Rivian Automotive Inc. is making significant strides in the field of autonomous driving, focusing on an end-to-end artificial intelligence approach that could redefine its vehicles’ capabilities. The company recognized the potential of transformer-based AI technology in 2021, prompting a strategic shift from its traditional rules-based driver assistance system to a more innovative framework. Rivian’s Large Driving Model (LDM) is about to go live. This innovative tech will enable Rivian vehicles to travel down tricky roads without requiring prior established direction such as lane lines.
Later this month, Rivian will finally be rolling out the Universal Hands-Free update. This change will improve the driver experience and set the stage for autonomy down the road. Future training of the LDM will use aggregated fleet data. This will allow it to run on more than 3.5 million miles of paved roads within the U.S. and Canada, focusing primarily on roads where painted lines are most apparent. This initiative aligns with CEO RJ Scaringe’s long-held vision, articulated since 2018, for Rivian’s vehicles to achieve full self-driving capabilities.
Rivian’s pivot towards an advanced driving software system took shape with the introduction of its second-generation R1 vehicles, which are equipped with Nvidia’s Orin processors. As Rivian continues to innovate, it plans to replace these chips with a custom autonomy computer designed for its R2 vehicles, expected to launch in 2026. If that’s the future, the new autonomy computer and lidar technology won’t be ready till long after the R2 debuts. The vehicle will continue to demonstrate a new “point-to-point” driverless capability driven by cutting-edge software. This feature enables hands-off driving of the vehicle while expecting drivers to stay attentive and be in a position to resume control if needed.
Scaringe acknowledges that the R2 model is critical for Rivian’s future success and emphasizes the importance of its sales performance. He notes that early buyers of the R2 may face limitations in terms of software features, stating:
“So [if] you’re buying an R2 and you buy it in the first nine months, it’s just going to be more constrained.” – RJ Scaringe
Additionally, the company anticipates some customers will want to wait for further functionality. At the same time, consumers could purchase the R2 today and exchange it in the future for a new and improved model. Scaringe commented on this consumer behavior:
“I think what will happen is some customers will say ‘that matters a lot to me, and I’m going to wait.’ And some will say ‘I want the newest, best things now, and I’m going to get the R2 now, and maybe I’ll trade it in a year or two.’” – RJ Scaringe
This forward-looking approach is designed to help consumers make informed choices about what they buy in a rapidly changing technological environment.
As everyone knows, Rivian has tremendous – some might say outrageous – ambitions. Scaringe admits that the timelines for launching vehicles and developing its autonomy platform don’t exactly align. He remarked:
“In a perfect world, everything times at the same time, but the timeline of the vehicle and the timeline of the autonomy platform are just not perfectly aligned.” – RJ Scaringe
As language goes, this LDM represents a radical break from Rivian’s previous formulations of control over their vehicle. In the past, they relied on deterministic programming and human-determined routes and strategies. Scaringe explains this transformation:
“Everything that the vehicle did was the result of a prescribed control strategy written by humans.” – RJ Scaringe
Under this new approach, with Rivian’s AAA support, it hopes to develop a testing framework that better accounts for more complicated driving conditions. And while the announcement was groundbreaking, Scaringe underscored that the company’s core focus remains on building autonomous capabilities for public roads. It won’t be hurrying to introduce this technology to more difficult settings just yet.
“We’re not putting any resources into rock crawling autonomously,” – RJ Scaringe
Instead, he pointed out that the aim is to ensure vehicles can reach challenging locations without fully autonomous capabilities on rough terrains:
“But in terms of getting to the trail head? For sure.” – RJ Scaringe
Rivian, like any newly public company, is eager to look ahead. They plan to deploy robotaxis within a few years as they iterate on their training models and expand their operational design domains (ODD). Scaringe remains optimistic about achieving true autonomy where drivers are not required to intervene:
“Then, it becomes a bit of a like, what’s the ODD [operational design domain]? Dirt roads, off road? Easy.” – RJ Scaringe
Rivian is clearly deep in development of AI-driven assistive tech and getting ready for the rollout this fall. Industry insiders are watching with bated breath to see how these advances will affect consumer demand and overall market success.