General Motors Company (GM) Earnings Call Transcript & Summary

Yes. We've had a mindset from, again, from the outset of our philosophy almost of we're -- we want to work on the difficult version of the problem and work on it in a way that allows us to take what we build there and then scale it to other environments. And so we've been quite careful from -- as we build infrastructure and how we architect parts of the system to not get ourselves in a position where you have to basically reinvent everything to go to a different environment. So we've had -- with the work that we're doing today and how we're building things today is done with an eye on that need to scale from there.

So at the Capital Markets Day a couple of weeks ago, you said that the herd of AV developers is thinning. So maybe there's 5 or 6 that are kind of still working full bore on this issue and kind of anybody can say that their technology is better, and it's really impossible for us to evaluate. Talk about some of the things outside the self-driving system that you feel like gives Cruise an advantage over others in terms of getting to that first-mover advantage that you've talked about before?

Yes. Well, I'd agree with you. The signal-to-noise ratio is low in terms of like trying to define what's going on amongst different companies here. The -- again, sort of back to strategically how we set this up from the beginning, building that core technology to the point of delivering superhuman driving performance is a big engineering challenge. And it's a critical part of building a business out of this, but it's not the only part. If you go back to those dimensions that I described that we're trying to compete on, we know what the existing product is in the market today. We want to build something that's safer, which is really the core technical challenge. We want to build something that's a better user experience. And so that gets to the in-vehicle experience and the nature of the environment that you're in. It gets to cost-to-cost dimension, which is, can we deliver something at a much lower price point than where the existing modes are today and can we do it in a way that's better for the planet? And those are the -- again, these 4 dimensions that we're trying to drive on -- drive improvement along on a continuous basis. And we're working on all of those in parallel. Because if you sold one of those and you show up, you still don't have a product, you have a piece of a product. And so our goal has been to work in each of those things in parallel so that we can show up with something that we can actually present to the user and say, "Here's something that's better than what you have today." And by the way, it's going to get better and continue to get better over time. And that also informs why we've configured Cruise the way we have and why we've built as much and acquired as much talent as we have and why we have put together as much capital as we have, is we want to be able to pursue all of those things in parallel so that when we show up, we show up with something that's complete.

Yes. That makes sense. That makes sense. If you're successful and this kind of proceeds a year, 1.5 years from now, you're in the first city in 2021, could you sort of take us out into 2025 and 2030? And what do you think the penetration of this service really becomes?

I think that depends a lot on back to this notion of how quickly can we continue to improve the product and how quickly can we get the cost point down and how quickly can we deliver something to the user that's better than whatever they're doing today. Because the faster we can do that, the faster that we'll have people say, "You know what, that's better than this, and therefore, I'm going to use this thing here instead." And so I think it's -- you can make predictions on penetrations. But the more important question is, how much conviction do you have in your ability to deliver a superior and continuously improving product. Because that is the thing that's going to drive, not just people converting from existing modes to this, but from whatever other competing AV offerings are out there. Because if today -- if you have 2 rideshare buttons on your phone, you push one button or the other, you kind of get the same thing showing up, sometimes it's even the same driver, right? But if you're at the point where you have a product that you push the button and it's like this seamless, magic carpet experience that's super inexpensive and its totally reliable over time; and you have this other one that's kind of a janky, minimum viable product on the other hand, you're always going to use the one that just works. And that's why this -- and that gets better over time. And that's why this rate of improvement and continuously improving thing is sort of fundamental to what we're doing.

Yes. Do you worry about being disrupted as well? Will there be disruptors to the disruptors? And maybe just to describe what -- there's a lot of talk about companies that are talking about AVs that will eventually be everywhere, right? The notion is if you and I can drive around with our 2 eyes, eventually, you're just going to have a camera with some computer that has human level or better than that level of cognition, and it'll be simpler and cheaper. Is that something that you think is realistic? And does that really threaten Cruise a couple of years after deployment?

It would threaten us if we were static and we weren't improving and we didn't have this objective of continually, in a way, disrupting ourselves even every year, every cycle of technology and so on. And I think this -- you've seen other technologies, the evolution of the spot firm where someone comes out, and like here's a thing and it becomes the standard. And it's standard until someone else comes out with something better. And so if you pull on that traffic not continuously improving what your product is, then you're at high risk of that kind of thing happening. If your mindset is we need to just make this better every single moment of every single day, then you're much less likely to have something sneak up on you like that.

We should see if there are some questions from the audience. But in the meantime, I think when you launched the Origin, you made a comment about the cost of that vehicle relative to a normal vehicle. Can you kind of reiterate that for us and kind of give us a sense of how the vehicle cost is built up in terms of what can you remove that's not needed that's in a normal vehicle? Like how are you thinking about kind of the initial kind of bill of materials for that vehicle?

So there are some obvious things that we don't need like steering wheels and driver controls and those kind of things. And that's a nontrivial thing to take out. There is the performance envelope of the vehicle. The Cruise Origin is obviously built to drive in cities and on highways and all sort of operating environments, but it doesn't need to have 155-mile an hour top speed or a 0 to 60 time of 2 seconds. So there are costs that you can take out because the performance envelope and requirements are different around that. And then there's other places where we put some costs back in to drive that million mile plus life cycle for the core of the vehicle. And so the mindset that I'd say that we brought to it is to put the money in precisely where it matters and to take all the money out precisely where it doesn't; and be very, very sort of ruthless in how we go about doing that because the goal is to deliver an awesome experience at a super low cost and consistent as well. So you know every time you push the button, you know what you're going to get. And so it's about being very deliberate about what the use of the vehicle is and what it's not. And because we control exactly how it's used and how it drives and so on, we can take out a lot of that sort of excess that's around that.

And I think like most things you hear people are talking about in terms of at $160,000, $170,000, AV-based mobility-on-demand can compete with incumbents. Or is that the magnitude of vehicle costs you're talking about? Or are we...

No. We're talking about less than that?

Yes. Okay. Okay.

I think we've got a question in the back.

Yes. You were very clear about the rate of improvement and reliability and technical performance and the trajectory. And I'm not aware of the history, so forgive me. But it does -- I was reading some historic articles. It does seem that there's been some slippage in the original plans. Have you always had that degree of visibility over the improvement? Or maybe just fill us in about what changed during the course of last year on that improvement trajectory.

Yes. I'd say -- it's a fair question. I'd say from a sort of philosophical approach, nothing has really changed. Like that underpinning of the importance of rate of improvement has always been sort of the foundation of our approach to do this. One thing that we have done over the time is we have expanded the operating domain and sort of the definition of success that we're initially targeting, so sort of raised the bar from a technical challenge point of view. And so when -- the nature of this, and this is the question everyone likes to ask is, "Well, exactly when is this going to happen?" When you're out on a very long tail of a thing that's been improved exponentially for a long time and then bringing exponentially more powerful tools and resources to that, predicting exactly how those 2 things are going to offset and therefore what the time line is to that moment of superhuman performance, it's hard to precisely predict. But -- so we improved -- we've been improving exponentially consistently for several years now. It doesn't always go on a nice linear, upward straight line. You have periods of incredible improvement, and then you have some periods of flat spots and then other periods of incredible improvement. So when you get to the last stage of this, trying to get super precise on time lines is challenging, but we have a very high degree of conviction in the rate of progress we have, the backlog of things that we had to work on to drive through this next sort of phase of performance improvement. All the output of the engineering machine of the company is -- feels very, very good. So I feel really great about where we are and the rate of trajectory that we're on, but that last little prediction of exactly how the timing is going to shake out is tricky.

Any other questions? There's one at the back.

You commented earlier on moving from single rider to shared rides. Other than cost, what variables or factors do you think come into play that will motivate and incentivize the consumer to be willing to take more rideshare -- shared rides?

Yes. I think cost is a huge driver and then the second one is the in-car experience; and am I sitting right next to somebody that I don't want to sit right next to? And am I crammed in the backseat of a car? And for example, these 2 guys here...

He looked at you first.

He looked right at me after saying that.

But the -- and so that's where the Cruise Origin plays such a significant role, which is if you get an opportunity -- or when you get an opportunity to sit in that, you're sitting not right next to the person and you're sitting pretty far away from the person that's across from you. And you really feel like you have your own personal space in there. And so all of a sudden, you've gone from being crammed in the back of a car with someone you don't know to all of a sudden, I'm sitting in this space. I've got my own personal space. I don't feel like anyone's intruding in it. And if this is costing me a fraction of paying for my own thing, we already have a fair amount of evidence that price elasticity is very high if you can offer someone a meaningful discount to do that. So it's simply pricing user experience, and that's a big part of what's behind the Origin.

We've -- yes, there's one right here.

Just on the earlier comment around 24/7 utilization. I mean the Origin, it's an EV. It needs to be charged. And I don't think you're using like a battery swap technology for it. So what are the implications for that for that sort of capacity, if you will?

Yes. So the -- obviously, they need to be charged. They need to be cleaned. They need to have other updates or software updates or other things will be going on, data download, all of that. So we're learning that today because the fleet we're running today is an all-electric fleet as well, basically. So we're learning about utilization. We're learning about charge time. We're running 24/7 today with our development fleet to gather those learnings. So I think we're already in the -- again, relatively early stages, but we're already learning a lot about how to do that, how to optimize; interesting things like are you better off to charge the car completely or there are times where you're better off to partially charge it and send it back out again and it is optimized that way. And so there's already a fair amount of sort of data science going on around that to understand how to optimize, how optimization model's getting built around that. So -- but that will be another layer of learning.

Can I ask you on -- yes, go ahead.

So I understand that long tails are extremely hard to predict and timing. But if I look at what's priced in and then I look at the stock, one could probably argue that either very few people are pricing in that it's going to work eventually or that financially, it will work. So help us get more comfort around not timing the fact that ultimately, yes, this will be successful, why you're so confident in that and it will be financially viable.

Well, it comes back to just basic premises. We know how big the market is today. We know what the existing [indiscernible] are in the market and how they perform in terms of safety, cost, user experience. And we're taking this, I think, fairly principled approach of if we can build a product that's better than that, then it's not too bold an assumption to assume that you will take market share from that market. And so we've talked at some length here about where we are on the safety dimension and getting to that initial superhuman level of driver performance. We've talked about where we believe we need to be on a cost curve and how much opportunity there is moving down that cost curve. We've talked about the user experience. The question on the shared rides and the Cruise Origin and how we're solving that. And then I think there's a sort of a higher purpose piece of this, which is that we are solving some of the fundamental problems of transportation and doing something that's good for society here. And I think that's important because the societal pull for this to occur as people begin to appreciate them more and more, I think, will be really, really strong. And so I think there's a -- in a way, if you look at it from that point of view, there's an inevitability that this has to happen, right? This has to get solved for all of those kind of reasons. And so we've approached it in this principal basis of: can we deliver something safer? Can we deliver it at a lower cost? Can we deliver with better user experience? And built everything we're doing around getting to those things. The timing of the thing can move around a little bit, but I think there's an inevitability to it occurring. And we have the road maps and robust plans behind each one of those dimensions on how to make those occur.

Can you talk maybe just about the behind-the-scenes work that's going on? And you alluded to this, there's a lot of learning on how fully do you charge the vehicles and that sort of thing. But presumably, there's a lot of infrastructure that need to be put in place. There's teleoperation that needs to be put in place, facilities where you're staging vehicles. What actually -- given that this may be a year, 1.5 years away, I would imagine that some of that is already underway, can you give us a sense of what is being done right now? And how management-intensive the back end of all that stuff actually is?

Well, I think one of the things that occurs when you get out the sort of far end of the initial challenge where we are is the amount of on-road testing that's necessary for validation and to build that final safety case as to where you are, necessitates that you have to build a sizable fleet and learn how to operate that fleet 24/7 and learn how to do that at an efficient rate of utilization. And so we're in that very sort of early phase of doing those things already. So to give you just one example of that. One of the many cost line items in the model is charging infrastructure and how does that work? And so between 2017, when we were doing some of our initial sort of first larger-scale charger installation and the installation we're doing at the moment, the operations team has worked with charger vendors to take the cost per kilowatt of charging capacity down by 85%. So not sort of down by a factor of less than 1/5 or something -- less than 1/5 of what it was. So a huge reduction in that just by working with different vendors, figuring out the optimal way to configure that, figuring going the optimal way to utilize the layout of the charging facilities. And so even in these early days, we're making huge steps forward on a lot of those learnings. Much more to go, but really encouraging progress.

Question?

So on telemetry and connectivity to the vehicle, we have multiple providers, multiple layers of redundancy in that. The current infrastructure works for what we need for it to work. So there's no reliance on 5G coming or anything like that to do what we need to do. On remote operation, we have our own capability, remote assistance, our own capability built around that, that we use every day.

In terms of like evaluating new cities. I mean to the extent that you can, what's kind of the funnel of criteria that you use? Like what's something that has to be there? And then as you kind of get down the funnel, how do you kind of choose, put together a priority list of cities?

I think it's sort of all the obvious things. It's: what's the commercial opportunity; what's the sort of regulatory environment and receptivity; what's the sort of environmental conditions, given weather constraints and things early on, what does that look like? So those would be fairly -- sort of some of the major things that we'd go in early on.

And I mean it seems like somebody the other day said that there are cities all over the place that want to bring us in and have us test. Is that your -- I guess my view was always that those are kind of maybe the smaller-, medium-sized cities that might not be as kind of lucrative or as kind of useful of a market for you. Do you see big, dense urban areas kind of receptive to this technology? Or are they more resistant to it?

I'd say it's a mix. It's very hard to generalize on that. Different cities have different priorities and different levels of sort of progressiveness around some of those. So we're starting in San Francisco, which is a challenging environment from a driving point of view and so on. And we'll sort of work our way up from there.

Got it.

Wanted to ask about -- I think we have maybe 2 or 3 more minutes. But what are your biggest worries? What are the biggest risks and hurdles that are non-AI-related, whether they're regulatory? Or competitively what are the sort of the things that are realistic obstacles that you still need to overcome?

I think a piece of it is -- an important piece is the -- creating that societal pull for this to occur. And what I mean by that is -- and we've started to do this in some of the recent communication work we've been doing is to really highlight for people, how's transportation working for you today, right? What are the pain points? What is the objective evaluation of current road transport in terms of safety and cost and environmental impact and congestion and all of these things and to force people to look at that and say, "Well, that's not as great as it could be and then here's this technology deployed in the right way that could make a really positive impact on that and make my life better," and create that greater acceptance and pull for this to occur. Because I think the -- for the average person on the street today, there's still a view that's self-driving, and you said it in your opening remarks, is this thing that's kind of far out there and maybe it's a science project and some technology companies are working on it, but it doesn't have any real direct relevance to me. And I think we need to advance the narrative to the point where people say, "Man, my commute sucks. And imagine how much better it could be if this technology was deployed at scale." And I think we have an obligation as industry leaders to tell that story and to build that sort of societal pull for this to occur.

Okay. And just lastly, let's -- I asked the question about how far you think the penetration is going to go in 2025, 2030? And the answer is, it depends. When you think about what the competitive landscape looks like, if you're successful in the next year or 2 and you pedal to the metal, you're expanding globally, how many competitors realistically do you see in this kind of a business? Is it sort of a natural monopoly or oligopoly in different cities? You see a few different companies that are -- they're going to be competing against you realistically.

I think, unfortunately, the answer is sort of the same to my earlier question, which is it depends how successful we and others are at driving this newest rate of improvement. And if you -- to get to that point where one thing is awesome and one thing is just okay, and the thing that's awesome gets used more and just keeps getting better, then that's going to be a pretty -- and if you were that company, that's going to be a pretty interesting place to be. I mean when you -- to go back to your earlier question on sort of how big can this be, if you look at the sort of rideshare companies, I think Uber's gross bookings, I think, got to $50 billion in the space of 7 or 8 years. So the -- if you build a product that is better than what people have today, you can build a very big business in a pretty short period of time.

Great. I think we're out of time. With that, I want to thank Dan Ammann and Cruise for this fireside chat.

Thank you.