Oct. 12, 2025 | Bonus Episode: Rivian CEO RJ Scaringe says software critical to automaker survival

Want to win the inventory game? With Kelley Blue Book Instant Cash offer, you can lock in trade-ins before your competitors even blink. Faster deals, better margins, lock in and load up today at b2b.kbb.com.

RJ Scorinch sat down with Axios reporter Joanne Muller for a fireside chat with the Automotive Press Association. On Thursday's show, we brought you a short story about his comments on the importance of software-defined technology architectures in the coming years. He says he's strongly convicted that automakers will need to have SDV architectures in the next decade to survive. But what exactly does that mean? And why is he so bullish on it? On this episode, we wanted to bring you

that full portion of the conversation, because his explanation is pretty interesting and illuminating, at least when it comes to his point of view on all of this. Here's that piece of Rivian CEO, RJ Scorinch's conversation with Joanne Muller.

There's a number of companies in China, I'd say, more than five, less than ten, that have technology that's very different than the rest of the world.

And it's because they started recently, and these are new companies, and if you're starting a company from scratch, you've designed things in a different way.

And if you allow me, I'll just describe what I mean by that.

Yeah, please, I'm very curious.

So, when you look at the electronics and software architecture that's in every vehicle sold in the United States with the exception of a Rivian or Tesla,

the beginnings of that network architecture date all the way back to the 1960s, and they go back to the very first piece of electronics that exists in the car, the first microcontroller.

And that first microcontroller was alongside of a fuel injection system.

So, for those car enthusiasts, remember like the old Bosch K-Tronic, and all these very early, 50, 60-year-old electronic fuel injection systems.

That was the first time the car went from a fully analog thing, having its first piece of compute inside the car.

And at the time, this is in the 1960s, car companies said, this is in our core competency, we're not going to develop this ourselves, we'll work with companies like Bosch to build these systems.

And that was the first emergence of what we now call an ECU, and that was an ECU, an electronic control unit that was focused on a function or domain.

So, we now would call that, like, that's a domain controller, that ECU is a domain controller.

And then, not in a planned way, not in a way that was architected, but nonetheless it happened over the next 60 years.

All these different areas of the vehicle started to add intelligence to themselves.

So, you would add a microcontroller with your HVAC system, you'd add a microcontroller with your seats.

And everything that had any form of intelligence or control logic started to have these little ECUs.

And, again, it wasn't like overnight, it was over a half a century.

And, fast forward to today, a modern car will have 100 to 150 of these little ECUs, these little domain controllers.

And it's like an awful field of weeds.

Each of these ECUs is its own hardware stack, each has its own little island of software that island of software is made by a supplier.

The suppliers, in fact, generally don't break the software themselves, they hand it to a third-party often in India.

And so, the layers of abstraction between the actual code and the manufacturer are two to three levels across 100 to 150 different controllers.

And that's a mess. It's hard to manage, but it was sort of livable.

You could live with it in a world where the software was locked, meaning you had to do all this incredible coordination efforts.

So, if you were involved in software and manufacturer at an OEM, your job would be coordinating amongst many different suppliers,

those suppliers be building these ECUs, those ECUs communicate over a CAN network, and you'd lock it.

And then, along came this idea of what if we want to update it.

And so, let me use, like, this is a very simple sequence to illustrate the point.

If I were to lock up to a Caribbean, or to any car, in the sequence of events that occurs, the lights, you know, change orientation, there's an awful noise in the exterior car, the car obviously unlocks, the security systems disengage, the seat adjust, the HVAC comes on, something happens with the music, the screens come on.

Let's see what else, the car may lower. And if you look on average, the number of companies involved with managing that sequence of events that occur when you walk up to a car for a ton lock is around 15.

And so, if you want to change that sequence, the exercise is like a mind numbing process of coordinating amongst 15 different companies.

Whereas in a Rivian, I could change the sequence of events this morning. We could agree right now that every Rivian on the road will have a different, like, unlocks sequence, and we could have it out by the afternoon with an OTA.

Because we control the entire stock. And so, what Rivian did differently in Tesla, and then more than five less than 10 Chinese companies, is they said it rather than having this distributed set of many, many ECUs.

Let's have a centralized compute platform where we own, obviously, the whole computer, but importantly, we own the whole software stack. So we own the operating system, the real time operating system, we own the applications layers, we own the middleware layer.

And by owning that entire platform, you can create new features very easily.

And the coordination calls come way down. And it was actually that software platform, I just described that we licensed to Volkswagen for $5.8 million.

And the reason for that licensing deal is what I just described is they need a solution quickly to have a technology that's competitive against the Chinese alternatives.

And so there's less than 10 companies, car companies and world that have this, there's two that have it outside of China. And for us, it was sort of an obvious decision.

If you're designing a system from scratch, you wouldn't say, I'm going to go coordinate with 100, 150 different suppliers to make these small little islands of hardware and smaller islands of software.

You have to manage it through these horrifically tall stacks of requirements documents that have to get disseminated across all these different suppliers.

You'd say, I'm going to build computers, I'm going to build a software operating system, and I'm on gold and applications that are there to be on.

And but to do that, you need thousands of software engineers, you need computer design engineers.

So it's just a very different skill set apology than what car companies have historically had.

Right, right. Okay. Well, so that's good segue to the Volkswagen partnership.

Is it maybe partnerships not the right word? I'd like to hear from you. Is it a pure licensing deal? You give them the software and they tweak it for their uses?

Or are you co developing anything with them? Obviously, they're giving you a lot of money, which is also helpful.

So they're also now are our second largest shareholder. So every quarter, I get like the shareholder listing.

And number one, our high, our largest shareholder is Amazon.

But number two is Porsche SE, which is, I grew up a Porsche fan and I grew up restoring Porsche.

So seeing that always makes me smile. It's sort of on a 20 billion dollars, right?

So, yes, so what we've structured is we have a joint venture that deploys the technology.

And something like this, it's not, it's, the reason I think for the scale of the steel is it's not an easy thing to integrate.

And then what do I mean by that? You can't have to do it. You can't say I'm going to do a half of a zonal architecture in half of a domain based architecture.

You have to really jump into it. And in the first go around, you may not be able to get it perfect.

But over time, you can scrape away a lot of those extraneous ECCs. What do I mean when I say you can't get it perfect?

Let's say you're a sourcing person and a traditional car company. And for the last few decades, you've sourced seats with a set of requirements that include functions.

And those functions are most of software. So the sheet will have six presets. It will move forward at this speed.

It'll go down at this speed. It'll come up at this speed. It'll, you know, have a torque release.

If it let's say squishing a leg in the back of this level, those are all software requirements.

And the seat then comes with an ECU. And so of course, we don't want to seek with an ECU. We want to dump seat because we could do one ourselves to do all the thinking.

So it takes a lot of work with the suppliers to reverse out of a lot of this quote-unquote value out of content that historically they would put into the car that is simply no longer needed.

And so if you do the full boundary diagram of what's the amount of cost save, I mean we're talking many thousands of dollars of cost save by improving this heap of unnecessary use.

And instead relying on those own controllers to do most of the decision and control logic.

And so to do that, it's more than just here's the zip drive with the software. It's like we have to integrate it.

We have to also work with the teams to understand that our sourcing approach is now different reliance on some of these big giant tier ones that historically, you know, they've had huge levels of control over a vehicle program.

You know the power shifted where no longer they have any, you know, they have very little control in a program like this.

And so it's a reset of almost entirely how you work on the autonomous car.

When we come back, we'll hear more from this conversation with Rivian CEO RJ Scorinch.

Inventory is tight, competition is fierce, but Kelly Bluebook Instant Cash Offer gives you the edge.

With direct access to consumer trade-ins, you can lock in more of the used inventory you need and load up your lot.

Dealers using Instant Cash Offer cut cost a market by 3.7% and close deals faster with new tools like counteroffer.

That means more margin less hassle. Whether you're looking to boost your used car pipeline or stream on your acquisition strategy, it's time to lock in and load up.

Visit b2b.kbb.com to request your demo and see how Kelly Bluebook can help you win more deals more often.

As a follow up to that, will you license your technology to other car makers as well?

We're certainly open to it. We've had a lot of interest. Right now, we're very focused on executing the large number of programs that we have with BW.

I say BW Group and we think of Volkswagen, but I think everyone here knows that that represents a very broad spectrum of brands.

That's Corsia, that's Bentley, that's Scoda, that's Cooper, that's Audi.

There's a very large spectrum of brands and a lot of products that come in.

We have 31 programs that need to launch through quarters of the decade and it's a ton of programs.

The way we've looked at it is to say, let's successfully launch a number of these programs before we engage in these other discussions.

One of the first programs, which I think is going to be just a remarkable car, something called the ID1, the VW ID1. You can google it as pictures of it.

It's a $22,000 electric car. It's very much utilizing our network architecture in ECutipology to allow that cost structure to be hit.

It's not going to be sold here in the next stage. I wish it was because I loved the own one.

But it represents just how much cost efficiency can be achieved through this type of an architecture.

And not get announced, but on the other end of the spectrum is a flagship product under one of the VW Group's brands.

That will also launch. And so the beauty of that is it's in some ways like the ultimate demonstration of the platform to show it from a $22,000 car all the way through a flagship.

And it's the proof that ribbing is capable of working with a large complex car company to deploy the technology.

So our view is launch those products well, and then there'll be lots of other options.

The thing I would note and open up for Q&A, I would say, I would say emphatically that it's inconceivable to me that a car company can exist and maintain market share.

By 2035 and not have a software to find architecture.

I think today you can get by you can get by with it with this mess of ECUs and really inefficient software architecture.

And by the way, it's an environment that is nearly impossible to hire capable software engineers into because you get hired and you're immediately just doing supplier management.

So I think every single car company, if they want to maintain market share, will have to make this transition over the next 10 years.

I think every car company has three choices. I think they can do it themselves. It's very, very hard to do.

It's very different than the skill sets that have been accumulated in gold over the last 100 years.

They can license it from a third party like us, or they can try to do it with existing suppliers.

But the very obvious challenge with existing suppliers is they are deeply conflicted.

So when you're a supplier that makes a lot of money selling ECUs, the very nature of his own architecture is to eliminate all those ECUs.

And so if you were to just look at it across the whole auto industry, every car should have around $3,000 to $4,000 less cost.

And so that cost is coming from somewhere. The OEMs, it's not their cost, it's the suppliers cost.

$3,000 to $4,000 per car that I think by 2035, certainly by 2040.

No at-scale manufacturer will be doing that anymore. They won't be able to burn in the car with that much cost.

Putting aside just the inefficiencies and the effectiveness of trying to do things like go with your updates.

And so I call that out because I do think there will be companies that come and say, hey, we'd like to work with you in this.

Yeah, that's very interesting. Some of them are likely to fall by the wayside if those revenues...

Anything call them will recognize. I think they'll try it. I think many will try. I think many will fail. I think some will be successful, but we'll see.

Rivian CEO, R.J. Scurringe, spoke this past week during a fireside chat with the Automotive Press Association.

At the EV Makers campus in Plymouth, Michigan, he spoke with exeos reporter, Joanne Muller.

Thanks for listening to this bonus episode of Daily Drive. We'll be back on Monday with a brand new full episode of the show.

Oct. 12, 2025 | Bonus Episode: Rivian CEO RJ Scaringe says software critical to automaker survival

About this episode

software-defined vehicle (SDV)

microcontroller

fuel injection system

Bosch K-Tronic

ECU

CAN network

Rivian

Tesla

OTA

Volkswagen

Porsche

domain-based architecture

zonal architecture

ECCs

autonomous car

Kelly Bluebook

Instant Cash Offer

used inventory

Audi

Volkswagen ID.1

electric car

network architecture

cost efficiency

flagship product

Request an Explanation

Want to learn more?

Help improve this episode

More from Automotive News Daily Drive