On with Kara Swisher - Inside Rivian’s Strategy to Survive a Hostile EV Market

Episode Date: January 29, 2026

Kara talks with Rivian founder and CEO RJ Scaringe at a tense time for electric vehicles: domestic sales have cooled, federal EV tax credits are gone, and tariffs are raising costs across supply chains. Rivian’s premium R1S SUV and R1T pickup helped establish the brand — the R1S is now the best-selling premium electric SUV in the country — but the company is still losing money. The R2, Rivian’s midsize SUV will hit showrooms this year, and Scaringe says it will be “an inflection point for us as a business.” Kara presses him on how to win buyers cross-shopping hybrids and gas vehicles, and what it takes to compete with both Tesla and low-cost, highly capable Chinese EV makers. They also dig into Rivian’s joint venture with Volkswagen Group, the economics of scaling an EV startup, and why Scaringe believes autonomy will eventually become as critical as having “tires on a vehicle.” Plus: Aurora CEO Chris Urmson asks why Rivian doesn't have CarPlay. Questions? Comments? Email us at [e-mail] or find us on YouTube, Instagram, TikTok, Threads, and Bluesky @onwithkaraswisher. Learn more about your ad choices. Visit podcastchoices.com/adchoices

It's on. Hi, everyone, from New York Magazine and the Vox Media Podcast Network. This is On with Caroswisher, and I'm Caroswisher. My guest today is R.J. Scouringe. He's the founder and CEO of the electric vehicle company, Rivian. They're making two cars right now, the R1S, a seven-seat SUV, and the R1T, a five-seat pickup truck. Rivian, though, is about to roll out a smaller SUV in the coming months, and the company has also announced a bunch of plans for more self-driving and autonomous vehicle.

heck. But it's not a great time to be an all-EV company right now. Sales of electric cars plunged at the end of last year. That was especially true for high-end EVs like the ones Rivian makes. There's lots of reasons why, but the big one was that President Trump and congressional Republicans ended the federal tax credit to help people buy EVs. Basically, a new EV got $7,500 more expensive overnight. Tariffs certainly haven't helped the situation, but it's a big deal for a company like Rivian, which still isn't turning a profit yet. I think I think Rivian's a fascinating company. I think it's a beautiful car. I've tested it many times. I've never met RJ, which is odd. But he's a very thoughtful CEO in a way that Elon Musk used to be and has some interesting thoughts and where things are going. He's certainly an important player in the area. And as most people know, I've been obsessed with autonomous vehicles and electric cars for 20 years and have driven in most of them many, many years ago and have been a real proponent of them, though I absolutely see the difficulty he and other. car companies face. I also want a world where there's more choice, where Tesla isn't the only

choice, and of course, consumers are already answering that question by not buying Tesla's sales are down quite a bit. So I always root for new competitors in this country and across the world, actually, and I certainly don't want it to be a situation where the Chinese makers get to dominate everything. All right, let's get into my conversation with R.J. Skarin. Our expert question comes from someone I have great regard for, Chris Earn. The co-founder, chairman and CEO of Aurora. First person I ever worked really closely with on understanding where electric cars and autonomous vehicles were going. So don't go anywhere.

RJ, thanks for coming on on. Well, thanks for having me. It's great to be here. So let's get started. Right now, Rivian makes an SUV and a pickup truck, and on the commercial side, you have a work van. Some baseline models cost a little more than $70,000, most are north of 80K. Talk about the price point on the consumer side. You're obviously in the luxury market.

So talk about who your customer is, who you're making vehicles for, and which models are driving the most revenue. Yeah, so we launched our first products at the end of 2021. It was our R1, what we call our R1T and our R1S. It's a sibling pair of vehicles. And these are flagship products. So they're intended to be, you know, premium products. As you said, the price points are high. So the average selling price on R1 is around $90,000.

And, you know, that was our handshake with the world. It was how we introduced Drivian as a brand to consumers. And we're now about to launch for like months away from launching what we call R2, which is a much lower priced vehicle, starts at $45,000. And it's really our first true mass market vehicle. vehicle, whereas the R1 being a flagship product, you know, it sells extremely well. It's the, it's the best-selling premium electric SUV in the United States by a significant degree. But just the nature of the price point limits, you know, how accessible the product is.

And so R2 for us represents such an important inflection point for the business. So who is the customer? Who is it for? Because, you know, as for the truck, it's not really a work truck. How are people using it? Yeah, I mean, that gets into like the deeper question of what's the brand's purpose. And there's, I think often when you think about products like this, like a vehicle, it's easy to sort of immediately jump to what are the specs? Like what's its zero to 60 or what's its cargo capacity? But there's a little bit more, I think there's a deeper meaning to what's the sum of the parts. What is it, what's the feeling it's trying to evoke? And so, for us as a brand, we've long talked about the desire to not only enable people to go do the kinds of things you want to have memories of for years and years or the kinds of things you'd want to take a photograph of, but also to inspire those things. And so to inspire this adventurous lifestyle or to inspire a life of saying yes to new experiences. And so R2 is just an extension of that.

It takes that same concept that we've delivered on in R1 of lots of little detailed design elements that all lead to this inspiration around adventure to something that's a smaller form factor, but also importantly, a much lower price point. So why, just explain for people who don't understand why start there at those price points. Obviously, you often start with wealthier people and then it trickles down in the car business, but sometimes it absolutely doesn't. In the very beginning, you know, so four years ago when we were launching R1, we didn't have any production volume, no supply chain scale or experience, not a lot of.

the leverage that comes with generating scale and generating momentum with suppliers for pricing. And so just by virtue of those constraints, it's helpful to have the first product which is going to be lower volume, be a higher price product. And then as we build scale and capabilities, the goal, of course, is to introduce lower price products. And then specifically with how we arrived at the price content performance tradeoff of R2, like we could have made it lower price, but not as high performance or lower price, but with a lot less features, is we were trying to identify,

where is that sweet spot in terms of volume, the buyers that are going to be looking for this type of a product, and found that to be in this, like, as I said, like this starting at $45,000, which allows it to be cross-shopped with a lot of internal combustion vehicles. Right, exactly. The analogy, of course, is an iPhone, which was quite expensive when it first started,

and then it's still expensive, but it's widely. used. If people are looking for an affordable car, because people are very price sensitive right now, how do you convince them? I have the most boring cars. I have a Chevy Bolt and a Kia hybrid. And I'm actually in the market for a car, and I'm looking at the Rivian, too. I absolutely am. But what would be the selling point for someone like me? Yeah, I mean, well, first, your question is just on pricing. You know, at the higher end, like where R1 sits, so the R1S gets, you know, it's cross-shopped with land rovers that are, or rangeovers

are much more expensive, you know, everything from Range Rovers, Esclades, Porsches, BMWs, and then into Tesla Model 3, Model Y. So you have like a pretty broad range of customers coming out of a broad range of price points. But I'd say pricing parity is there with R1's competitive set.

And actually, the evidence to that is take California or state of Washington. The R1S isn't just the best-selling premium electric SUV. It's the best-selling premium SUV. in the state of California. So for SGVs priced over $70,000, R1S outsells everything. So then the question is, okay, if we move into the category of $45,000 and up, which then you're suddenly in the world of more mass market cars,

the kinds of things you'd expect to see broad portions of population buying, and then it's like there's a wonderfully broad aperture through which you're attracting and potentially appealing to customers. as they're making the decisions as to what they buy. And I think one of the things that's, maybe one of the most important things that's often missed around electrification is there is a surprising lack of choice.

And so at the, at the, call it, $45,000 to $55,000 price point, there's a handful of choices that are compelling. And you contrast that with truly hundreds of internal combustion choices that are compelling. And, and a wide. array of foreign factors and design decisions that have been made by each of the companies. And so, you know, for us to see adoption grow beyond what we have today in the United States

around 8 percent, I really believe we need a lot more choice. And you see so much concentration of market shares today with one company with Tesla. And that's not a reflection of a healthy market. It's a reflection of a market that's being underserved. It's a reflection of just not enough choices. You were the first automakers, as you noted, to bring a fully electric truck to the market. You'd be out Ford, GM, and Tesla and begin delivering the R1T in the fall of 2021, as you said. This has the last couple of years has been tough on all carmakers and especially EV makers. You've had to deal with pandemic, supply chain issues, chip shortages, inflation, now tariffs while building your brand. And also consumer disinclination,

which is still, I believe that will change. I think that's a shifting kind of thing, but it's still there. A lot of these carmakers have pulled back, obviously. Talk about the time now and about delivering the R2 in the next few months. A lot of people, including the publication, InsideEVs, describe the RT as a make-or-break product for you. So talk a little bit about that. Because as most people know, Ravion isn't profitable yet. Huge losses. I mean, nothing compared to AI, so don't worry about it.

But go ahead. Yeah, I mean, R2 is, you know, you could call many things. It's a make or break product. It's the product that's really given me an inflection point for us as a business. You know, and there's nothing, that's not at all surprising to us. In fact, that was the plan. You know, certainly many things have gone differently we expected, but the plan was always to establish the brand with a flagship product

that by virtue of its price would be more limiting in its scale. But invest in a whole host of vertically integrated capabilities from our technology, so that's electronics, software. autonomy, AI, all the way through to our distribution channels. We're building sales channels, service channels. We're building a charging network. So essentially be investing in a very heavy level. So with this level of vertical integration.

So lots of fixed costs, both on the building of these teams to do that, but also just in terms of running the enterprise, that in the beginning, when it's just R1, isn't enough to offset the cost to run all those big systems. but Rivian isn't being designed as just for R1. So like if the goal was to build a company that's only going to make 50,000 cars a year, it would look very different as a business, but we've really invested in the business,

contemplating and planning for Rivian to be a much larger company. And so because of that, R2 does need to be successful to support the rate of capital consumption of the business because of how much tech we're developing and how much infrastructure we're building. And so because of that, we've also really wanted to make sure R2 is dialed. And so there's been a lot of iteration in the product, a lot of learning from the R1. I just drove one of the early manufacturing validation builds off the line last week.

And I mean, the car is just incredible. So we're quite bullish on it, but we also recognize, you know, it needs to get into the market. And we're very hopeful that it will resonate in the same way. that the R1 products have resonated, meaning... Although it cannot just merely match sales of your R1s, correct? No, but if... Yeah, correct, on a gross basis, but what I'm saying is on a market share basis.

So if I were to take the R1 relative to, let's say, a Model X, we significantly outsell the Model X. They're priced about the same, very different in terms of how they've been interpreted from a design, aesthetic, design decision, or engineering decision point of view. but covering a very similar ground, if we could even have a fraction of the market share

we have on R1. So taking it from them, presumably. Taking it or I think more likely, some of that will inherently happen, but more likely just creating an option for folks that haven't been able to jump from an ice into an EV yet because the choice wasn't there for them.

If you look at it through the lens of a customer, if I want a mid-size SUV today, and I want one that's like very robust technology, but I don't like the way a Tesla model I looks or I'm not attracted to the brand. I'm really without a choice. Johnny Ive once told me that Tesla looks like the inside of an egg and I tend to agree with him. It was so eviscerating in such an elegant way. But go ahead.

That's funny. So this has to sell, right? The company's future rests on this car, correct? Yeah. Yeah. If it doesn't sell, we've got to rethink some things about the business. Yeah, absolutely. So just for, in addition to the R2, Rivian Set to release a smaller crossover, the R3, you have a joint venture with Volkswagen to develop software and electronics, and the investment is about $6 billion. VW will use the technology across its brands, including Portia. A bit of a struggling company, too, but they're trying to jigger themselves. But talk about being the software and electronics company that works with legacy automakers, not a bad business. Yes.

Yeah. Volkswagen Group, this is the second largest car company in the world by volume, did a $5.8 billion deal with us, a software licensing deal, as you said, for us to take our electronics and our software and deploy them across their portfolio brands and products. And that's really exciting to us because it's so deeply aligned to the mission, but it answers this question we talked about a moment ago, which is the last, lack of choice and the need for more choices and highly compelling choices at that. But, you know, one of the things about automotive that is, I think often maybe overlooked,

is just the nature of how car companies and the skill sets embedded within car companies evolved over the last, call it 67 years. And of course, cars, if you go back, were very analog. And so prior to the 1960s, there was not a single computer in a car. There wasn't a single line of code. No, I was there because I so miss rolling up my window. Yeah. So all that, all the electronics that have entered into the vehicle actually were borne out of, very ironically, fuel injection systems. It was the first computer in the car was to control fuel injectors. And car companies at that time, this was in the 1960s, made the decision to push the electronics for the fuel injection systems to third parties to suppliers.

And not necessarily planned and, you know, you could do a whole study on this, but over the course of the subsequent 60 years, anything that started to enter the vehicle that was electrically controlled or controlled through software had a little computer come with it, what we now call ECUs. And, you know, over the fullness of time, cars went from having no computers to today having 100 to maybe 150 little ECUs that have their own little island of software that's written by a supplier to control the function or the domain that it's associated with. And it's almost the exact opposite of how you design a software architecture and a network architecture if you're going to do it from scratch today, where you'd say, I want to have a very small number of computers that have a very broad function set that are located geographically within the vehicle or in a zone of the vehicle to control everything in that. zone. And so if you're starting with a clean sheet, you'd very quickly arrive at this idea of

centralized compute, more of a zonal architecture. And so Tesla, of course, has an architecture that looks like that, and Arriviant has that. And we're the only two companies today that have this, you know, very unique. But could you see a situation where it becomes just a software and electronics only company that works with Legacy automaker? I don't, it's certainly not the plan. We want a car. Yeah, we would do both. And the business of, like, we're quite bullish on the nature of the business, meaning if you have, once you make these investments in these core technologies and provided you're building a brand that has a level of desire building and excitement around it, there's really an enormous opportunity to capture customers and market share. And in the margin structure, we think particularly with the introduction of more technology in the form of artificial intelligence and self-driving only further grows what we think. is the business. We'll be back in a minute. Support for the show comes from Ship Station. As your business grows, so do the challenges, especially when it comes to filling orders, one of the things

your success can really hinge on. With Ship Station, you don't have to tackle it alone. Ship Station gives you everything you need to manage your shipping and get orders to customers all in one place. That includes order management, rate shipping, inventory, and returns, warehouse systems, and comprehensive analytics. So instead of toggling between five to seven disconnected tools, you only need one. Shipstation says their time-saving automations can save customers 15 hours per week on fulfillment, and they also do the legwork of comparing rates across major global carriers, finding you the best shipping option on every order. If you already have a negotiated carrier rates, no problem, just bring them over to Ship Station. That means you can keep your discounts while

adding Ship Station's automation and intelligent features to help things go even smoother. You can try ShipStation free for 60 days with full access to all features no credit card needed. You can go to Shipstation.com and use the code Kara for 60 days for free. 60 days gives you plenty of time to see exactly how much time and money you're saving on every shipment. That's shipstation.com code cara. Shipstation.com code cara. Right now in the world of AI, two things are happening simultaneously. One, the technology is getting better fast.

people are finding new uses for it. It's more popular than ever. And two, every company that makes AI is absolutely hemorrhaging cash. On the Vergecast this week, we're talking about what OpenAI and other companies are doing to try to finally figure out how to make some money off of this technology. Spoiler alert, it's mostly ads. And we're talking about whether any of it's actually going to work. All that, plus some stories about the Chinese company that appears to be beating Tesla on the Vergecast, wherever you get podcasts. This week on Net Worth and Chill, I'm revealing the exact playbook rich people use to set their kids up for financial success and how you can do it too. Even if you're not a trust fund baby.

From adding your newborn as an authorized user on your credit card to give them a near perfect credit score by the age of 18 to opening a 529 account that saves you thousands and taxes on everything from books to college tuition. Plus, I'm answering your questions about whether your debt transfers to your kids, tax credits worth thousands that new parents are missing out on, and how to give your children money completely tax-free, whether you're expecting, already have kids, or just want to understand how generational wealth transfer actually works. This episode will show you how to break the cycle and build a legacy. Get ready to give your kids the financial head start you wish you had. Listen wherever you get your podcasts or watch on YouTube.com slash your rich BFF. So let's talk a little bit about some of the obstacles you're facing, though,

because let's talk about consumers, because that's really where the, I hate to make metaphors, but rubber meets the road. The big one right now is that EV sales in the U.S. took a note. side at the end of last year, Ravians' fourth quarter sales were down more than 30 percent compared to the same period. Now, that's not unique to Rivian, by the way. Ford stopped making F-150 lightning, which I tried out. Terrific car. GM paused production of its electric hummers, which I also tried a cool car, and Cadillac escalates, same thing. Beautiful car, actually. The escalade was stunning. Stalantis canceled plans for an all-electric ram pickup truck. Now, I get that because when I wrote

a column six years ago and said, someday you're not going to own a car. Most of the letters I got were from truck people who were, like, in the Midwest. We want our cars. I'm like, fine. I think it was like having a horse, something like that. And I said, fine, have your fucking horse. I don't care. But this is where it's going.

It's like the people who like a phone in their house. I don't think these are going there. I'm quite bullish on them, too. But what do you do to deal with the, obviously, the drop in the man? You've had to cut 600 employees. Many car makers are doing this. Tesla's fallen off a cliff, really, even though the stock keeps rowing. that's a whole different meme situation.

But one of the reasons for the drop in sales is because Trump administration and congressional Republicans ended federal tax credits for people buying EVs, which has been around for a long time. Trump has three more years in his term. So how do you navigate those years given if the federal government is working against efforts to push more people and companies towards EVs? Yeah. Well, the way we think about it is ultimately our job. And that is, I think this is true for any company that's making products, is to develop products that are highly compelling.

And in order to truly drive electrification adoption, and the products need to be not like slightly better or as good as. They need to be a wildly more compelling choice than an ICE alternative. And so that was the philosophy that, you know, drove me to start Rivian, it was before any of these policies existed. This philosophy still drives us today as we think. about R2. And so when we look at it, it's the performance that's like zero to 60. The driving dynamics need to be exceptional best in class. The storage and functionality and features that allow you to fit your gear, your pets, your stuff need to be really thoughtful. The design needs to be something that's compelling. And then range in the time to charge the vehicle need to be really nice.

And this is perhaps the one that's, I mean, most difficult to explain until you experience. it, which is once you try an EV and eliminate the need to go to a gas station, and if you have charging at your home, which most folks do in the United States, of course, there's challenges in highly urban environments, but in most of the United States. It's not that hard anymore. Yeah, it's not that hard. It isn't. And it's really nice.

Your cars, you don't have to think about, did I put fuel in it? You charge it at night. And there's like all the benefits, the intrinsic, sort of hard-to-measure benefits, if it's quiet, it's smooth. And so we do think that with the existence of more choice, that demand will naturally come up. And there's something here that, in speaking to the consumer, I think, is really maybe unexpected if you were to just listen to all the rhetoric and all the noise, which is, we look at our customers today, our R1 customers. And if you were to just extract out some of the tone you hear around these topics, you would expect 95% of our customers to be Democratic.

and a small fraction of a Republican. And the truth is actually it's about 50-50. And I think that's great news. And what it tells me is customers are going to choose something that appeals to them. And regardless of maybe political tone. There's only one political car that exists is a cyber truck. But go ahead. There's zero others, really.

So when we see that, it's encouraging. And we try really hard. a company to not politicize the idea of electrification. It's strangely become political in ways that I wouldn't have expected and somewhat. Well, Elon is the problem for you, for all of you, honestly, in that regard. Yeah. Well, nonetheless, so we just have said, let's focus on making the best cars. And so in doing that, there are externalities that have happened in the last six months. What is the greatest consumer sentiment against them? Obviously, I think most people are

intrigued. There's no question. Everybody now knows about them. And when I started pushing electric cars, many, many years ago, most people didn't or wouldn't consider it. Obviously, range anxiety is one of the issues. But from your perspective, what has shifted? Yeah, I sort of bucket it maybe into three categories. I think there are people who are open to it and looking for the right car. So they want an EV. Maybe they own one. Or maybe they don't. In many cases, they don't because they just haven't had the form factor, you know, what they need for function, coupled with price. coupled with something that they emotionally connect to. And so we believe there's a lot of folks there.

And we see this firsthand with customers. Most of our R1 customers are first-time EV buyers. And they were in that category. The product they were looking for hadn't arrived yet. The second category are people that are curious, but have those concerns, range, anxiety. You know, my folks live six or miles away. I want to be able to drive to their house at Christmas and not have to worry about charging along the way, those kinds of things. And then the third category, which is by far the smallest,

but I do think the most vocal, are folks that for some reason have identified electric vehicles as being bad or wanting to be against them. And I do think this is a very small percentage, but it can generate noise. And I've reflected a lot as to what's the psychology driving that. And I don't know. I'm certainly not a psychologist, but if I had to like project a gas, would say it's some version of, I think folks that feel like a solution's being pushed onto them or forced onto them are inclined to naturally reject that. Well, the haters include the President of the United States. Clearly, he talks about them quite negatively.

How do you navigate that next period, especially if they continue to be hostile? And on top of axing the federal tax credit, Trump's also slapped tariffs. That may change, but it'll take a while. So talk a little bit about the Trump impact here on. your business? Well, and to that last category, so folks that are, let's say, against electric vehicles, I think the first thing we do from a consumer point of view is we make it very clear that we want there to be choice.

And I think the existence of choice helps eliminate some of those tensions and to say, look, if you don't want to buy an electric car, you don't need to buy an electric car. But we're making these vehicles to be really compelling and hopefully you drive it and you experience it and it sways you or pulls you in. I think as it pertains to policy and working with the administration, what we've really focused on, which is an important point, is the indisputable view that over time, and we can debate how long that is, but in the fullness of time, high technology vehicles where there's lots of software, lots of electronics, they're deeply integrated together, the vehicles increasingly drive themselves, the skill sets necessary around

power electronics and battery systems, those are things that the U.S. should continue to maintain a leadership position in. And so in order for that to happen, it's necessary that companies like Rivian, not just Rivian, but Rivian and many other companies like Rivian need to exist to be active participants in the future technology development that we're going to see play out over the next few decades. And we found that it's a really, I think it resonates extremely well on all sides of the aisle. It's also really helpful to look at it through that lens of skills. Is Trump undermining that idea? I don't think he's...

You're being careful. I think there's a lot of... Yes. The answer is yes, but go ahead. You answer anyway. I think there's a lot of, what we try really hard to do is there's a lot of noise. There's like the signal to noise ratio right now is it's, you know, if you're looking at this is like a signal processing problem, it's the noisiest the signal has ever been. And so we're trying really hard to ignore or avoid riding up and down some of the peaks of the noise and look for the signal.

But credits did matter, so do terrorists. Oh, yeah. Yeah, credits, credits, like, created it, like, to be clear, I mean, you talked about the Q4 drop-off relative to Q3. There's no doubt, yeah, the credits going away impacted everyone. But I don't think that's going to change the end state. It's a, it's like a speed bump. It's a bump along the way, but something that we all have to navigate. Right, which could take your entire chassis off compared to other carmakers with more money. But, and this is a very, for people who don't understand, car making is so expensive compared to other. I mean, you joke. I joke about AI, but it's really is a high level of – it's a very difficult jump for everybody. So it's actually quite astonishing how we make cars in a lot of ways. Now Canada has agreed to slash its 100% tariffs on Chinese EVs, which is not a good news for you all. You argue that customers need more choice in the EV market, and that's not a zero-sum game. But, wow, it's a real problem for you from a global market point of view, because you can't compete with these Chinese.

companies because of the price and innovation really. Yeah, I mean, I think first, just to frame the overall Chinese market, it's very, very, very different than what we have in the United States. There's a couple of ways that's very different. The first is there are many choices there, like an overwhelming number of choices. So there's an excess of 100 different companies making electric vehicles in China. And actually, a vast majority of them are not particularly compelling in terms of technology or innovation.

but a subset of those are both affordable and impressive technology. Companies like Xiaomi or B-Y-D very much technically demonstrating a lot of leadership and then from a price point of view quite competitive. And there are two things to take away from that. The first which I think is often overlooked is for a lot of the same reasons, Rivian or Tesla have vertically integrated technology stacks. distinctly different than legacy car companies is we had the benefit of starting with a clean sheet in a world where software and electronics existed. And so we, like, Rivian was built as a software

and electronics company first, and we make cars. And so we have more software engineers than we do mechanical engineers. And that same dynamic exists for a number of the newer Chinese car companies where they approached it as a technology first business. Sure did. They made architectural decisions that are advantaged relative to legacy manufacturers. And if you were to take those cars apart and compare them to a Rivian, you'd see there's technically, there's actually a lot of similar in-zonal architectures, really heavy focus on in-vehicle inference and compute for enabling AI features at scale over time, you know, like removal of a lot of the Tier 1 supply base and bringing a lot of that in-house,

a lot of consistency. And then you'd say, well, why are the cost structure are so different. And here, I do think too often we look at this as if it's like some, some like black art, it's like magic happening. But it's not magic. It's, you can build it in a spreadsheet. And it's, it's a combination of things. The first is the cost of capital in China. So the cost to build the plant and invest in the products is very low, near zero, because in many cases, the plant and the equipment is paid for by the government. So whereas we have to raise capital with equity or some debt. In China, you get sort of given the capital to go build capacity.

Yes, they have decided this is an area they want to dominate. And so it's like, it's very much a decision by the Chinese government to say we're going to be leaders in a space. And so they've heavily subsidized the capital initiative to build these businesses. That number one. The second is that's not just happening at the OEMs. That's happening across the tiered supply. And so you take the cost of capital to close to zero, the whole supply base suddenly becomes cheaper because your amortization of your fixed cost becomes a lot less. Then you have a labor cost that's much, much lower and labor policies that are very different. So you have people that go and work and live at a plant working 50 to 60 hours a week and making, you know, somewhere between one-fifth and one-seventh of what someone in the United States would make. And there's no way with that difference in labor cost to really have, and the cost of capital, to have a like for like one to one.

I get all that. They're also very innovative. I think one of the canards. Well, there's my point first. That's why I wanted to call out. Like, I think sometimes we just say, oh, it's low cost. There's a handful of companies that are very, very strong in terms of technology. And they have these advantaged cost structures. So here you have Trump slashing, slashing the credits, putting on the tariffs, which makes it very difficult for all kinds of companies, and Canada agreed to slash this. It might have been favorable to you had we had a pleasant relationship. Now it's toxic and largely due to the Trump administration. In fact, completely due to Trump. Why am I being nice about it? So how do you compete if this is the case, especially this Canadian thing, is a big deal? Well, yeah. So we have 100% tariff for people don't know on Chinese EVs.

Yeah, yeah, that's an important point to note. Yeah. I think first, as I called out, there's two vectors here to consider. So first is we have to compete on technology. And you made that point a couple of times. I couldn't more fully agree with you. And it's for that reason, you know, Rivian has many of the cars we just described. We own them.

We study them. We take them apart, just like the Chinese have our cars and take them apart and study them. And so we need to continue to be on the tip of the spear in terms of technology and leading there. Number one, on cost, there's not a way to be competitive where labor costs is so much lower or where cost of capital is so much lower. And so I think what will happen in the United States, and I can't speak for other countries, I think it's very, very, very likely that the tariffs, the United States have put in place will remain.

And for the Chinese to really participate in this market, they'll need to locally produce. And if they decide to locally produce, those advantages I just talked about, both in terms of cost to capital, but also in terms of labor costs, just go away. And there's not, they're not building the cars differently. You're to take a Chinese car apart and compare it to a rig. Like they're stamping and castings. It's all very similar processes. Again, there's no like, there's no like magic happening.

It's just, you know, there's two big factors. And so I think in, you know, the United States, which is by far the second largest car mark in the world, the first being China, is maybe the one topic that the right and the left are very aligned on, which is a pro-US policy. Now, I think Europe's a big open question, as you've already called out, the interesting thing there is they have a large domestic auto industry to protect as well. But a lot of the countries that don't have domestic auto industry are likely to be more open or fully open to Chinese vehicles. I do think it's going to see a lot of market share shift to the. those brands. Now, what's also happening, which is we haven't touched on yet, but is an important point, is just the nature of how cars are produced is going to change over the next 10 years.

And this is an area where we're going to start to see industrial automation and high dexterity robotics come in that this is truly the only way the United States can compete with low labor cost countries for manufacturing is by developing robotics to essentially. operate manufacturing plants with far fewer people. And so I do think that's something that is going to happen. Which of course takes time. And to that point, I started a robotics company recently called Mind Robotics that specifically focused on industrial automation industrial robotics. Separately from your company. It reverevan's a shareholder in it, but it's a separate entity. Yeah, I've been talking a lot about everyone's focused on AI. I'm like, you need to focus on robotics.

is what you actually need to focus. I couldn't agree more. I keep saying it. And everyone's like, I was like, stop with the AI will be a critical part of robotics, but it's robotics in terms of lots of things will matter. Great deal. We'll be back in a minute. Every episode we get a question from an outside expert, by the way.

And yours comes from Chris Irmson, the co-founder, chairman, and CEO of Aurora. I rode in one of my first cars of his, and I tried to run him over, as he knows, with the electric car in Google parking lot. This was almost 20 years ago. So let's listen to him. I'm a big fan of Chris. Yeah, me too. Hey, this is Chris Irmson.

My question for Raja is this. I've been a Rivian truck owner for many years now. I love it. It looks great. It has awesome performance. The one area that could really be improved is the media experience. Can you talk a little bit about the strategy behind why you haven't leaned into the carplay ecosystem and why you think that leads to a better experience for customers,

particularly give them for the new car buyers somewhere two and a third and a half of them aren't interest in buying a car if it doesn't have car play. Anyway, cheering for me from the sideline, huge fan. Thanks for all you doing. He gave you an easy one, but I would agree. I have to say it matters. I'm just as superficial. Oh, it's a great question.

Chris and I've talked about that a lot in the past. We have the view that the integration of vehicle capabilities, So vehicle status, vehicle health, vehicle, so like window position, vehicle speed, anything that has to do with how the vehicle is operating into, as context, into the media or user interface experience is going to be really important. And we want to control that whole, essentially the whole platter. And we think that becomes more important as AI enters more and more into the existence of how you run the vehicle. And so the challenge with CarPlay is, of course, it's an app that runs in, in, you know, on the screen, but it sort of takes you out of the vehicle experience into a separate app that's very media-centric. But it's very difficult to create consistency across all the different platforms and to create consistency with every part of the user interface. And so what's missing, you know, what draws like that question is like I want to have the exact same ease of using my text or ease.

of using functions around my phone. And so what we've done is we've worked really hard with Apple and similarly with Google to create the same menu of options, but we're Rivian is the, we're setting the table. And so, for example, we have an Apple Music integration. We just did an Apple Watch integration as a key. We have Google Maps integration. You know, two years ago, three years ago, this was a bigger issue because we didn't have all those integrations and some of those discussions we have with our partnership at Apple and with

Google took some time. But we're now in a place where we have a great working relationship with Apple, a great working relationship with Google, where we can serve up all those items. And now we're doing the backend work, which we just demonstrated, to really enable, and more of an agent-to-agent basis, all the integration of the back-ends of these apps to create a more seamless experience. So you could say to the car, I want to go to dinner tomorrow. I want it to be after my meeting, that meeting with Bill, and I need to be home in time for this. And it can collect all that information between your Google calendar, the state of charge of the vehicle, where the vehicle is. But that integration and aggregation of all the information is really hard to do if you leave the vehicle and go into a separate app that doesn't have access to the same types of information.

So that's the reason. It certainly has blocked, as Chris said. Some customers have said, I don't want to buy a Rivian because I want car play, which is like we recognize. Like, we're not blind to that. We recognize that. Right. So that gets into the final thing I want to talk about, about AI, because AI is part of that. And autonomous vehicles, which is another area I'm very, I'm very bullish on. Rivian recently announced some new AI features coming to its call.

As you've developed a new custom chip. Their plans to roll itself driving capabilities, obviously, you've announced to Ambition center, the Robotaxy space. Talk a little bit about your ambitions. Tesla's been promising customers will be able to turn their cars into robo taxis for years, but it's turned out to be Google and Waymo, and Chris Irmson was the original person there. Talk a little bit about what you'll be able to deliver when Tesla has just not been able to do so, and certainly should have been able to do so, although he has some obsession with the lack of cameras. So I think there's an important thing to note here, which is the nature of how we develop self-driving has changed really

dramatically in the last couple of years. And if you go back to, you talked about Chris, if you go back to the beginning when Chris first started working in this area, the way self-driving systems were designed is you had a perception platform, which depending on the design of the overall system, could be cameras, but more likely it was cameras plus radar plus LIDAR that would perceive the world, identify objects in the world, classify those objects, associate factors with those objects, and then hand all that classified object information to a planner. And the planner was a rules-based environment that was written by humans.

Humans were trying to codify or describe how a vehicle operates, what are the rules of the road. And those were the systems that were designed and built up through the late, you know, the late, you know, to call it 2020, 2021. And even Rivian's launch vehicle, we first launched within late 2021 was one of those systems. So I would call that like an AV1.0 solution. These were not neural net-based solutions. These were not systems that used, you know, things like transformer-based encoding. But around 2021, 2022, the state of the year. are in part informed by some of the things that had been happening in the LLM space, the large language

model space, started to shift to say, can we take more of a neural net based approach and build a foundation model that describes how the world works through a much more complex, large parameter model? And to do that, you need to control the perception platform, so you can't go through third parties that process this signal. You need raw signals from your cameras, your radars, and if you have them, your LIDARs, you need to be able to trigger interesting events that you can then train off of, then move those interesting events offline to build this very, very large parameter model that describes how the world works. And so we made the decision at the end of 2021, beginning of 2022, to start working on that system, on this neural net-based approach. And to do that, you need to first have the hardware, and then you need to launch it.

And then you need to put enough vehicles in the car park to have a data. You need data. You need lots of data. So here's where it gets interesting. So we launched that hardware platform in the middle of 2024. That was what we call our Gen 2 vehicle. And the biggest change was our autonomy stack. So in-house compute platform, in-house perception. And then over the course last year, started to build enough vehicles in the car park to build this really robust data flywheel. But we approached it in a way that is probably not fully understood. So I'm so glad I get to talk about it here, which is the sensors that are on the vehicles are not just there to drive the vehicle. They're very importantly there as part of a large-scale data acquisition

platform for building a model. And if I were to characterize how I think about autonomy over the next call it three years, relative to the last three or four years, the rate of progress that we will see as an industry is going to be substantially higher. And that's because of this inflection point towards neural net-based approaches. And so I'm, like, beyond bullish on this. So why do you think you'll be able to deliver on this vision? Tesla hasn't been able to. They're trying, he's been trying to do that. And I mean, I know he's busy making non-consensual sexual material. But why has that not been able? Because they do have all those elements, right? They certainly do. Yeah. So if we look at the spread of how you build this, you need large data set, lots of vehicles deployed. You have two engines.

of the spectrum, which are interesting know. One is Waymo. It's a relatively small fleet, but with a very robust sensor set. Right, absolutely. Lots of megapixels, lots of data. And then the other end of spectrum, you have Tesla with a very constrained sensor set. So, you know, the cameras are calling, they're okay, but they're dynamic range. You know, low light performance, bright light performance isn't great. So it's like very limited sensor set. We took the approach of saying we're going to have a much bigger fleet than Waymo, a smaller fleet than Tesla. A smaller fleet than Tesla, but we need our vehicles to learn faster. And so every one of our R2 vehicles, when we launch the LIDAR variant, will have a LIDAR.

And so the LIDAR is an important thing because it gives us a, it protects a lot of corner cases, particularly very long-range corner cases, but it also is a great platform for turning the entire fleet into ground truth. And so that's our goals, radar, LIDAR plus camera. I always note to people when one car learns they all do as opposed to a human being. Yes. So last couple of questions, you were just on the podcast access and you said, you think within the next three to five years, autonomy is going to be quote as critical as having tires on a vehicle and that not buying a VE capable car would be like buying a cell phone without Wi-Fi.

Now, as I noted back in 2019, I wrote a column about how owning a car would be like owning a horse down the line. Why even own a car when you have an insurer, maintain, pay tens of thousands of dollars for not to mention deal with, with legal fallout if it gets in an accident. And also, humans suck at driving at the very base level. Talk about that. You have visions of robotaxies. Now, again, the only people who have delivered consistently has been Waymo. And Zooks to a lesser extent. There's stuff happening, obviously, in China that's very quite robust, actually. Talk about why you think that's a critical part. You don't have to comment on not owning a car. I just don't think people are going to own cars, but that's just me. Yeah. Well, to the last point, I think that the business model, the customer model on how you, let's call it, purchase mobility, purchase transportation, I think could change a lot. But putting that aside, because recognizing regardless whether you own the car, you're purchasing a ride, you're purchasing a subscription, all the different ways you could conceivably consume transportation, the level four, the ability for the vehicle to operate itself, and importantly operates itself empty. is the enabler. And so our view is that that's going to go from a very, very small subset of vehicles today, as you said, the only deployed at scale variant of that is Waymo, to becoming something that if you're buying a vehicle, having what we call personal level four is going to become

very important. And of course, if you want a different mobility model, it's a necessity to enable that mobility model. And so I think this is going to represent the biggest, shaking of the nature of the auto industry. Mobility. Of just the whole thing. Like, market shares can shift dramatically with this. And so we've decided, and this has been a long effort for us, that's the most important investment area for the business.

So it's by far where we invest our most capital. As we think of like our R&D budget, it's shifted from a lot of the traditional vehicle level innovations, which are now fairly stable, electronics platform, so on, and moving those dollars to autonomy. To autonomy. And what will be the critical thing for your company to do so? We need a robust data flywheel. We need a lot of GPUs to train the model.

And we took the decision. This is the decision we took, geez, almost four years ago, to bring inference, to bring silicon in-house. And the reason for that, there's so much focus on the sensors. But the compute is actually the really expensive part. Correct. So we said, look, we need to bring compute in house versus a dependency on Nvidia. to allow us to put very high levels of compute into every car. And so we want every car to have a really capable brain.

But LIDARs are no longer $1,000 things. No, they're not. It's low hundreds of dollars. Radars are low tens of dollars. Cameras are low tens of dollars. So the perception platform, as much as that gets visual, you see it, as much it gets a lot of attention, it's the little silicon wafer that's the most expensive part of the car,

or the most expensive part of the stack. And so we brought that in-house so that we could deploy everywhere as part of this overall philosophy on building a data fly-will. In that environment, what you're talking about, which is, I agree with you,

this is where cars are going. How do you, that, you still have to run a car company. You still have to sell cars. You still need money. Does that require you get purchased by a larger, like an apple or something else?

Or can you do this on your own? Because you also have to make money, as I said. You can't just have this. and to get, you know, this valley you need to go through is quite long and difficult. And then the second one is what then happens to a competitor like Tesla? I have not asked you one Elon question because I'm so tired, sick and tired of him. But he's still critical.

And you can see him talking about it even though it's in this sort of strange way. When he's talking about robotics, I understand what he's saying. Why wouldn't a Tesla be able to do this better eventually if he ever focused on it versus you? What will it take to do that? You don't have the benefit of a meme stock that doesn't represent your business. Yeah. So, and when we said it before, Artu is really important for us. It demonstrates that like a path to the automotive side of our business being profitable and the revenue associated with that vehicle covering our fixed costs. That's really important. But I think maybe at the maybe in a more philosophical way the deeper part of the question is how many companies

ultimately can be successful so you know if we look historically automotive is a very big it's a huge market you know the tam is enormous but it's not one in which there's a single dominant player you know very different than tech so if you look at tech in each segment there's one two maybe three companies that occupy the vast majority of market share. In automotive, on a global basis, the largest manufacturer represents around 10%. And so the question is, does that continue? Or does it consolidate down to like one or two companies? And my view has continues to evolve on this, maybe without saying exactly how many companies I think there will be in the end. I think it's helpful to think. How many? Well, I think it's helpful to think about what are the ingredients necessary to be

a car company at scale. And I do think there'll be more than one. But I think there's probably going to be less than there are today. But the ingredients that I think like a must have is the vehicle to be a successful at-scale vehicle manufacturer, putting aside like very niche, call it like, you know, almost like a horse-like vehicle, vehicles that are just like anachronisms as the past. But to be at scale, the vehicle must be software to find, which allows for robust new features to be deployed, which to do that well, it means breaking from the network architecture and

ECU topology of 100 to 150 different companies writing little islands of software that then gets integrated on a canvas architecture. I think that's, there's absolutely no way a company can be at scale in five to 10 years and still have that architecture. Now, that statement sounds subtle, but that's a big statement because there's only two companies in the Western world that have a software-defined architecture, Tesla and Rivian. Of course, the deal we did with Volkswagen unlocks this for them. But that means every car company, if you believe what I just said, will have to find a path to moving to a very different software and electronics architecture than what they've had historically. Number one. The second thing, which requires the first thing

to really do the second thing well, is the vehicles will need very high levels of both AI integration and autonomy. And I think that if you miss autonomy as a car company, when if you're a customer and you have a choice of I can buy a car for $35,000, $40,000 and it can drop me at the airport, it can go to the grocery store to pick up stuff for me. It can drop a friend at a house. It can do all those things or a car that doesn't do that. It's going to be very binary. I think there will be very few people that will self-select, say I don't want those features. Even folks who are not comfortable with the idea of self-driving, once you experience it one or two times, it does. I try to say that to everybody.

It's so sticky. Because you get your time back. Suddenly you can be reading a book on your phone. It's just so sticky. My one way of convincing one person who likes to party, I'm like, you can, you can text and drink. I don't know what to say. There's my, that's my sale for you. But go ahead. But I do think this just represents like a big, like it's like taking all the different companies or market shares and shaking them up and saying, this is going to be a different looking outcome. And here's the part that I don't know the answer to and is, you know, it's provocative just to think about it, which is if market shares of some of these larger incumbent companies drops by even a small percent, these are very, very high fixed cost businesses.

And they're big machines. They're not easy to reduce costs. And we have an example of what happens when market, you know, the revenue drops by 20 to 30, 35 percent, and that's 2008, 2009. And so that's the part that's hard to predict is if, if some of these changes I talked about, the combination of being a software-defined architecture and having very high lows of AI integration are not, if you're not front of the curve and you start to lose 20, 30 percent of your volume, you just have the risk of being clobbered, like just completely clobbered by your fixed cost.

And that could lead to like this slowly, slowly, slowly, suddenly changes in the dynamics within the space. You know, interestingly, and I'll let you have the Westford. It's like Netflix, Circa, when they got off of discs. Yeah. And I remember talking to media companies. They said, you're finished what they're doing. You don't understand what they're doing here. And so it's like that.

And of course, here they are poised by up one of the greatest brands, the former. brightest brands if they make it through. So if that comes to pass, can Rivian be one of those things or just a junk heap on the... Yeah, well, you're asking a highly biased perspective here, but of course, like our view is, and that's what we're working towards, is to be one of the companies in the end that's both creating products, you know, for consumers that are just incredible and obviously software, obviously very high levels of autonomy and AI. But the second part is we do believe, and part of the strategy with our Volkswagen deal, we do believe there's an opportunity to help some of the existing legacy companies that weren't born the way that Rivian was born

and don't have this deep technology backbone to come up on the learning curve. But does that mean you might need an even bigger partner, like an Apple or someone else? No, no, I more meant like the existence proof that Rivian technology can be deployed into existing OEM. If you were to think of what's the ultimate existence proof, you'd probably say, let's pick one of the largest, most complex OEMs in terms of brand proliferation and market participation. And if you're looking for that, that would be Volkswagen Group. It's a very large group. It's got lots of different brands. And so our ability to demonstrate that we can deploy in a productized way are software and electronics. And this is not yet even talking.

about autonomy, just software and electronics, is the existence proof, or it is the demonstration that should show other manufacturers, hey, one, Ravion can be a great partner. We've demonstrated this now with Volkswagen. We've, of course, demonstrated this with another very large company, Amazon. And so I do think the opportunity for us to not only have our own products, but to emerge as a very, you know, very capable technology partner and provider that has both, both the ability to work with and the humility to work with very large existing companies is something we hope to demonstrate. So very last question, 20, 30 years, that puts you in Tesla at a complete competition,

presumably, if he can get his attention back to this. What does Rivian in 20 years say it survives this? What does it look like? Last question. I know you don't think in those time frames, but... I mean, you have to think about it somewhat. I mean, the way we make investments in technology, the coal is to be one of the most impactful and, you know, important businesses in transportation technology. And like, I do think we're at an inflection point aside from businesses, like from a societal point of view around how our vehicles are powered, what's going to happen as AI interests of the physical world.

And a huge motivator for me is to help be part of writing that story and to do it in a way that's, you know, to the benefit of society. Great. All right, RJ, thank you so much for your time. Thank you. Today's show is produced by Christian Castor Roussel, Michelle Alloy, Megan Bernie, and Kaelin Lynch. Nishat Kerwa is Vox Media's executive producer of podcasts. Special thanks to Catherine Barner. Our engineers are Fernando Aruta and Rick Kwan, and our theme music is by Trachidemics. If you're already following the show, you get to drive autonomously with me in San San Francisco. That's all I do when I'm in that city. If not, you get a horse. Go wherever you you listen to podcast, search for On with Carous Wisher and hit follow. Thanks for listening to On

with Caroushisher from Podium Media, New York Magazine, the Vox Media Podcast Network, and us. We'll be back on Monday with more.