TAM is a story; the business model is the fact that decides it. Perrone Robotics' October 10, 2023 grant US11782442B2 claims a general-purpose robotics operating system with autonomous-vehicle extensions — a platform meant to run on other people's machines.
Read the model in the claim. The CPC tags — G05D 1/0088 autonomous control, B25J 9/1661 robot programming, G05B 2219/40393 motion control, Y10S 901/06 robot software — describe portable autonomy middleware. The asset is software that licenses, not a fleet that depreciates.
The reconciliation matters because the capital intensity of autonomy is mostly in the hardware and the fleet. A company that licenses an autonomy OS to integrators and manufacturers carries software economics: high gross margin, low capex, no vehicle depreciation. That is a fundamentally different burn profile than a fleet operator.
For a deal desk, the question is whether the platform actually gets adopted. A licensable OS is only valuable if integrators build on it, and the autonomy-software space is crowded with open and proprietary stacks. The model is attractive on paper; the moat depends on adoption.
The honest limit: a platform patent does not disclose license revenue or customer count. It establishes the model — software licensing over vehicle building. The economics depend on uptake the claim cannot promise.
The takeaway for the money desk: distinguish autonomy companies by where their capital goes. A licensable OS is a software-margin bet that dodges fleet capex; reconcile the valuation to adoption, not to the elegance of the platform.