No Priors: Artificial Intelligence | Technology | Startups Rivian’s Roadmap to AI Architecture and Autonomy with Founder and CEO RJ Scaringe
161 snips
Feb 12, 2026 
RJ Scaringe, founder and CEO of Rivian and leader of its electric vehicle and autonomy strategy, discusses Rivian’s shift to vertically integrated, software-defined vehicle architecture. He covers the move from rules-based stacks to neural nets, building in-house chips for affordable onboard inference, monthly OTA feature updates, and the R2 plan to bring a $45k mass-market EV to life.
AI Snips
Chapters
Transcript
Episode notes
Complete Reboot To Neural Net Architecture
- Rivian shifted from rules-based autonomy to a full neural-net, vertically integrated stack with a Gen 2 clean-sheet reboot in 2024.
- RJ Scaringe argues this AI-first architecture and data flywheel will accelerate autonomy far faster over the next few years.
Integrate Core Autonomy Components
- Vertically integrate critical vehicle systems when they are core to differentiation and require a closed training loop.
- Control sensors, on-vehicle data capture, and offline training to build a defensible autonomy stack.
The Brain Drives Cost, Not Sensors
- On-board inference compute is the dominant cost in an autonomy system, more so than sensors like cameras or LiDAR.
- Rivian built an in-house chip primarily to reduce cost and deploy high-level autonomy in every vehicle.