Self-driving truck company Kodiak AI announced on October 30, 2026 that its Autonomy Readiness Measure, an internal metric tracking progress toward a fully driverless launch, has reached 99%, up from 84% in February. The company says it remains on pace to launch unsupervised, driver-out commercial highway trucking by the end of 2026, starting on Interstate 45 between Dallas and Houston, a 219-mile corridor, with IKEA as the launch customer, according to Kodiak AI.
What exactly is the Autonomy Readiness Measure?
It's not a measure of miles driven or crashes avoided. Kodiak defines it as the percentage of claims and evidence in its safety case that are materially complete. A safety case is the formal, documented argument a self-driving company builds to justify why its system is safe enough to operate without a human behind the wheel, broken into individual claims, like "the system detects stopped vehicles in time to brake safely," each of which needs supporting evidence, like simulation results, test-track data or real-world miles, before regulators and the company's own engineers consider it closed.
Kodiak's score has climbed steadily through the year: 84% in February, 86% in April, 91% in July, 93% in August, 96% in September, and now 99% in October. Director of safety Frank Fratrik described the final stretch plainly: "Each claim closes on evidence, and that evidence is nearly there." The remaining percentage points represent the hardest, most stubborn claims to prove, not an evenly distributed final lap.
Why measure readiness this way instead of just announcing a launch date?
Target launch dates for driverless vehicles have a history of slipping, often because a handful of safety claims take longer to validate than engineers expect. By publishing a granular, updating score instead of a single promised date, Kodiak gives investors, partners and regulators a way to track real progress rather than wait for the company's word that it's ready. It's a transparency play aimed at a sector where skepticism about driverless timelines runs high.
Underpinning the score is Kodiak's BreakPoint AI tool, which runs simulations designed to surface edge-case driving scenarios, rare situations like a tire blowout next to a merging lane, that a truck might encounter once every few hundred thousand miles on the road but that the safety case needs to account for before launch. Kodiak also uses a Probabilistic Risk Assessment methodology to quantify how often and how severely those rare scenarios could cause harm, turning hard-to-observe risks into numbers that feed directly into the readiness score.
How does this fit into Kodiak's broader rollout?
The 99% score isn't Kodiak's only evidence of progress. The company already runs 35 trucks with no humans in the cab in the Permian Basin oilfields of West Texas, hauling for energy producer Atlas Energy Solutions, a lower-speed, more controlled environment than interstate highway driving. The I-45 corridor for IKEA represents a harder test: higher speeds, denser traffic and the unsupervised long-haul trucking Kodiak has been building toward. CEO Don Burnette called the moment the final stretch: "Our momentum is compounding as we conduct our final preparations ahead of launching driverless long-haul service."
If Kodiak hits its end-of-2026 target, it would mark one of the first unsupervised, driver-out commercial highway trucking launches in the U.S., a milestone competitors including Aurora and Waabi are racing toward on their own timelines with different corridors and customers.