Kodiak AI (NASDAQ:KDK) is focused on building self-driving technology for commercial applications, with an emphasis on long-haul trucking, industrial and off-road operations, and military and defense use cases, founder and CEO Don Burnette said during a discussion at a company event. Burnette, who said he has worked in autonomous driving for more than 17 years, described Kodiak as an “AI solution to driving vehicles” that has been operating for nearly eight years.
Current deployment and near-term expansion plans
Burnette said that, as of the company’s latest reported quarter (Q3), Kodiak had 10 driverless trucks operating for customers with no one in the cab and no required remote monitor. He emphasized that these trucks are owned and operated by the customer and are running “around the clock” in Texas, including through challenging weather conditions such as dust storms.
Sensor redundancy and “multi-pathway” validation
In contrast to camera-only approaches discussed in the broader autonomous vehicle market, Burnette said Kodiak uses radar, cameras, and LiDAR for commercial deployments, arguing that multiple sensing modalities improve safety when the business case can support the added cost. He added that redundancy is built throughout the system, including not only sensing but also compute and vehicle controls such as steering, braking, and power, which are critical when no human is onboard.
Addressing questions about how autonomous systems resolve conflicting sensor inputs, Burnette said Kodiak’s implementation relies on “confidence through consensus.” He described a structure with “many, many different parallel pathways,” citing 20-plus learned AI pathways that reason about the environment using different inputs (for example, camera-only, LiDAR-only, and radar-only). He said those parallel outputs help validate end-to-end AI models and mitigate concerns about “black box” decision-making by providing additional systems that can be queried and probed.
Permian Basin operations and product learnings
Burnette discussed Kodiak’s industrial deployment in the Permian Basin with Atlas, noting the operations are primarily on private roads. He said private-road operations allow significantly heavier loads than public highways, citing public road restrictions around 80,000 pounds. Burnette added that Atlas has announced it is pulling doubles and is working on triples, which would not be permitted on public roads. He characterized the choice of private roads as driven by economics and legal limits rather than technical constraints, adding that “a road is a road to the robot.”
He also highlighted the operational setup supporting Atlas’ efficiency, including a 42-mile conveyor belt across the desert called the Dune Express, which moves sand above and around public roads to reach well sites and maximize truck utilization.
Beyond autonomy technology, Burnette said Kodiak has learned significant “product” lessons from real-world deployment—how customers actually use a driverless truck, how interfaces should work, and what is required to “industrialize” the system for day-to-day operations. He framed the company’s progress around “three pillars of autonomy”:
- Technology
- Safety case
- Product
Burnette said Kodiak views the practical learnings from real customer use as a competitive advantage that can transfer to future long-haul deployments.
Unit economics: labor, insurance, utilization, and safety
Burnette said customer interest in autonomous trucking is influenced by multiple factors, including driver availability—particularly a shortage of “quality drivers”—and the costs and operational burden of hiring, retention, turnover, and sign-on bonuses. He said Kodiak aims to provide early adopters an “immediate cost discount,” acknowledging there can be inefficiencies in early-stage deployments.
He also pointed to insurance costs and “nuclear verdicts” as a growing industry concern, arguing that continuous 360-degree data capture could help reconstruct incidents with detailed information about positions and vehicle dynamics. Burnette said the company is already seeing insurance parity with human-driven fleets in its Atlas deployment, though he expects pricing to improve as more driverless miles accumulate—something he said insurers need before changing rates materially.
Burnette also emphasized asset utilization as a potential benefit, noting that freight networks have been designed around human hours-of-service constraints. He said autonomy could ultimately enable network redesign and higher utilization by allowing vehicles to operate continuously, subject to refueling needs.
Scale-up efforts, partners, and technology approach
When asked about revenue and profitability, Burnette said the company is in a quiet period ahead of upcoming Q4 earnings and could only discuss Q3. He characterized revenue as “still on the small side” given the scale of deployments, but said generating revenue at all is a milestone in an industry he described as largely pre-revenue.
Burnette said Kodiak has been building out the supporting ecosystem needed to scale, including a manufacturing partner, Roush, and noted the company recently brought on Bosch as its tier-one supplier for future generations of its system. He said that relationship “unlocks global scale” for future volumes.
Explaining the company’s “Kodiak Driver” package, Burnette described SensorPods mounted near where truck mirrors would be, housing LiDAR, cameras, radar, and related electronics, along with a compute stack inside the truck. He reiterated the company’s focus on redundancy, including dual independent battery banks and electrical interfaces and backup systems for steering and braking. He said the system is modular and has been adapted to different vehicle platforms, including heavy trucks, passenger vehicles, and military vehicles.
On differentiation, Burnette said Kodiak does not rely on high-definition maps, which he described as costly to create and maintain. He also discussed Kodiak’s work on bringing modern AI architectures—drawing on data-center-scale research—down to low-power compute suitable for trucks operating in safety-critical environments.
In discussing adoption timelines, Burnette said that within five years, autonomous trucks should be a visible presence on certain corridors, particularly across the southern United States, though not ubiquitous across all vehicles.
About Kodiak AI (NASDAQ:KDK)
We are a blank check company incorporated as a Cayman Islands exempted company for the purpose of effecting a merger, share exchange, asset acquisition, share purchase, reorganization or similar business combination with one or more businesses, which we refer to as our initial business combination. Our only activities since inception have been organizational activities and those necessary to prepare for this offering. We have not selected any business combination target and we have not, nor has anyone on our behalf, initiated any substantive discussions, directly or indirectly, with any business combination target.
