Kodiak lands first commercial deal to supply driverless trucks for public highways

Kodiak AI will deliver 75 self-driving trucks to DTL Transport beginning in 2028. The deal matters because it is Kodiak’s first binding commercial contract for driverless big rigs operating on public highways. It moves the company closer to selling autonomous trucking technology instead of operating every vehicle itself.

DTL is a California midmarket carrier that transports fresh produce and other goods to grocery chains. It plans to deploy the trucks on several interstate routes, including Dallas to El Paso and Interstate 20 between Dallas and Atlanta. Kodiak has already tested and operated on both routes.

The trucks will initially be part of DTL’s own fleet under Kodiak’s driver-as-a-service model. DTL will own and operate the vehicles, while Kodiak licenses its self-driving system and charges by the mile. The contract’s value was not disclosed, but Kodiak estimates $150,000 to $190,000 in annual recurring revenue per truck.

Here, “driverless” means the truck’s self-driving system is intended to control highway driving without a human safety operator in the cab. That is the important distinction in this contract: the vehicle is not merely assisting a person. Kodiak plans to remove human operators as it begins a driverless highway service.

During DTL’s California pilot, Kodiak autonomously hauled perishable goods for about 225 miles, but a human safety operator remained behind the wheel. The operator provided a fallback while Kodiak demonstrated that the system worked on Interstate 5 and California State Route 99.

The transition is not complete yet. Kodiak plans to start removing safety operators in December, beginning with freight for Ikea on Interstate 45 between Dallas and Houston. The DTL trucks are planned for delivery from 2028, so the contract represents a future fully driverless deployment rather than the supervised pilot already completed.

DTL’s contract covers 75 self-driving trucks. The scale matters because it gives Kodiak a sizable commercial customer while testing its shift toward an asset-light business. Instead of owning the vehicles, Kodiak would earn licensing revenue as DTL deploys the technology across multiple interstate routes.

DTL plans to start with Dallas to El Paso, Texas, and Interstate 20 between Dallas and Atlanta. Kodiak has already conducted testing and operations on both routes. The company’s forecast assumes annual recurring revenue of $150,000 to $190,000 per truck.

Using the lower end of that forecast, 75 trucks could produce at least about $11.2 million in annual recurring revenue once delivered. The estimate depends on assumptions including a licensing fee of about $0.97 per mile and roughly 225,000 miles per truck each year. Kodiak did not disclose the contract’s actual value.

Before removing safety operators on public highways, Kodiak must prove three things: that the technology works, that it is safe, and that it integrates with a customer’s fleet. Those demonstrations matter because DTL will eventually own and operate the trucks rather than rely on Kodiak as the carrier.

Kodiak used this progression with DTL. During a California pilot, its system autonomously hauled perishable goods about 225 miles between Fresno and Los Angeles. The route used Interstate 5 and California State Route 99, and a human safety operator remained behind the wheel.

Kodiak plans to begin operating a driverless highway service in December, carrying Ikea freight on Interstate 45 between Dallas and Houston. The article does not describe specific regulatory tests or approval standards. It does show that Kodiak is moving from supervised pilots toward driverless public-highway operations, while using commercial routes to demonstrate readiness.

Driver-as-a-service, or DaaS, makes Kodiak a technology supplier. The customer owns and operates the self-driving trucks, while Kodiak licenses its system and charges by the mile for the contract’s duration. DTL is Kodiak’s first on-highway DaaS customer.

Trucking-as-a-service, or TaaS, works differently. Kodiak owns and operates the autonomous trucks as a third-party carrier on behalf of customers. It currently carries freight for Charger USA, Roehl Transport, and Werner Enterprises on several Texas routes, although Burnette described those operations as pilots.

The difference is who controls the physical trucking operation. DaaS leaves fleet ownership and operations with the customer; TaaS leaves them with Kodiak. Kodiak used TaaS to prove its technology with its own trucks, then pursued DaaS to become asset light. Atlas Energy Solutions has also owned and operated Kodiak-equipped trucks since 2024, but in off-road West Texas operations.

Kodiak wants customers to own and operate the vehicles because it does not want to run a trucking company. Its stated goal is to become “asset-light,” meaning it sells or licenses the self-driving technology instead of owning and managing the trucks itself.

The company’s current TaaS model requires Kodiak to own and operate trucks as a third-party carrier. It carries freight for customers such as Charger USA, Roehl Transport, and Werner Enterprises on Texas routes. Under DaaS, customers take over vehicle ownership and operations, while Kodiak licenses its system and charges by the mile.

That change could let Kodiak focus on its technology while customers handle fleet operations. DTL’s agreement is an important step because it is Kodiak’s first on-highway DaaS contract. The article also identifies Aurora Innovation as a competitor pursuing the same asset-light direction. Kodiak’s shares have fallen 87% from their opening-day price, making the deal a potentially important benchmark.

Autonomous trucks generally use sensors to detect lanes, vehicles, road edges, signs, obstacles, and changing traffic conditions. Cameras, radar, and sometimes other sensing equipment provide the system with information about its surroundings. This lets the truck build a current picture of the road instead of relying on a human driver’s eyes.

Software combines those sensor readings with detailed maps and positioning data. Artificial intelligence helps identify objects, predict how nearby vehicles may move, and select actions such as steering, braking, accelerating, or changing lanes. Separate control systems then turn those decisions into movement. Maps help the truck understand the planned route and road layout.

The provided article does not describe Kodiak’s specific sensors, maps, or AI methods. It only identifies Kodiak’s self-driving system and its planned highway operations. In practice, the system would need to repeat this perception-and-decision cycle continuously, while meeting the company’s requirement to show that the technology works, is safe, and integrates into a fleet.