TechCrunch has learnt that Rivian has been developing its own AI helper for almost two years; this endeavor is distinct from its multibillion-dollar technological joint venture with Volkswagen.
Rivian has not disclosed when customers would be able to use the AI helper. However, Rivian’s software director Wassym Bensaid told TechCrunch earlier this year that the company was aiming for the end of the year. During its next AI & Autonomy Day, which will be livestreamed beginning at 9 a.m. PT on December 11, the business will probably reveal more.
Rivian’s intentions are in line with the current state of affairs, where sectors are rushing to catch up with the rapid speed of development from fundamental AI businesses, which include tech giants and startups like Anthropic, Google, Microsoft, Meta, and OpenAI that are constructing the key models and infrastructure.
However, as Bensaid pointed out to TechCrunch earlier this year, this is hardly a hasty attempt to keep current. It is also more than just an infotainment system chatbot. Bensaid noted that the device is intended to be connected with all vehicle controls and that the firm has invested a significant amount of time, money, and thinking into it.
According to Bensaid, the firm began with the fundamental idea of creating an overarching architecture that is platform and model neutral. The Rivian AI assistant team, which operates out of the company’s Palo Alto headquarters, quickly came to the conclusion that development of the software layers that assist in coordinating different processes as well as the control logic that resolves conflicts should also receive time and attention.
Bensaid added, “And that is the in-vehicle platform we have constructed.” “We employ what the industry loves to term an agentic framework, but we have considered that design from very early so that we can connect with diverse models.”
The internal AI assistant initiative aligns with Rivian’s efforts to increase vertical integration. Rivian completely redesigned their flagship R1T truck and R1S SUV in 2024, altering everything from the electrical architecture, sensor stack, and software user interface to the battery pack and suspension system.
The business has also invested a significant amount of money in creating and enhancing its own software stack, which consists of a layer pertaining to the infotainment system and everything associated with real-time operating systems (RTOS) that control the vehicle, including thermal dynamics, ADAS, and safety systems.
Bensaid stated that the AI helper consists of a variety of models that perform distinct functions, but he did not go into specifics. The end product is a hybrid software stack that blends cloud AI, which uses distant servers to handle complex models that need more processing power, with edge AI, which does jobs on the device.
This should result in an adaptable, personalized AI helper that divides the work between the cloud and the edge.
A large portion of the AI software stack was created by Rivian itself, including its own proprietary models and the “orchestration layer,” which functions as a kind of conductor or traffic cop to ensure that the many AI models cooperate. Rivian hired other businesses to perform particular agentic AI tasks.
According to Bensaid, the goal is to create an AI assistant that boosts consumer involvement and trust.
The AI helper will remain in Rivian for the time being. The company’s software-focused joint venture with Volkswagen has nothing to do with automated driving or AI assistants.
The basic electrical architecture, zonal computation, and infotainment are the main focus of the $5.8 billion technological joint venture with Volkswagen, which was announced in 2024. As early as 2027, the Volkswagen Group is anticipated to get software and electrical architecture from the joint venture, which formally began operations in November 2024.
Although AI and autonomy are now distinct, “that does not imply that it may not be in the future,” according to Bensaid.
