Einride is expanding its collaboration with NVIDIA to evaluate a new artificial intelligence driving model designed to help autonomous trucks interpret complex driving conditions and make decisions. The company plans to begin evaluating the technology on the road in the second half of 2027.
The collaboration will incorporate NVIDIA’s Alpamayo AI reasoning model into the Einride Driver, the company’s proprietary autonomous driving system. Einride plans to be the first trucking company to deploy Alpamayo on NVIDIA’s Hyperion platform, building on an existing technology partnership announced in September.
Alpamayo is a vision-language-action (VLA) model that interprets driving environments, proposes vehicle movements and explains the reasoning behind those decisions. The technology is intended to supplement Einride’s autonomous driving capabilities while operating within the company’s existing safety architecture, which retains full control over the vehicle.
Einride will adapt the model using its own operational data collected from vehicles and freight routes. This process will include additional training and optimization to better align the technology with the requirements of heavy-duty trucking.
The company plans to initially evaluate Alpamayo in “shadow mode,” allowing the AI model to operate alongside the existing driving system without controlling the truck. During testing, engineers will compare its decisions with those generated by Einride’s driving planner and human drivers.
This approach will allow Einride to assess the model’s performance using real-world driving information while maintaining the authority of its established vehicle control and safety systems.
“Being the first trucking company in this engagement gives us the opportunity to help shape the model for freight,” said Henrik Green, chief technology officer at Einride.
The project builds on the companies’ September 21 announcement outlining plans to develop the next generation of the Einride Driver using NVIDIA’s Hyperion autonomous vehicle platform.
That partnership also includes NVIDIA’s Halos safety system and the use of NVIDIA Cosmos and Blackwell computing infrastructure to train and validate autonomous driving models.
According to Einride, incorporating Alpamayo could help shorten development cycles and reduce the cost of adapting its autonomous technology to additional vehicle types and freight routes. The company’s longer-term objective is to expand autonomous operations across highway and point-to-point routes within its existing freight network.
The collaboration also gives NVIDIA access to Einride’s operational experience and freight data as the companies evaluate how reasoning-based AI can address more complex commercial vehicle driving situations.
For fleet operators considering autonomous trucking, the development highlights ongoing efforts to improve the ability of autonomous driving systems to handle different operating environments and route conditions. The planned shadow-mode testing will provide an opportunity to evaluate the technology before it assumes any driving responsibilities.
Einride is targeting a live road evaluation in the second half of 2027. The company has not identified the specific routes or customers that will participate in the evaluation or announced a timeline for commercial deployment of Alpamayo-equipped trucks.
