Product-first defense startup playbook

Anduril worked because it sold the government a finished thing, not a promise to build one. The early product was a border surveillance system that could be deployed fast, priced like a commercial product, and improved over time through recurring software and support, which let Anduril skip years of custom R&D contracting. That is also why Forterra moved from applied research toward a repeatable product stack sold across defense and commercial vehicle programs.1 2 3 4

• The real wedge was not defense in general, it was one urgent buyer with budget and authority. Customs and Border Protection had a live surveillance problem, Boeing's earlier approach had been slow and expensive, and newer edge GPUs made Anduril's tower economics finally workable.1 3 4
• Recurring license basis means the money model looked more like software than like a classic defense prime. Instead of billing hours on a cost plus development contract, Anduril could sell a fixed product, keep ownership of the tech, update the software, and layer new use cases like counter drone on the same core stack.1 2 3
• This is why the broader market shifted from research shops to product companies. Palantir and SpaceX helped force agencies to buy commercial technology, Anduril proved a startup could field hardware plus software quickly, and Forterra is following that path with one autonomy kit reused across military and industrial vehicles.1 2 3 5

The next wave in defense tech will be less about copying Anduril and more about finding one sharp entry point, then expanding from a shared software and hardware core. Companies that can turn an urgent program into a repeatable product line will compound fastest, because every deployed system becomes a reference, a dataset, and a cheaper starting point for the next contract.1 2 5