Rivian CEO says third-party AI agents will run in its cars

Rivian has already launched an in-car AI assistant that lets drivers and passengers control their EVs by voice. But CEO RJ Scaringe said at TechCrunch Disrupt this week that the company eventually wants to let AI agents built by other companies, such as Anthropic, run inside its vehicles. He described that future as "magical," though he did not give a timeline and pointed to Rivian's AI & Autonomy Day in December for more details.

Rivian's current system uses an in-house agentic framework that decides which model to use for a request, routing simple tasks to a local edge model and more complex ones to the cloud. It can adjust temperature, activate seat warmers, summarize news, and, as of May, access external apps like Google Calendar. Scaringe compared the routing to a traffic light.

Adding third-party agents will not be easy, he said, because it requires "a lot of plumbing" both inside Rivian and with outside AI labs. Safety is a major constraint: agents will not be allowed to control braking or steering. Instead, the idea is to extend the assistant's existing context, such as pulling a conversation from another device to help plan a meal while driving.

Rivian is not the first American automaker to add AI agents to its cars—Tesla already integrated Grok. But Rivian's approach stands out because it built its own software stack from the ground up rather than relying on Apple CarPlay or Android Auto, giving it more control over how agents interact with the vehicle.