Rivian at Goldman Sachs Conference: R2’s Strategic Leap

On Wednesday, 10 September 2025, Rivian Automotive (NASDAQ:RIVN) took center stage at the Goldman Sachs Communicopia + Technology Conference 2025. Founder and CEO RJ Scaringe highlighted the company’s ambitious plans, focusing on the upcoming R2 vehicle as a pivotal point for growth. While the R2’s lower price is set to broaden market reach, challenges such as manufacturing scale-up and supply chain readiness remain.

Key Takeaways

• Rivian’s R2 vehicle aims to significantly expand market reach with a starting price of $45,000.
• The company plans to achieve adjusted EBITDA break-even by 2027.
• Rivian is focusing on expanding into the European market and enhancing its direct-to-consumer sales strategy.
• A $5.8 billion licensing deal with Volkswagen Group underscores Rivian’s technological prowess.
• Rivian anticipates meaningful growth in its commercial vehicle business with Amazon.

Financial Results

• R2 Pricing and Demand

- The R2 starts at $45,000, attracting over 70,000 pre-orders within 24 hours of its unveiling.

- Rivian holds approximately 35% market share in the premium electric SUV segment with the R1, which has an average selling price of $90,000.

• Commercial Vehicle Business

- Amazon’s initial order of 100,000 commercial vehicles by 2030 highlights Rivian’s potential in this segment.

• Guidance and Targets

- Rivian targets achieving adjusted EBITDA break-even by 2027, driven by the R2’s lower cost structure.

Operational Updates

• R2 Production

- Production is set to begin in the first half of 2026 at the Normal, Illinois facility, which can produce up to 150,000 R2s annually.

- Current plans for 2025 include manufacturing between 40,000 and 46,000 vehicles across the R1 and commercial van portfolio.

• Supply Chain Management

- Rivian is emphasizing supplier readiness to support the R2’s production ramp-up.

• European Expansion

- The company is evaluating exporting R2s to Europe, contingent on favorable trade policies.

Future Outlook

• R2 Impact

- The R2 is expected to drive significant volume growth and improve unit economics, with production biased towards this model.

• European Market

- Rivian sees Europe as a promising market for the R2, depending on trade agreements.

• Commercial Vehicle Growth

- Anticipated growth in the commercial vehicle business with Amazon and potential expansion to other fleets.

• Autonomy and Technology Licensing

- Plans for Level 3 autonomy by 2026 and potential for further technology licensing to other OEMs.

Q&A Highlights

• R2 Demand

- Strong customer interest with substantial pre-order numbers.

• R2 Manufacturing

- On track to start production in 2026, leveraging the Normal, IL facility.

• Commercial Vehicle Business

- Continued partnership with Amazon, with prospects for expansion beyond Amazon.

• Technology Licensing

- Volkswagen Group’s $5.8 billion investment in Rivian’s technology stack.

For more detailed insights, please refer to the full transcript below.

Full transcript - Goldman Sachs Communicopia + Technology Conference 2025:

Mark Delaney, Analyst, Goldman Sachs: Okay, great. Thank you, everybody, for joining us. My name is Mark Delaney, and I have the pleasure of covering Rivian Automotive for Goldman Sachs. I’m very pleased to have with me today RJ Scaringe, Rivian Automotive’s Founder and CEO. Thank you very much for joining.

RJ Scaringe, Founder and CEO, Rivian Automotive: It’s great to be here.

Mark Delaney, Analyst, Goldman Sachs: RJ, I wanted to start off with the R2, really exciting future product. Tell us a little bit more about the R2 and why that’s so important for the company. On the topic of R2, the company had talked about having over 100,000 non-binding pre-orders for the R2. I realize you’re not disclosing pre-orders at this stage, but tell us a little bit more around how you think demand will be for that product and how you’re assessing that.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I mean, R2 for us is, it’s really the most important program we’ve had where it represents, you know, we’re going from our flagship product with R1, which has been, you know, wonderfully successful in terms of demonstrating the brand, connecting with customers, as we often say, being our handshake with the world. Nonetheless, it’s a high price point flagship product, ASP, average selling price for around $90,000, whereas R2 starts at $45,000. It massively broadens the aperture of demand for us. It really represents an inflection point or a step change for us to go to much higher volumes. With that, of course, cover our fixed costs and allow us to get to, you know, it’s a big enabler for us to getting to positive free cash flow. In terms of demand, the reaction to the product has been outstanding.

You know, it really, you know, leverages and jumps off of the brand that we built with R1. As you said, you know, a lot of orders have come in. We haven’t announced the numbers. We put a number out after the first 24 hours. We launched it in the spring of 2024. We unveiled it, and within 24 hours, we’d had around 70,000 orders. We just put that out there just to say, hey, people really love it. We didn’t want to get into the practice of having to update the order number, you know, every month or something. We haven’t said anything since then. I think there’s been some reports shortly thereafter that it eclipsed 100. I think there were some reports that it eclipsed further numbers. The excitement is really strong for it.

For us, we’re taking that as a really key motivator for making sure we get the product right, have a smooth ramp, making sure the supply chain is healthy and ready to support our growth.

Mark Delaney, Analyst, Goldman Sachs: You guys were kind enough to have me at the launch event. I can vouch for the fact that it’s a really awesome looking vehicle. I’m not surprised to see you guys rack up that many pre-orders so quickly for it. Of course, you have to manufacture it, right? Maybe we can talk a little bit about that. You’re supposed to begin shipping the R2 in the first half of 2026. It’s less than a year away now. Tell us more, what’s giving you confidence you can hit that target in the first half of this coming year.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I mean, we’re doing all the things you’d expect us to be doing. We’ve got validation builds that are running today. The health of the program relative to everything that we’ve done before is really high. We’re just such a mature organization relative to, let’s say, when we launched our R1 product. That means our suppliers and the bring-up of those suppliers and the inclusion of production content and prototypes gives us a lot more confidence in the ability of all those suppliers to ramp up. The contracted cost that we have to support the bill of materials in the vehicle and the associated cost, which is for the bill of materials, which is a big driver of the overall COGS, has been really robust.

The plant that we’re going to be building this in, we’re building it at our campus in Normal, Illinois, which is where we build our R1 and our commercial vans. We built out a new general assembly area for R2 and a new body shop, and that’s come together really nicely. Today we’re like head down. The whole organization’s focused on it. Everyone realizes the importance. Getting the enthusiastic response from customers, as I was saying, has been great. Great motivator. I’ve been driving an R2 for a little while now, including picking up my kids from school. It’s a camouflage vehicle, but it is so awesome. It’s, without question, I’ve never been this excited about a product for anything that we’ve ever developed. It’s really cool.

Mark Delaney, Analyst, Goldman Sachs: I think the Normal facility has the potential at full capacity to make 150,000 R2s per year. This year, the company is planning to make about 40,000 to 46,000 vehicles and deliver that many across the R1 and the commercial van portfolio. How fast do you think you can ramp R2 to that kind of scale? Are there any key learnings you’re incorporating from the R1 ramp?

RJ Scaringe, Founder and CEO, Rivian Automotive: One of the biggest challenges we’ve had with R1, we launched R1 at the very end of 2021. That was, of course, we built the plant, set up the plant, and then launched the product in the middle of COVID. Subsequent to the initial launch, we then had a lot of the supply chain crisis, which seems like it was a long time ago, but in 2022 and extending into 2023. As a new company with a new supply base, we had just a whole host of challenges with getting enough parts to build our vehicles. With R2, we spent a lot of time making sure, as I described it before, not only is the part design and the system design robust, but are the suppliers capable of supporting our ramp up?

That is ultimately the gating factor for how fast we can ramp production on this new line in our facility in Illinois. The work we’re doing now is to make sure that that’s ready. We’re not waiting until the last minute. We’re doing supplier validations. We’ve got lots of teams that are now built processes around being on site and working with those suppliers. There’s a whole host of proof points and gateways, as we call them, that look at readiness of each supplier to support this pretty aggressive ramp that we have going into the end of 2026.

Mark Delaney, Analyst, Goldman Sachs: You said before R2 is a vehicle that could make sense for the European market. Tell us a bit more on that and what sort of timing might investors expect for when you could be shipping R2 to the European geography.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I mean, R2 and its sibling product, R3, were both designed from the get-go to support not only U.S. sales, but also European sales. We haven’t announced the European timing, but amongst all the things that are changing in terms of our trade environment and trade policy, it looks like the likely outcome of 0% tariff to go from the United States to Europe is going to stick, in which case that makes Europe an even more attractive market for us to export R2s from the United States into Europe. We’re watching that really closely. Provided that sticks, I think it’ll be a really strong tailwind for us to be more aggressive in rolling out in Europe.

Mark Delaney, Analyst, Goldman Sachs: You said you could export out of the Illinois plant to serve that geography, but is there a lot of local infrastructure you would need, distribution, service, things like that? Maybe talk a bit more on some of those other aspects if you do go into the European market.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I mean, we’re direct to consumer, and a big area of investment for us as we’ve grown has been building out service infrastructure, distribution infrastructure, of course, sales infrastructure. As we go into Europe, we don’t have a lot of that infrastructure that has to be built. The environment there is quite a bit different than what we have in the U.S., where you can have more of an omnichannel-based approach, where in the U.S., you can either be entirely direct or you can leverage third parties to sell your product. You can pay them a pretty significant amount. In the U.S., we’ve been very specific on staying direct to consumer. That is still the plan for Europe, but it’s just worth noting that the ease of accessing things like service infrastructure is a little bit easier.

The other thing to note is we do have some footprint in Europe today. We have service to support the rollout of our commercial vans with Amazon. We’ve got some experience in both setting up infrastructure and then operating that infrastructure. In Europe, we have our European entities to do that already, but those need to be scaled dramatically to support R2.

Mark Delaney, Analyst, Goldman Sachs: I want to talk a bit about manufacturing technology, not necessarily specific to the R2, but just as you think about Rivian and the industry. A couple of your competitors, including Tesla and Ford, have talked about moving to a more parallel type of manufacturing. Tesla’s called it unboxed. Ford called it an assembly tree. Is that kind of manufacturing technology and approach something that Rivian would consider in the future?

RJ Scaringe, Founder and CEO, Rivian Automotive: I mean, our biggest focus when we think about vehicle program is we look at it, we look at the cost structure in a really holistic way. By far, the biggest cost driver for every manufacturer, ourselves, including Tesla, including Ford, is the cost of the bill of materials. As much attention as the plant gets and as important as the plant is, it represents on the order of 20% of your overall COGS. With that said, a lot of our focus around innovation on R2 has been in simplifying the product design through either part consolidation or part elimination. In terms of the exact sequence that that then manifests in terms of how the car comes together, a lot of that ladders off of simplifying the part design or simplifying the product design. Sometimes you have competing objectives.

To make the plant flow easier, you may want to more modularize the vehicle, but then you have more joints or fastened assemblies. Versus to simplify the bill of materials, you want to have less parts and less joints. We’ve come up with, on the launch of R2, something that still uses a linear line, so less parts being fed to the line as the car gets built up, but has dramatically reduced the amount of cost in the vehicle through this part consolidation, part elimination. As we look at Georgia, which is our next phase of R2 production, we are looking at further ways to evolve our manufacturing model. It’d be premature to say what those look like today. It’d also be premature to categorize them as being, let’s say, highly similar to what Tesla’s doing or to what Ford’s doing.

Mark Delaney, Analyst, Goldman Sachs: Okay, very interesting. Maybe we could talk a bit on some of the current products you’re selling, starting with the R1. In your Illinois facility, you have the potential to make 80,000 R1s a year. This year, something sub $40,000. What would have to happen for R1 to get to those kinds of volumes where you’d use up that level of manufacturing space? Would it require a much lower priced SKU of the R1, in your opinion?

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, as I said, the R1, this is public, the R1 average selling price is around $90,000. In terms of market penetration, it’s by far the best selling premium electric SUV in the country. If you look at the premium SUV market, let’s say in the state of California, we actually have the best selling premium SUV, electric or non-electric, in California. That’s also true in the state of Washington. If you look at premium electric SUVs, we have the best selling premium electric SUV in the U.S. It outsells everything else, Model X, products from European brands. It does really well. Of course, the boundary diagram matters when you’re thinking about market share, but the way we’ve drawn the boundary diagram, we have, we believe, about 35% market share there.

If we can take that type of market share success that we’ve had in a segment that has a relatively limited number of buyers because of the price point and apply that to where the R2 is, which is the biggest market. The average selling price of a car in the United States is around $49,000, and the most popular configuration is a five-seat, five-passenger crossover SUV. We only need a fraction of the market share success that we’ve had in R1. We’re very, very bullish on the market for R2. I think the natural question then becomes, as you put it, how we divide our volume. What we’ve done in our plant in Illinois is we have some fungibility between building R2s or R1s. The constraint for us is our paint shop.

We could either build, as you said, up to 80,000 R1s, but that would come at the cost of some of the R2 capacity, or we could build more R2s. Between R1, R2, and our commercial vans, we can build a total of 215,000 units. What we’re planning to do is to bias as many of those as possible towards R2, given its really advantaged costs and therefore really strong unit economics.

Mark Delaney, Analyst, Goldman Sachs: Okay, that makes sense. R1 is a premium product, and you have some premium trims. You launched the quad motor variant in July. On your last earnings call, you talked about the tri motor, and you said demand there has been quite strong. Can you talk a little bit more on what you’re seeing for those products? Really, it’s the premium, but I imagine it also has some good margin and obviously really good performance.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, we’ve seen the take rates for our premium variant, in particular this tri motor that you just referred to, be much stronger than we had anticipated, which is a great outcome. That’s what we hope for. We’re seeing our ASP trend upward as a result of that. I think with the launch of R2, we think that R1 will further move further up market because it’ll take some of the more price-sensitive customers, and they would be more naturally drawn into an R2. That gets back to this volume split as we have the opportunity to move R1 up market, let its ASP grow a bit, and then capture more volume-sensitive customers and therefore some of the larger number of customers with the R2 vehicle.

Mark Delaney, Analyst, Goldman Sachs: Okay, makes sense. I just wanted to speak a bit more on the commercial part of your business. Amazon initially ordered 100,000 commercial vehicles by 2030. The estimated deliveries to date would imply a step up in EDV deliveries from here to reach that total. Is that 100,000 total still on track?

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, yeah. It continues to go well. It’s ramped slower than we had originally hoped for a lot of reasons, but year over year, we expect it to grow quite meaningfully as we look at going into 2026 and 2027. We’re also thinking about what comes beyond that initial 100,000 unit contract. As I think all of you know, Amazon’s fleet is considerably larger than that, and there’s real opportunities for us to continue to penetrate across the fleet. We are a vehicle that the drivers very clearly prefer. This is pretty well documented. There are lots of videos about this out there on just the driver preference for our vehicle. One of the big cost drivers for Amazon is driver retention or driver turnover. The fact that the turnover levels are much lower with our vans is just emblematic of how much the drivers enjoy being in the vehicle.

Mark Delaney, Analyst, Goldman Sachs: You talked about the potential to expand the commercial business to other fleets and businesses. What do you think it would take to see that kind of response and increased traction beyond Amazon in the commercial part of the business?

RJ Scaringe, Founder and CEO, Rivian Automotive: I think in the fullness of time, we’re going to see the commercial space will electrify. I think for a lot of reasons, it’s gone slower than we thought it would. I think the natural big buyers of commercial vehicles have been slower to adopt. There are lots of compounding factors for that, one of which is some of the businesses that would be buying these are capital constrained in the moment. You also have a big shift that’s happened geopolitically, and a lot of these companies are sort of aligning to that and being careful in how quickly they electrify. I think over the course of the next five years, we’re going to start to see a number of folks jump in. What’s going to drive that is nothing other than the economics.

The total cost of ownership and the total cost of, let’s say, delivery is notably lower in an electric vehicle if you take a long enough time horizon. The operating costs are much lower, of course, but you also have much lower service intervals. As I said, because of the way the vehicle has been designed, we have much higher levels of driver retention.

Mark Delaney, Analyst, Goldman Sachs: No, that makes a lot of sense. Let’s talk about some technology topics. I’ve really been looking forward to the opportunity to speak with you on some of the things Rivian has underway in that broader area. Maybe we could start with the electrical and electronic architecture. You have an agreement with Volkswagen Group and a joint venture to provide about $6 billion of capital in various forms to Rivian over time. It gives Volkswagen Group access to utilize your electrical and electronic architecture. I guess the question that comes to mind for me is why do you think Volkswagen Group chose Rivian as a partner? Why would they utilize a partner at all? Why Rivian as opposed to doing this in-house or maybe even using a tier one?

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I mean, it’s worth taking a little bit of a step back and talking about how electronic architectures, as we know them today, got to where they are today. This is really important as we start to think about software-defined vehicles and vehicles that have regular updates and new rich features being added to the vehicle set. The way that electronics first made their way into cars started actually in the 1960s and started ironically with fuel injection systems. Prior to the early 1960s, every car made was completely analog, meaning there were no semiconductors in a car. There were no computers in a vehicle. The first computer system that made its way into a car was a small little chip that was part of an electronic fuel injection system.

For those car enthusiasts in the group, these are things like the Bosch K-Tronic platforms and some of the things that, if you’ve restored old cars, you may know of. That set off the very beginning of a highly organic and, I would say, very unplanned proliferation of electronics into vehicles. What subsequently happened over the last 60, 70 years has been a bunch of different functions went from being analog to being electrical. As those functions became electrically controlled, they transitioned to that with a controller. Let’s say your fuel injection system was the first. Then you had electric seats that suddenly had a little microcontroller that came with them. You had an HVAC that had a little microcontroller that came with it. You had electric windows that had a little electric microcontroller with it.

Before you knew it, over the course of many decades, a vehicle with a lot of content may have 100 of these little, what we now call ECUs, electronic control units. Each of those ECUs represented an island of hardware, so a little compute module, and then, importantly, an island of software. That little island of software was provided by the supplier. Over the currents of decades, many of the software started to outsource that as well. The supplier that provided the ECU doesn’t, in fact, even do the software themselves, but a third party does it. Fast forward to today, and never designed as an architecture, but ultimately the result of what we have as an industry, with the exception of Rivian Automotive and Tesla, is vehicles. Like you take a high-content vehicle, it may have 120 little computers, 120 islands of software.

It is not only not what you would design if you were to look at it from a clean sheet. I’d say it’s almost precisely the opposite of what you would want to design if you took a clean sheet approach. If you were to take a clean sheet approach using first principles, you’d say, I’m going to have the smallest number of computers in the car possible. Those computers are going to do many different functions. They’re going to be doing functions that are geographically close to those computers. Depending on the size of the car, you might have one computer, you might have two computers, you might have three computers, but you’re certainly not going to have 110 computers. You’re certainly not going to have 110 little islands of software written by different teams, often by different teams abstracted, two layers away from the OEM.

The reason all this matters is when you have this highly abstracted system of many islands of software written by many different teams, it becomes very cumbersome to do an update to the vehicle. If you think of it in terms of a customer-facing feature, making a change to, let’s say, the sequence of events that occur to what happens when you walk up to the car in a classic architecture, like what you find in essentially every vehicle other than a Rivian or Tesla, you’ll have to coordinate amongst maybe 15 different suppliers who ultimately may have multiples of that of suppliers who provide the software. If you want to change the exterior sound when you walk up, the lighting sequence, what happens to the HVAC, the seat setting, maybe the interior lighting, each of those is a different computer, each of those is a different stack of software.

It is very, very, very challenging to do meaningful over-the-air updates that add new features and new capabilities. I’m describing it as a reference example around the sequence of what happens when you walk up to the car. There are much more complex sequences that are being contemplated. Out of that has been born this concept of a software-defined vehicle where the vehicle itself has much more sophisticated computers, much fewer of them. Like our R1 has three computers that run the whole car. The ability, because we control the entirety of the software stack from the base operating system all the way up to the middle layer, all the way up to the applications layer, we can do updates very easily, and they can be very big. They can be robust. They can touch many parts of the car.

I think every car company in the world needs to be thinking around how do they move away from this legacy-based system that was born out of the 1960s with fuel injection systems and never designed to something that’s highly architected and highly thoughtful around a much more software-heavy environment. You have a couple of choices as a manufacturer. You can build it yourself, but it takes skills that typically car manufacturers don’t have. You could go to a tier one, but there’s a very obvious conflict of interest here. The tier ones love selling lots of little computers. Like this architecture I’ve just described with the zones and much more capable computers doing many functions, you’re removing many thousands of dollars of costs. You go from 110 little computers to, depending on the size of the vehicle, maybe two or three slightly larger computers.

It’s a much, much cheaper architecture. The third option is you go to someone who has it and licenses that technology. In the case of companies outside of China, companies operating in the Western world, there’s two choices. That’s it. It’s very limited. It’s us or Tesla. All that said, it led Volkswagen to doing this very large, one of the largest software licensing deals in the history of the auto industry, a $5.8 billion deal for us to provide them this technology stack that supports their whole portfolio of brands. It’s Porsche, Audi, VW is the brand, brands that exist in Europe and not in the United States, brands like Skoda or Cupra. These are a portfolio of different brands that are now going to be utilizing this platform. Our view is every manufacturer is going to be faced with the need to do this.

One of those three choices will be a path they’ll take. They’ll try to build it themselves. It’s very hard to do. They’ll try to convince their tier one suppliers to do it. That’s, I would say, extremely hard to do. Or they license it from us.

Mark Delaney, Analyst, Goldman Sachs: You talked about the potential over time to also license this potentially to other auto OEMs. If this architecture is software-defined, why does Rivian need to hold off on bringing other OEMs in to use this product?

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, right now, we built a platform. It’s very scalable. What we’re doing at Volkswagen really represents the existence proof of its scalability. One of the early programs that’s going to be launched with Volkswagen, which is now publicly announced, and I was just in Munich yesterday, and they have the vehicles on the floor there at the show, is what they call the ID.1. It’s a $22,000 EV that is achieving that price and cost level really heavily through the use of this advanced architecture, where you dramatically reduce the amount of electronics in the car, really simplify the network architecture. That serves as, in many ways, like the ultimate billboard for why a software-defined vehicle is important. One, it creates a vehicle that’s very compelling to consumers, but it does so at a much lower cost basis.

As all of you know, the nature of the auto industry is people look to the left and right and see who’s doing things better. When that car is available, you can imagine every car manufacturer in the world is going to buy it. They’re going to take it apart. Usually, they’ll be like, oh, this is really impressive how much I like this. Then they’ll take it apart and they’ll say, holy cow, how are there only, how’s there so few controllers and how’s this architecture so simple?

I think it’ll be a bit of a wake-up call to say either we have to go figure out how to do this on our own, which is very hard, and they’ll then be competing with someone like us that’s got, you know, well in excess of a decade of working on this problem, or they’ll come to us and say, can we also license it? We do see an enormous amount of opportunity for additional licensing of this technology. We designed the joint venture to accomplish that.

Mark Delaney, Analyst, Goldman Sachs: No, that makes a lot of sense. One of the important applications these architectures support is autonomy. Rivian plans to offer situational or Level 3 autonomy next year when drivers can at times take their eyes off the road. Talk a bit more about how that feature may look for Rivian in 2026. Is this an eyes-off and a traffic jam kind of scenario at low speeds on a very limited number of highways, or are you thinking Level 3 in a wider way and maybe operating at full highway speeds?

RJ Scaringe, Founder and CEO, Rivian Automotive: At start, it’ll be specific domains. It’s going to feel like when you’re waiting for it, it’ll feel like it’s taking a while, but before we know it, in the fullness of time, I’d say in a couple of years, the expectation that a vehicle can drive itself with your hands off the wheel, of course, but also your eyes off the road and it performing turn by turn. I think by the end of this decade, by 2030, is going to be a really clear expectation from consumers. To do this well, it requires the entire vehicle to be architected to support this. At a foundational level, you need a vehicle that’s software-defined. You need a robust network architecture. Think of that as like that’s the groundwork, that’s the site work to even enable some of these types of features.

At the autonomy level, you need a platform that’s been designed around a data flywheel, meaning the perception stack, whether that’s cameras or cameras plus radar plus LIDAR, needs to have all those signals provided in raw form into an onboard inference platform that’s running a distilled version of a very large multi-billion parameter foundation model, large neural net that’s been created through a set of thoughtfully designed triggering events for the fleet that’s deployed, meaning the fleet that deployed is the training platform. We often do this called like an end-to-end approach where the vehicles are training this large model end-to-end. This end-to-end trained foundation model approach, the amount of progress we’re going to see on this in the next five years is markedly different than if you look at the amount of progress that’s happened in this space in the last five years.

We have to recognize it’s because there’s been massive technical breakthroughs that have completely changed how we develop self-driving. Relative to, let’s say, pre-2021 or pre-2022, where we were operating in a much more rules-based environment where the cameras or the perception stack would identify and classify objects, assign vectors to those objects, and hand that to a planner, and the planner would be based upon a set of rules written in code written by humans. We now have that perception stack feed a model that’s being trained with the benefit of the full fleet to build this large multi-billion parameter model that represents how driving is executed. The rate of progress on an AI-centric model versus a human-built planner is, I mean, it’s just, it sort of gloves off.

It’s going to be, I’m very, very, very bullish on what we’re building, and I’m very bullish just on this space in general. Our expectations as consumers by the end of the decade are going to be that the vehicle can drive itself, give me my time back, do point-to-point directions. I get into my car, say, take me to work, and it takes you to work. This is our biggest investment category as a business, much like what we’ve done on the rest of our software and technology platform. We’re very strong in software and electronics, and this is a huge initiative for us as a business.

Mark Delaney, Analyst, Goldman Sachs: There’s been a lot of partnerships announced in this area of autonomous vehicles because there is so much potential as an industry here. We’ve seen a number of announcements already. I’m curious if Rivian would consider partnering in this space. We’ve seen a lot out of some of the rideshare companies like Uber and Lyft, and that may even be a way you guys bring in capital. Talk a little bit more about how much you guys do yourself and is there partnerships to be had and maybe even with some capital behind it.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, I think in terms of the go-to-market, Robotaxi versus personally owned, Robotaxi gets a lot of attention, but the rideshare business is actually much smaller than the personally owned business. If you look at it in terms of miles, about 99.5% of the miles driven in the U.S. are in personally owned vehicles or household-owned vehicles. Only about 0.5% is in rideshare. I don’t say that to diminish the importance of rideshare. It obviously is important, but the application for a self-driving capability, a self-driving capability that’s able to go hands-off, eyes-off, and give your time back, extends from personally owned all the way into Robotaxi. Our focus today is on the technology and first applying the technology in our vehicles that are personally owned.

We’re, of course, open to and thinking about the types of partnerships that would allow that to eventually be used in Robotaxi applications as well. It’s likely the number of rideshare miles is going to increase as this capability grows. Does it increase by 10x, 20x? It’s hard to say today, but I think the core is solving the technical problem. Once the technical problem is addressed or solved, then there’s a plethora of different business models. There’s rideshare, there’s charging for autonomy, there’s charging for autonomy upfront, there’s per mile, there’s a lot of really interesting go-to-market debates that are going to happen and we’re going to see play out in front of us. Foundational to that is first solving the technical issue.

Mark Delaney, Analyst, Goldman Sachs: I’ll do a plug. I know you’ve got an AI Day you’re planning for December, so we’ll look forward to learning more there. If I could just sneak one last question in in the last less than minute we have here, you’ve targeted to be adjusted EBITDA break even in 2027. The industry is facing tariffs, IRA purchase credits are going away starting next quarter. As you think about some of these changes, in your opinion, is it possible to still hit that target? If so, what would get you there?

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, it’s not like we believe it’s possible. The R2 program was designed with, as we talked about before, a much lower cost structure, and that lower cost structure supports a healthy positive gross margin on the R2 program. We did that contemplating a world in which a lot of the incentives that had existed previously, we did this contemplating a world in which they wouldn’t exist. Certainly, you know, some of those going away create a short-term headwind. In the long term, in the fullness of time, we think a highly compelling product at the right price with the right margin structure is really what’s going to drive the business. We’ve made decisions around the vehicle, around the product specifically for that.

The ramp-up of R2 that occurs in 2026, what that then does to enable 2027, as you said, with positive EBITDA, that’s, you know, we’re working towards that. We still, as you said, we’ve continued to guide with confidence towards that as well.

Mark Delaney, Analyst, Goldman Sachs: Great. Unfortunately, we are out of time. RJ, really appreciate you joining us.

RJ Scaringe, Founder and CEO, Rivian Automotive: Yeah, thanks for having me.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.