretend to settle for you. It is, however, the question the entire sector rests on.
The investment case, even-handed
Anduril is private, so most people cannot buy it directly; the relevance here is how sophisticated investors and analysts frame the bet, and the disagreement is sharp. (None of this is investment advice.)
The bull case, as its backers make it. Anduril has done in under a decade what defence primes took generations to do: build a real, software-first weapons company with billions in revenue that is doubling, a defining software platform embedded in the Pentagon, and a factory strategy to produce at scale. Its investors, led in the prior round by Founders Fund, are betting it becomes a new prime contractor for an era of autonomous warfare and rising geopolitical tension. If that thesis holds, $61 billion will look cheap.
The bear case, as the sceptics make it. This is where the deferred authority earns its keep, because the doubts come from the financial press, not from us. Barron’s has estimated that Anduril’s $61 billion valuation works out to roughly 14 times expected 2026 sales, against around two to three times sales for established contractors such as Northrop Grumman and L3Harris. Fortune ran the blunt headline that the defence-tech boom “has become a bubble, or it will be soon”, and reported early-stage defence startups fetching 17 to 50 times revenue. The larger worry, as several analysts put it, is that capital is moving faster than governments can turn new technology into recurring production orders, and that defence-tech valuations are now being priced on the same scarcity-driven logic as AI infrastructure, a category the Bank for International Settlements has already flagged as showing bubble-like characteristics.
The neutral read. Both can be true at once: Anduril can be a genuinely important company and an overpriced one. The things to watch are concrete and public: whether that doubling revenue keeps compounding, whether the Fury and other flagship programmes convert from prototypes into recurring production contracts, and whether the test-failure headlines become a pattern or a footnote. A 14-times-sales price leaves little room for delayed programmes or a messy factory ramp. Judge it on delivered contracts, not on announcements.
What this means
Anduril is the defining company of the defence-tech boom, and it carries all of that boom’s promise and all of its risk at once. It has genuinely changed how the Pentagon buys and thinks about autonomy, it is growing at a pace old contractors cannot match, and it is priced for a future it has not yet delivered, in a field whose ethics remain fiercely contested. That is exactly why it belongs at the top of a ones-to-watch list: not because the story is settled, but because almost nothing about it is. (None of this is investment advice; do your own research, and if it is a serious sum, speak to a professional.)
Related on Top Tool Stack: Helsing: Europe’s answer to Anduril · Shield AI: the autonomous pilot
Did you know: Anduril’s valuation implies roughly 14 times its expected 2026 sales, while century-old defence primes trade nearer two to three times. The market is paying for a future, and the question is whether the future arrives on schedule.