Audio · 2026-02-12 · 32m · 6 moments

Rivian’s Roadmap to AI Architecture and Autonomy with Founder and CEO RJ Scaringe

Autonomous vehicle technology has moved past human-coded rules and into an era of neural networks and custom computer chips. And to solve the most difficult driving scenarios, electric vehicle company Rivian abandoned its original technology platform to build a vertically integrated data stack. Sarah Guo sits down with Rivian Founder and CEO RJ Scaringe to explore the seismic shift in the automotive industry toward AI-driven, software-defined vehicles . RJ discusses the move away from function o ✦ AI generated

Apple Podcasts ↗

GUEST RJ Scaringe HOST Sarah Guo HOST Elad Gil

timeline · colored by role

Mechanism◆

Autonomous driving technology has moved past rules-based systems into an era of neural networks and custom silicon, making the last three years and the next three years look completely different in terms of progress.

RJ Scaringe explains that Rivian completely reset its autonomy platform in 2022, moving from a third-party rules-based system to a clean-sheet neural network approach with in-house hardware, and argues that the rate of progress in autonomy from 2025 to 2030 will far outpace the prior five years.

transcript

RJ Scaringe: We used a third-party, a front-facing camera that was essentially a third-party solution that then plugged into an overall frame like that we've built, but it was all rules-based. So the cameras fed a rules-based planner, the planner would then make a bunch of decisions around the feeds from the perception. And it was, you know, the moment we launched, we knew it was the wrong approach, but it was the thing we'd started working on before the launch. And so at the end of 2021, beginning of 2022, we made the decision to completely reset the platform. ... We made the decision to redo it like clean sheet, no legacy of what we had built in the Gen 1. ... Not a single line of shared code, not a single piece of common hardware on the perception or on the compute side. ... The rate of progress that we saw in autonomy between, let's say, 2020 and 2025, or 2021 and 2025, and what we're going to see between today and let's say 2029, 2030 are completely different slopes, and that really comes back to entirely new architectures now being used to develop self-driving, actually truly AI architectures, whereas before these were not AI architectures in the true sense, they were rules-based environments that we defined as humans, we codified them.

Claim◆

Fewer than five companies outside of China have all the necessary ingredients to succeed at neural-network-based autonomous driving: a large vehicle fleet for data, control of the perception stack, and the capital for GPUs and custom inference chips.

Scaringe argues that vertically integrated autonomy is the only viable path, and that there are more than one but fewer than five companies (including Rivian, Tesla, and Waymo) that possess the fleet size, sensor control, data architecture, and capital to compete in neural-network-based autonomy.

transcript

RJ Scaringe: When you really look at what's necessary to be successful in a neural net-based approach, there's a core set of ingredients that very few people have. ... First and foremost, you need to have complete control of the perception platforms. ... The system needs to be capable of triggering unique or interesting or noteworthy events that you can then use to train. ... Companies that are either developing independent solutions that are not a car company, they typically don't have access to the type of mileage that we do. ... I think there's more than one, less than five companies outside of China that have the necessary ingredients to do this, the capital, the GPUs, the car park with enough vehicles, generate enough data. ... I would include all three of those [Tesla, Waymo, Rivian], yeah, and there's maybe one or two others in the mix. ... A lot of the solutions that are more 1.0 based and are sort of stuck in that framework, I think have a truly a 0% chance of progressing to be competitive with a neural net-based approach.

explains mechanism · 3provides context · 1

Fact

The onboard inference chip is the most expensive part of an autonomous driving system — an order of magnitude more costly than the entire perception stack — which is why Rivian built its own custom chip.

Scaringe explains that while sensors like cameras, radar, and LiDAR have become cheap, the onboard inference compute is the dominant cost driver, motivating Rivian to design a custom chip in-house to enable deployment across every vehicle.

transcript

RJ Scaringe: We design and spec and build the cameras. Radars are extremely cheap. LiDARs are now very, very cheap. But the really expensive part of the system is actually the onboard inference. And so that's like an order of magnitude more expensive than any of the perception stack. I think people focus on the perception because it's the things we can visualize. But the brain is actually the most expensive part. And so we brought that in-house as a way to remove cost from the system so that we can easily deploy this on every car.

provides context · 1

Prediction◆

By 2030 it will be inconceivable to buy a car that cannot drive itself, and the difference between Level 2, 3, and 4 systems has collapsed to just how many extreme corner cases the system can handle.

Scaringe describes how the traditional delineation between Level 2, 3, and 4 autonomy has faded — the systems now share the same perception and compute hardware, differing only in coverage of rare corner cases. He predicts that within a few years, autonomous capability will be as expected as airbags or air conditioning.

transcript

RJ Scaringe: What's happened is those two worlds just, I think, have just started to very clearly merge where the delineation between a level 2, a level 3, and a level 4 in terms of perception and in terms of compute has started to fade. And it's now essentially just remove like how capable the system is at addressing all these corner cases. ... If you're driving a level 2 system or a level 3 system or a level 4 system, for 99.9999, like 49, they're identical. The difference is like the 5th or 6th or 7th 9 on that is these like extreme corner cases. ... I think by 2030, it'll be inconceivable to buy a car and not expect it to drive itself. ... Sort of like it's hard to imagine buying a car today without airbags or buying a car today without air conditioning. These things at a moment in time were optional. I think in not too much time, a couple of years, it'll be hard to conceive buying a car that can't drop you at the airport or pick up your kids from school.

extends · 1supports · 2

Mechanism◆

Every car on the road except Tesla and Rivian uses a domain-based architecture with 100-150 separate ECUs, which is a disastrous legacy that makes software updates nearly impossible — the reason Volkswagen paid Rivian $5.8 billion to license Rivian's zonal architecture.

Scaringe contrasts the traditional domain-based architecture (100-150 separate computers, each running supplier-written software on its own island) with Rivian's zonal architecture (1-3 computers running a single OS). He cites Rivian's $5.8 billion deal with Volkswagen Group as evidence that legacy automakers recognize they cannot build this capability internally.

transcript

RJ Scaringe: With the exception of Tesla and Rivian, every car on the road has what is called a domain-based architecture. ... In a modern car, you might have 100 to 150 of these. And each of these run their own little island of software. And that little island of software is written by a supplier, more likely a supplier to the supplier. ... That's why it's impossible to debug like a software system. It's also why it's really hard to do an update. ... In contrast, on an approach where you build a zonal architecture, where you have a very small number of computers, really, one, two, maybe 3, depending on the size of the car, that are running one operating system that control everything, it's very easy. ... It actually goes back to fuel injection systems. ... Up until early 1960s, like every car on the road was completely analog. ... The first computers were there to drive the fuel injection systems and car companies said, this isn't a core competency. Let's push that little computer to run the fuel injection system to a supplier. ... And it just like grew into this absolute disastrous mess that is today the network architecture that's in truly every car on the road with the exception of two companies. ... We did a large software licensing deal, a $5.8 billion deal with Volkswagen Group, the second largest car company in the world, to essentially leverage our network architecture and ECU topology for all their various brands.

Claim

EV adoption in the United States is low not because of consumer resistance but because of an extreme lack of compelling choices outside the Tesla Model 3 and Model Y, and the industry has failed by copying the Model Y instead of creating differentiated products.

Scaringe argues that the 8% EV adoption rate in the US is driven by insufficient product variety — there are over 300 ICE vehicle choices under $70,000 but fewer than three great EV choices. He criticizes competitors for copying the Model Y外形 instead of offering distinct form factors, and positions Rivian's R2 and the Volkswagen partnership as expanding consumer choice.

transcript

RJ Scaringe: The overall adoption rate in the United States of EVs is around 8%. The vast majority of vehicle buyers are buying vehicles that are under $70,000 with the average sale price of about 50. And so if you look at the number of vehicle choices you have at a price point that's under $70,000, ... there's well in excess of 300 different vehicle model line choices. ... And in the EV space, I think, and this is, I think there's more than one, less than three great choices. ... Because of the success of the Model Y, in particular, the EV choices that do exist that are outside of Tesla are often very similar to a Model Y. ... The world doesn't need another Model Y. The world needs another choice. ... So I think my view is the EV adoption in the United States is a reflection of the lack of choice.

Highlight slides

Rivian's Autonomy Reset✦ from: Autonomous driving technology has moved past rules-based systems into an era of neural networks and custom silicon, making the last three years and the next three years look completely different in terms of progress.Two Eras of Progress✦ from: Autonomous driving technology has moved past rules-based systems into an era of neural networks and custom silicon, making the last three years and the next three years look completely different in terms of progress.Vertical Integration Is the Only Path to Neural-Net Autonomy✦ from: Fewer than five companies outside of China have all the necessary ingredients to succeed at neural-network-based autonomous driving: a large vehicle fleet for data, control of the perception stack, and the capital for GPUs and custom inference chips.The Three Essential Ingredients✦ from: Fewer than five companies outside of China have all the necessary ingredients to succeed at neural-network-based autonomous driving: a large vehicle fleet for data, control of the perception stack, and the capital for GPUs and custom inference chips.Who Makes the Cut (Outside China)✦ from: Fewer than five companies outside of China have all the necessary ingredients to succeed at neural-network-based autonomous driving: a large vehicle fleet for data, control of the perception stack, and the capital for GPUs and custom inference chips.The L2/L3/L4 Delineation Has Collapsed✦ from: By 2030 it will be inconceivable to buy a car that cannot drive itself, and the difference between Level 2, 3, and 4 systems has collapsed to just how many extreme corner cases the system can handle.Autonomy Will Be Standard by 2030✦ from: By 2030 it will be inconceivable to buy a car that cannot drive itself, and the difference between Level 2, 3, and 4 systems has collapsed to just how many extreme corner cases the system can handle.The legacy car: 100–150 isolated computers✦ from: Every car on the road except Tesla and Rivian uses a domain-based architecture with 100-150 separate ECUs, which is a disastrous legacy that makes software updates nearly impossible — the reason Volkswagen paid Rivian $5.8 billion to license Rivian's zonal architecture.Rivian's zonal architecture vs. the old way✦ from: Every car on the road except Tesla and Rivian uses a domain-based architecture with 100-150 separate ECUs, which is a disastrous legacy that makes software updates nearly impossible — the reason Volkswagen paid Rivian $5.8 billion to license Rivian's zonal architecture.

Related episodes

Video · a16z Podcast

Autonomous driving technology has moved past rules-based systems into an era of neural networks and custom silicon, making the last three years and the next three years look completely different in terms of progress.

Fewer than five companies outside of China have all the necessary ingredients to succeed at neural-network-based autonomous driving: a large vehicle fleet for data, control of the perception stack, and the capital for GPUs and custom inference chips.

The onboard inference chip is the most expensive part of an autonomous driving system — an order of magnitude more costly than the entire perception stack — which is why Rivian built its own custom chip.

By 2030 it will be inconceivable to buy a car that cannot drive itself, and the difference between Level 2, 3, and 4 systems has collapsed to just how many extreme corner cases the system can handle.

Every car on the road except Tesla and Rivian uses a domain-based architecture with 100-150 separate ECUs, which is a disastrous legacy that makes software updates nearly impossible — the reason Volkswagen paid Rivian $5.8 billion to license Rivian's zonal architecture.

EV adoption in the United States is low not because of consumer resistance but because of an extreme lack of compelling choices outside the Tesla Model 3 and Model Y, and the industry has failed by copying the Model Y instead of creating differentiated products.

Is Defense the Next Trillion-Dollar Category? | a16z American Dynamism Summit

Fable's Back, AI Engineer Recap, & SambaNova

The New Rule for Picking AI Winners | The a16z Show