Kodiak AI Partners With AMD To Introduce High-Performance Processors For Autonomous Truck Platforms!

AMD EPYC processors will be responsible for aggregating and preprocessing sensor data from LiDAR, cameras, and millimeter-wave radar, while also providing path planning support for the Kodiak Driver autonomous driving system. Kodiak stated that the company is the first company to deploy AMD EPYC processors in its driverless truck hardware platform. Kodiak stated that physical AI requires powerful computing capabilities to quickly process massive sensor data while supporting increasingly complex AI workloads. AMD offers industry-leading CPU technology that meets the demands of advanced autonomous driving applications in terms of performance, scalability, and energy efficiency.

Compared to the previous generation, AMD EPYC processors feature higher clock speeds, enabling faster data processing speeds at lower power consumption. These processors handle sensor processing, positioning, and path planning workloads that are highly latency-sensitive and difficult to parallelize. AMD's Senior Director of Automotive Markets stated that this collaboration fully demonstrates the tremendous potential of combining autonomous driving software with high-performance computing power.

Kodiak currently operates a fleet of driverless vehicles in the Permian Basin, with no drivers required onboard. The company stated it is working to bring driverless trucks onto public highways later this year. Kodiak's commercial

partners and customers include well-known companies such as Atlas Energy Solutions, IKEA, J.B. Hunt, Bridgestone, Werner Enterprises, C.R. England, General Dynamics Land Systems, and Roehl Transport. In 2024, Kodiak has deployed autonomous technology on customers' own driverless semi-trucks.

Raymond James has upgraded AMD's stock rating to a strong buy. The agency extends its AI factory analytics framework to the server CPU market, forecasting revenue to grow at a four-year compound annual growth rate of 44%, reaching approximately $201 billion by the 2030 calendar year. This growth is driven by three key drivers: traditional data center demand, CPUs that carry and coordinate accelerators, and CPUs supporting agent workloads.

Market Insight:

The CPU increasingly takes on the role of managing accelerators and handling workloads around each model call such as retrieval, databases, applications, security, sandboxing, and tools. Because higher utilization, improved software efficiency, custom chips, and task offloading can absorb some incremental growth, workload growth may not correlate one-to-one with processor shipments.