Kodiak AI adopts NVIDIA DRIVE Hyperion to scale driverless trucks
Company logs 10,700 driverless hours as it targets late 2026 long-haul launch
Kodiak AI, Inc. announced on March 16 that it will integrate NVIDIA's DRIVE Hyperion architecture into its next-generation autonomous trucking platform, a move the company says will accelerate deployment as it targets a commercial long-haul launch in late 2026.
The DRIVE Hyperion platform is built on two NVIDIA DRIVE AGX Thor centralized computers and delivers up to 1,000 INT8 TOPS of AI compute. It is designed to support generative AI, transformer, and VLM workloads at the edge.
"The key to deploying physical AI is harnessing the power of datacenter-scale AI and optimizing it to run at the edge," Kodiak AI Founder and CEO Don Burnette said. "At a time when our core AI capabilities and product footprint are rapidly expanding, NVIDIA DRIVE Hyperion delivers robust performance across a wide range of platforms for meeting the challenge of scaling autonomous vehicles."
Kodiak AI, a Mountain View-based autonomous vehicle technology company founded in 2018, currently operates 20 customer-owned driverless Class 8 trucks with no one in the cab, all deployed with sand logistics company Atlas Energy Solutions in West Texas's Permian Basin. The company reported Q4 2025 revenue of $1.1 million, a 37% quarter-over-quarter increase, driven by growth in driver-as-a-service (DaaS) revenue.
Those 20 trucks have logged more than 10,700 revenue-generating hours and delivered more than 12,600 loads, an 87% increase compared to year-end 2024. Burnette said on the March 10 earnings call that the company remains committed to delivering 80 additional trucks to Atlas over the coming quarters.
The NVIDIA partnership fits into Kodiak AI's broader hardware strategy. The company is also working with Bosch on a next-generation autonomous platform built for automotive-grade manufacturing. Rishi Dhall, Vice President of Automotive at NVIDIA, said the DRIVE Hyperion architecture is designed to support Kodiak AI's Level 4 (L4) autonomy stack across highway and off-road environments.
Expanding routes and customers
On the commercial side, the company launched a second driverless route with Martin Brower, a food and beverage logistics provider, between Dallas and El Paso. Burnette said the route brings total weekly operational lanes to eight and extends beyond a single Hours of Service window. "This new route represents a core tenet of our strategy of increasing our penetration with our existing customers," he said on the earnings call.
Kodiak AI also announced a new pilot with an undisclosed Fortune 500 private fleet hauling between Dallas and Houston, and said it is exploring international markets including Australia, Canada, and the Middle East.
The company's autonomy readiness measure, a proprietary metric tracking progress toward its long-haul safety case, stood at 84% at the end of Q4. Burnette said closing the remaining claims is a testing process, not a feature development one, and that the company remains on schedule.
CFO Surajit Datta guided full-year 2026 free cash flow at negative $160 million to negative $170 million, with spending concentrated on AV hardware and R&D ahead of the long-haul launch. "We are entering 2026 with momentum, a clear execution plan, and a capital-efficient business model designed to scale, drive towards profitability, and generate free cash flow over time," Datta said.