Is the robot fleet actually funded? Kodiak's path to that answer runs through planning IP like this. Kodiak Robotics' December 2, 2025 grant US12485917B2 claims systems and methods for path planning of autonomous vehicles — turning perception into safe trajectories.
Read the planning depth in the claim. The CPC tags — B60W 60/001 autonomous operation, B60W 30/09 collision avoidance, B60W 40/02 environment estimation, plus B60W 2554 and B60W 2556 perception-context classes — describe a planning stack that consumes perception and produces driving decisions. That is the core of removing the human.
The burn read is that planning is the gating capability for driverless operation, and gating capabilities are expensive. A truck cannot drop its safety driver until its planning handles the full range of highway situations safely, and proving that takes sustained engineering and validation spend.
For a runway-focused reader, Kodiak's commercialization push into driverless trucking is exactly the kind of capital-intensive bet where the question is whether revenue arrives before cash runs out. The planning IP confirms technical seriousness; the order book and the cash-flow statement confirm whether the runway reaches the launch.
The honest limit: a path-planning patent documents capability, not cost, runway, or commercial traction. Planning is competitive. The grant proves the program is real, not that the economics close.
The takeaway for the money desk: driverless trucking is gated by planning, and planning is a sustained spend. Read planning patents as the cost of the capability that removes the driver — then check whether the balance sheet can fund the wait.