The Anduril Playbook

In 2017, a new company started building finished defense products that no government had ordered. No contract. No cost-plus check funding the lab. Just its own capital, poured into hardware the Pentagon hadn't asked for yet. Nine years later it's worth $61 billion — and it doubled that number in nine months. Same playbook the whole way: build it first, on your own dime, then hand the customer a price tag, not an invoice.

The business behind the hardware. Not what they build — how they built it.

Issue 01 was Baykar: built the metal, refused the money. Issue 02 was Helsing: took the money, then bought the metal. Anduril is the third operating system — own every layer and fund the R&D yourself, so you answer to no one. Baykar and Helsing both arrived at the full stack from opposite doors. Anduril walked in through a third: it refused the one thing every defense incumbent is built around — letting the government pay for the research. The result is a valuation that makes last issue's raises look quaint.

Here are the seven moves.

This is the whole company in one decision. For seventy years, the way you built defense hardware was "cost-plus": the government funds your research, you bill your costs, you add a margin, and your incentive is to spend more, not ship faster. Anduril inverted it. It pays for its own product development — with venture and internal capital — builds the finished thing on spec, and sells it to the government at a fixed price. The lab bill is Anduril's problem. The price is the customer's only number.

That single inversion changes every incentive downstream. When you fund your own R&D, speed is suddenly your interest, not a concession. You own the IP outright. You set the roadmap. And you keep the margin you'd otherwise have handed back. The catch is obvious — you're spending your own money on products that might not sell — which is exactly why most incumbents never tried it.

Takeaway: Whoever funds the R&D owns the roadmap, the IP, and the margin. Pay for your own development and you stop being a vendor and start being a company.

The founders said it as a category, not a tagline: a "defense products company," not a "defense contractor." A contractor sells effort — bodies, hours, programs that run for a decade. A products company sells a finished, working thing you can buy off the shelf and field now. The distinction sounds semantic until you trace where the money and the talent flow. A products company is judged on whether the product works, not on whether the program survives the next budget cycle.

Reframing the category did the same job Helsing's "software-first" thesis did last issue: it let Anduril recruit, raise, and sell against a story incumbents couldn't tell about themselves. You cannot out-contractor a fifty-year-old contractor. You can make "contractor" the old category and put yourself in a new one.

Takeaway: Don't compete inside the incumbent's category — rename it. Sell a finished product where they sell effort, and the comparison stops being apples to apples.

Anduril builds drones, undersea vehicles, surveillance towers, command systems — but the asset that ties them together is software, an operating layer it calls Lattice. Every piece of hardware is a node; Lattice is the nervous system that makes them act as one. Strip it down and the strategy is a tech-company strategy wearing a defense costume: the hardware is the razor, the software platform is the thing that compounds.

This is why the breadth of products isn't sprawl. Each new system makes the platform more valuable, and the platform makes each new system easier to field. A pure hardware company adds products and adds complexity. A platform company adds products and adds gravity. Anduril has been, in its own words, a software and a hardware company since day one — and it's the software half that makes the hardware half defensible.

Takeaway: In a hardware business, build the software layer everything plugs into. Hardware can be copied; the platform that coordinates it compounds.

Anduril didn't grow every capability organically. It bought the hard, physical layers. It acquired an undersea-vehicle company (Dive Technologies, 2022), a solid-rocket-motor maker (Adranos, 2023), and an autonomous-aircraft designer (Blue Force Technologies, 2023) — propulsion, airframes, subsea, pulled in-house. The pattern is deliberate: acquire capability, not revenue. Each deal added a physical competence that's slow and painful to build from scratch and instantly useful once it's plugged into the platform.

This is the same destination Baykar and Helsing reached — owning the whole stack — but Anduril got there partly by checkbook, using the capital it raised to compress years of in-house engineering into a closing date. Buy the metal layers; build the software layer; let the platform make the acquisitions worth more than they were standalone.

Takeaway: Acquire capability, not customers. The physical competences that take years to grow are the ones worth buying — and a platform makes each one worth more than its price.

Most companies treat manufacturing as a cost center. Anduril treated it as a product. It's building a hyperscale factory — Arsenal, a roughly five-million-square-foot plant in Ohio — designed to turn out tens of thousands of autonomous systems a year. The twist is the software: Anduril runs the factory on its own manufacturing operating system, treating the plant as reconfigurable, software-defined infrastructure rather than a fixed line bolted to one product.

That means the factory isn't tuned for a single item; it's a platform for whatever the company designs next. Read it as the physical expression of Moves 1 and 3: if you fund your own R&D and your moat is software, then the smart move is to make the factory itself a piece of software-defined IP that any future product can run on. Helsing localized factories to sell sovereignty. Anduril abstracted the factory to sell scale.

Takeaway: Treat manufacturing as a product, not an expense. A software-defined, reconfigurable factory is a moat the next product inherits for free.

Capital-heavy hardware usually means slow, debt-laden, incremental growth. Anduril financed it like a software startup instead — raising venture-scale rounds on a tech cadence and staying private to keep moving. The ladder: more than $11 billion raised in total; a $2.5B round in mid-2025 at roughly $30.5B; then a $5B round in May 2026 at $61B, led by returning backers. The valuation doubled in about nine months.

The point isn't the size of the checks — it's the structure. By funding a capital-intensive industrial business with growth equity rather than government programs or debt, Anduril could outspend incumbents on plants and people while keeping the optionality of a private company. Revenue followed the capital: roughly $1B in 2024, about $2.2B in 2025, with the company projecting $4.3B for 2026. Tech-style capital, industrial-scale build.

Takeaway: Match your capital structure to your ambition, not your industry. Growth equity can fund a heavy industrial build faster than the sector's usual debt-and-program model — if you're willing to stay private and spend.

Anduril didn't start from zero credibility. Its best-known founder had already sold a hardware company — Oculus, the VR headset maker — to Facebook for about $2.3 billion. Its executive chairman came straight from the venture fund that had been hunting, explicitly, for the next Palantir or SpaceX in defense. That combination — a proven hardware operator plus an investor with a thesis already loaded — let Anduril raise its first money and recruit its first engineers on conviction, before there was a product to point at.

A prior exit is not just a line on a bio; it's spendable capital. It buys the benefit of the doubt from investors, the willingness of senior talent to take the risk, and the patience to self-fund R&D (Move 1) long enough for it to pay off. Most founders underuse the credibility they've already earned. Anduril spent it deliberately, on the two things money can't directly buy early: belief and people.

Takeaway: Your last win is working capital for your next one. Spend the credibility — on the early hires and early believers who won't show up for a pitch deck alone.

Issue 04, and a change of format. We've now torn down three giants who each tried to own everything — the metal, the money, the factory, the stack. So here's the question we've never asked out loud: what if the single most honest document a defense-tech company publishes isn't its funding announcement or its product page — but its careers page? We pulled ten real ones apart, line by line: the titles, the perks, the things they beg for and the things they quietly hide. One of them tells you a company is about to miss its numbers. Most are broken in the same four ways. Next issue: why every defense-tech careers page is broken — and the hiring moves worth stealing.

Commander is the business behind the hardware. Not what they build — how they built it.

New transmission every two weeks. If it earned its place in your inbox, forward it to one operator who’s building something — that’s how Commander spreads.

COMMANDER · ISSUE 03 · commander.media · Istanbul