Anduril's Defense Product Engine
Ross Fubini, Managing Partner at XYZ Capital, on the defense tech opportunity
The key insight is that Anduril did not start as a one product company, it started as a fast defense product factory. In the early years, the team could dream up hardware, get it working on a test range, sell a small first deployment to an operator who saw it in person, then promote the winners into repeatable programs like Sentry Tower and later counter drone systems. That shortened the path from idea to revenue and taught the company what buyers would actually fund.
-
The early wedge was not a huge Pentagon program. Anduril first sold into border and homeland security use cases, where a painful surveillance problem already existed and the company could ship a working tower quickly. That gave it a live customer, recurring software usage, and proof that the product approach could work in defense.
-
What made the pipeline valuable was shared technical reuse. The same core Lattice software, sensors, and autonomy stack could support multiple products. Counter UAS started as a tower pointed upward with similar software underneath, which meant each new product was not built from zero.
-
This is the opposite of the usual defense contractor model. Instead of waiting years for the government to define specs and pay cost plus, Anduril funded R&D itself, brought a finished product to demos, and sold fixed price systems. That model can produce much higher margins, but only if a few experiments become scaled programs.
Going forward, the winners in defense will look less like single product startups and more like product engines with one shared software base and several hardware lines. Anduril's later scale, from hundreds to thousands of units and revenue reaching $1B in 2024, shows how a pipeline of experiments can become a new kind of prime contractor if enough early prototypes graduate into repeatable procurement programs.