Reject Custom Builds That Don't Scale
Scott Sanders, Chief Growth Officer at Forterra, on autonomy for every vehicle
The real filter on early defense revenue is whether it creates a repeatable product, not whether a customer is willing to pay for a pilot. In autonomy, a custom build for one buyer can soak up engineers, lock the roadmap to edge cases, and leave the company with no standard kit to sell to the next program. Forterra’s strategy has been the opposite, one core autonomy stack that can go on military trucks and yard tractors, because venture scale comes from shipping the same system again and again.
-
Forterra had to wind down 17 legacy contracts while consolidating around a single product. That is the concrete version of the tradeoff. Services and one off projects bring revenue, but they can pull engineering away from the standard hardware and software package that eventually scales into programs of record and OEM deployments.
-
This lesson came directly out of the Anduril playbook. Early traction mattered, but the winning products were the ones that solved a large urgent problem with an off the shelf system and a clear path from small contracts to bigger production buys, not projects that needed endless custom development for each office inside DoD.
-
The comparable in air autonomy is Shield AI. Its growth came from selling repeatable drone systems and then licensing the same autonomy software across more platforms. That is the shape of a venture scale defense company, one product base reused across many buyers, rather than bespoke work for fragmented demand pockets.
Going forward, the winners in defense and industrial autonomy will be the companies that say no to attractive but narrow custom work early enough to preserve a standard product. As more budgets open for robotic systems, the companies with a deployable kit, production discipline, and multiple adjacent markets will compound fastest.