A capital-light “Driver-as-a-Service” bet on Physical AI—Kodiak leads in real-world industrial autonomy today, but long-haul scale and near-term funding will decide the outcome.
Overview
Kodiak AI (KDK) develops autonomous driving technology purpose-built for heavy-duty use cases—industrial hauling, long-haul trucking, and defense logistics—making it a levered bet on the emergence of “Physical AI.” Founded in 2018 and publicly listed in Sept 2025 via a business combination, the company’s core product is the “Kodiak Driver,” a unified virtual-driver platform designed to be vehicle-agnostic and deployable across multiple verticals. Kodiak’s go-to-market emphasizes the toughest environments (e.g., Permian Basin) where autonomy can deliver immediate ROI and where the company can collect differentiated edge-case data. It sells autonomy through a Driver-as-a-Service model intended to generate recurring revenue without owning fleets. Operationally, Kodiak is scaling: by end-2025 it logged 10,700+ paid driverless hours and expanded its Atlas deployment to 20 driverless trucks, but it remains deeply unprofitable with material funding and execution risk ahead of the pivotal H2 2026 long-haul launch.