Silicon and lidar and AI, oh my!
Announcements around these topics headlined Rivian’s first “Autonomy and AI Day” in Palo Alto on Thursday, where executives highlighted a slew of tech advancements the EV startup is slated to roll out in the near future.
Let the chips fall: Rivian execs announced the development of custom silicon that will underpin its next-generation autonomy platform and that they said would deliver cost, performance, and speed benefits.
“Our updated hardware platform, which includes our in-house 1600 sparse TOPS inference chip, will enable us to achieve dramatic progress in self-driving to ultimately deliver on our goal of delivering L4,” Rivian founder and CEO RJ Scaringe said in a statement.
Execs unveiled what they’re calling the Rivian Autonomy Processor, or RAP1, described in a news release as “a custom 5nm processor that integrates processing and memory onto a single multi-chip module.” The RAP1 will support Rivian’s third-gen autonomy computer, dubbed ACM3, which the company said will have the power to process 5 billion pixels per second.
The company plans to roll out the new third-gen autonomy hardware in R2 models beginning in late 2026.
“With our in-house silicon development, we’re able to start our software development almost a year ahead of what we can do with supplier silicon,” Vidya Rajagopalan, Rivian’s SVP of electrical hardware, said during a livestream of the event.
Additionally, executives announced that future models of its second-gen vehicle lineup, R2, will have lidar sensors integrated into the sensor suite, supplementing cameras and radar. Rajagopalan said the decision to introduce lidar came down to improvements in the sensors’ cost, performance, and design in recent years.
“Camera is the main workhorse of our sensor suite, generating the bulk of the data fed to the models,” she said, “but the radar and lidar are critical to addressing the edge cases which would otherwise create the long tail of problem cases.”
Hands off: Rivian unveiled the company’s Large Driving Model, or LDM, “a foundational autonomous model trained like a large language model.”
“With the deployment of our Gen 2 R1s, we began the process of building our data flywheel to grow and build our large driving model,” Scaringe said.
“Our approach to building self-driving is really designed around this data flywheel, where a deployed fleet has a carefully designed data policy that allows us to identify important and interesting events that we can use to train our large model offline before distilling the model back down into the vehicle,” he added.
To that end, Rivian plans to roll out software updates to second-gen R1 vehicles, soon to expand hands-free assisted driving across 3.5 million miles of roads in the US and Canada. Next year, the company will launch a subscription package for autonomy features priced at a one-time cost of $2,500, or $49.99 per month.
Scaringe said the Gen 3 autonomy platform also will be able to do point-to-point journeys, and sketched out a vision in which Rivian’s autonomy capabilities incrementally improve over time, eventually enabling eyes-off driving and fully autonomous driving—and maybe even give Rivian an opportunity to hop into the robotaxi race.
“While our initial focus will be on personally owned vehicles, which today represent a vast majority of the miles driven in the United States, this also enables us to pursue opportunities in the rideshare space,” Scaringe said.
Let’s chat: In early 2026, Rivian plans to debut a new AI voice tool called Rivian Assistant on Gen 1 and Gen 2 R1 vehicles. While the assistant will be able to connect to third-party apps like Google Calendar, it will be underpinned by Rivian Unified Intelligence, a “shared, multi-modal, and multi-LLM foundation.”
“Directly controlling our network architecture and our software platforms in our vehicles has of course created an opportunity for us to deliver amazingly rich software,” Scaringe said, “but perhaps even more importantly, this is the foundation of enabling AI across our vehicles and our business.”
This report was originally published by Tech Brew.
