Rivian Presents Custom Silicon and New Autonomy Architecture

Rivian announced its custom silicon, next-generation autonomy platform, and AI-driven software roadmap at its first Autonomy & AI Day. Learn what was presented.

Rivian has outlined one of its most ambitious technology roadmaps to date, unveiling a tightly integrated stack that spans custom silicon, next-generation autonomy, and deep artificial intelligence integration. The announcements were made during the company’s inaugural Autonomy & AI Day in Palo Alto, marking a clear shift toward full vertical control of its vehicle technology.

At the center of the presentation was the debut of the Rivian Autonomy Processor (RAP1), a custom-built 5nm chip designed specifically for vision-centric physical AI. The processor delivers up to 1600 sparse INT8 TOPS and can process five billion pixels per second, underscoring Rivian’s focus on camera-based perception at scale. RAP1 powers the company’s third-generation autonomy computer, the Autonomy Compute Module 3 (ACM3).

The RAP1 architecture introduces RivLink, a low-latency interconnect that allows multiple chips to be linked together, making the platform inherently extensible. Rivian also highlighted its in-house AI compiler and platform software, developed to optimize performance across its proprietary hardware. While the company confirmed compliance with automotive functional safety requirements, it did not disclose a specific ASIL level.

Rivian also confirmed plans to expand its sensor suite by adding LiDAR to future R2 models. The technology is intended to complement the existing multi-modal approach by adding high-resolution 3D spatial data and additional sensing redundancy, particularly for complex edge cases. The Gen 3 autonomy hardware platform, combining ACM3 and LiDAR, is currently undergoing validation, with production deployment on R2 vehicles expected by the end of 2026.

On the software side, Rivian detailed its Rivian Autonomy Platform, built around an end-to-end data loop for training. At its core is the Large Driving Model (LDM), a foundational autonomous driving model trained in a manner similar to large language models. The system uses Group-Relative Policy Optimization (GRPO) to distill effective driving strategies from large-scale datasets directly into vehicle behavior.

Some of these capabilities are set to reach customers in the near term. Rivian announced Universal Hands-Free assisted driving for second-generation R1 vehicles, enabling extended hands-free operation across more than 3.5 million miles of roads in the United States and Canada, including select non-highway routes with clearly marked lanes.

The new autonomy features will be offered through Autonomy+, a subscription service launching in early 2026. Pricing has been set at $2,500 as a one-time purchase or $49.99 per month. Rivian framed autonomy as an evolving software product, with a clearly stated trajectory from point-to-point driving toward eyes-off operation and, ultimately, personal Level 4 autonomy within defined conditions.

Beyond driving, Rivian introduced Rivian Unified Intelligence (RUI), a shared AI foundation designed to connect vehicle data, service operations, and user-facing experiences. A key outcome of this architecture is the Rivian Assistant, a next-generation voice interface scheduled to launch in early 2026 on first- and second-generation R1 vehicles.

The assistant runs on Rivian’s own edge models, augmented by frontier large language models, and relies on an in-house agentic framework to connect vehicle systems with third-party applications. Google Calendar was named as the first confirmed integration. The same intelligence layer will also support service diagnostics, analyzing telemetry and vehicle history to assist technicians and improve maintenance workflows.

Taken together, the announcements signal Rivian’s transition from isolated feature development to a unified, scalable platform approach. By aligning custom hardware, autonomy software, and AI-driven services under one architecture, the company positions this moment as a turning point toward making advanced autonomy a foundational element of the Rivian ownership experience.