Kodiak bet driverless trucks would sell first in oilfields; Atlas took 100 of them

Kodiak builds the Kodiak Driver, an autonomous driving system it sells to fleet owners as Driver-as-a-Service: modular sensors, software, monitoring and updates on a subscription.

Driver-out autonomy sells first in unstructured off-road oilfields — no HD maps, private lease roads — reaching revenue years before regulated long-haul trucking.

About $243M raised as a private company before the merger with Ares Acquisition Corporation II at roughly $2.5B pre-money (announced April 2025), supported by over $110M in committed financing from investors including Soros Fund Management, ARK and Ares, plus about $551M of trust cash.

Don Burnette worked on self-driving at Google, left in early 2016 to co-found the truck-autonomy startup Otto with Anthony Levandowski and Lior Ron, and watched it get absorbed into Uber just as the Waymo lawsuit exploded. In 2018 he co-founded Kodiak Robotics to try again, with an asset-light plan: sell the driver, don't run the trucks.

Kodiak ran piloted freight on Texas highways but aimed its commercial launch at an unstructured, off-road environment: hauling frac sand in the Permian Basin for Atlas Energy Solutions. The partnership was announced in July 2024 after Kodiak's first driverless run there. In December 2024 Atlas received its first two Kodiak-equipped trucks and began driverless operations; by January 2025 it had delivered 100 loads with nobody in the cab. Burnette called it the first time a customer — not the AV company — owned and operated driverless vehicles. Atlas committed to an initial order of 100 trucks, and in April 2025 Kodiak announced a merger with Ares Acquisition Corporation II at about $2.5B pre-money.

By the end of 2025 Kodiak — public as Kodiak AI (Nasdaq: KDK) — had 20 customer-owned driverless Class 8 trucks, which it called the largest such deployment globally, had delivered more than 12,600 loads (up 87% year over year), and had logged over 10,700 driverless revenue-generating hours, roughly 700,000 miles. Q4 revenue was $1.1M, up 37% quarter over quarter, with mid-single-digit-millions ARR from Driver-as-a-Service — early money against $34M of quarterly free-cash burn.

Off-road oilfields offer constant, short-haul demand on private lease roads — autonomy can operate there without public-road regulators slowing deployment to a crawl.

The customer-owned, subscription model turns capital-hungry fleet operations into a software-like business and makes each deployment a case study the next customer can buy.

Burnette had already lived the failure mode of operating trucks at a startup (Otto-Uber) and chose to sell the driver rather than run freight.

The Permian deployment doubled as proof for the harder prize: the same Kodiak Driver architecture underpins the long-haul highway safety case now 93% complete.

Going public while revenue-generating — even at $1.1M a quarter — separated Kodiak from the pre-revenue SPAC class that destroyed earlier AV names.

Regulated public roads are the slow lane to revenue. Kodiak sold autonomy on private oilfield roads first, with the customer owning the fleet — and the niche became the reference case for long-haul.

As of September 2026, 35 Kodiak-equipped trucks were running driverless in the Permian Basin (end of Q2), the Atlas order was heading toward its 100-truck commitment, and the long-haul safety case stood at 93% complete, up from 84% in February, with driverless long-haul service targeted before year-end. The company added Bosch (redundant platform) and Verizon (fleet connectivity) partnerships, launched a route with Martin-Brower, and expanded to eight weekly lanes. TipRanks noted liquidity visibility only into late 2026, with cash burn still the central risk.