Okay, great. Thank you, everybody, for joining us. My name is Mark Delaney, and I lead coverage of the U.S. Autos and Industrial Tech sector at Goldman Sachs. I'm very pleased to have with us Kyle Vogt, the Co-founder and CEO of Cruise. Thanks for joining us yet again this year.
Thanks, Mark. It's good to be here. I know Mary Barra was going to be here today, but there's some other things going on. In particular, she's working with her team on UAW activities, very important. But if she was here, I would say, you know, thank you for the continued commitment to Cruise, especially during our period of rapid scaling and as we work to bring the Origin to market. That's really exciting. And before we get started, I also wanna just point out Mike Abbott, sitting right over here. He's GM's new EVP of Software, joined from Apple, where he ran engineering for all of their cloud infrastructure and products. So pretty exciting to have him at GM. Glad to have you here. Off to a great start. Yeah.
Yeah, looking forward to talking about Cruise, the amazing progress we've had and our collaboration with GM.
Yeah. Well, it's really exciting times in the AV industry, and a lot of stuff going on at Cruise, so I've been looking forward to this. You know, Cruise cars are now operating 24/7 in San Francisco. You know, started off only in the evenings. I think, you know, last year some of us were fortunate enough to do some evening rides, and now it's around the clock in San Francisco.
Mm.
Also operations in Phoenix, Austin, Dallas, and Houston, so, so quite a bit of progress over the last 12 months. You know, I know, as you mentioned, Mary wasn't able to make it today, but maybe you can talk a little bit about why GM is so bullish on autonomous vehicles when some of the other companies in the industry have scaled back.
Well, I think, to be bullish on this, you have to have a leadership position, right? So Cruise is the largest in terms of fleet size, scaling the most quickly, and we have line of sight to really good, you know, economics. You know, there's cost per mile, which is really critical for this to be a successful and profitable business. But if you take a step back and look at this, you know, autonomous vehicles are something where customers absolutely love it, and we can talk more about that in a little bit. Just the demand is off the charts.
Yeah.
The product is better than the status quo. It's safer, and it's more comfortable for people. They prefer it. It has a better cost structure, and it's a multi-trillion-dollar market. So if you're sitting on that thing in a leadership position, I think you should be doing everything you can to establish a defensible position in this massive market, and that's exactly what GM's doing.
Let's talk a bit more on San Francisco. You know, big milestone in the AV industry, being able to operate 24/7, but also be able to charge for rides and be hailed on a smartphone, you know, somewhat akin to calling an Uber or a Lyft. Talk a little bit more around how the Cruise fleet has been performing, and what's feedback been like from your customers?
It's incredible. Like you said, in the last year we were operating in a corner of the city late at night. Now it's a fully operating service, and people love it. You know, we hear, I heard just yesterday, someone told me that they had that sort of interaction with a consumer product for the first time, that was magical. It was like their first time using the iPhone, and they actually compared it to their first time using Uber when it first came out, I don't know, 14 years ago, and it was such a jump, you know, over the status quo.
But, you know, for context, in, you know, in our fleet now, we're driving over 1 million miles a month, so this is reaching, you know, significant scale. In San Francisco, tens of thousands of residents have tried Cruise, so this is not a science experiment. It's an operating service. We have over 100,000 five-star reviews at this point, so there's clear evidence this is resonating with people. And, you know, if you dig a little deeper, one of the things that matters most for a consumer product is the stickiness and the retention, that people come back and use it after they try it for the first time.
And what we're seeing with the Cruise fleet is retention numbers in our second year of operation that are matching what Uber and Lyft achieved after 14 years of operation. So the stickiness of this product is really, really good, even in the early days, and this is, you know, just-
Mm.
Just getting started. And of course, other metrics like NPS, world-class, like, people perceive this as, as a premium product.
That's very interesting and good to hear. Maybe we can talk a little bit more on San Francisco. Have Cruise's expansion plans been affected by the DMV's request that Cruise reduce its driverless fleet temporarily in the city?
No, but we can talk about that. So, you know, having the largest fleet, driving more miles than anyone else, this is sort of simple logic that if there are rare events that can happen on the road, we're probably gonna see them first. And that's what has happened here. So anytime... Well, I guess in this case, if you're a regulator, and the first time there's a novel event, a certain type of collision, I think it's healthy, you know, public discourse and regulatory scrutiny for them to want to investigate that event and understand it. And so we're working closely with them, you know, as they investigate that, and I'm sure when they see all the data, you know, they'll have more comfort with how this technology is performing.
But I guess taking a step back, we're at a unique moment in time where anything an AV does, it, you know, even if it's just awkward or something kind of interesting or like, "Oh, that's, you know, I didn't think of that," maybe a human wouldn't do it exactly that way, it becomes a national headline. You know, like an AV came into some contact with some wet concrete in a poorly marked construction zone, and that was a national headline. Meanwhile, just a couple weeks ago in San Francisco, there was a four-year-old girl who was tragically run over by a driver on Fourth and King Street, just you know, a few blocks from here. Didn't even make the news. It's devastating to me, I have a five-year-old son that we're not talking about that.
There's a massive double standard here between the attention and sensationalism around AVs versus, you know, the tragic status quo on our roads that we've normalized. And, you know, we're sitting here looking at this data, showing that AVs are, they get in significantly fewer crashes than human drivers, almost a 75% reduction in the types of crashes that can cause injuries, so we should be celebrating that. And so if we don't keep the focus on safety, road safety, I worry that we're gonna set society back, you know, a decade. And when it comes to road safety, that's just something we can't do.
Yeah, no, this is so impactful for many, many people. Let's talk on safety a bit more if we could. I mean, you commented a number of times around how Cruise's safety performance is actually better than human drivers, and especially when it comes to things like emergency vehicles or even atypical road scenarios. Can you elaborate a bit more on that?
Yeah. So, you know, it's a tough thing to do, to figure out. To do a comparison, you have to figure out how good humans are at driving. And intuitively, we probably all think we're good drivers, but it's the times that we're distracted, drunk, or drowsy, that you have problems, and that actually happens quite frequently. So we did this study to figure out exactly how good human drivers are, not just by looking at the collisions that are reported to insurance companies or to the federal government, because that's a tiny fraction of what actually happens. You know, people don't-
Yeah
... want to report crashes. So we collected this data set by instrumenting human-driven vehicles and seeing how often they actually get into collisions in San Francisco. And then we compared it to our driverless AVs across the first 1 million miles, and they're what you'd expect. You know, these are, these are cars that don't get, drowsy, distracted, or drunk. Half as many collisions, 75% reduction in the more severe collisions that could cause injury. And out of the collisions we had, 91% of them, the AV was actually not the one at fault. It was the other party, the, the human driver, who actually, was the primary contributor to those collisions. So that was just our first 1 million miles, right?
Now we've driven almost 5 million, and we're seeing that even as we increase complexity, like moving to daytime driving, 24/7 operation, those safety benefits are persisting. So this is massive. This is not a, like, a 10% gain over the status quo or something like that. This is, this is game changing. This, this is a massive improvement in safety, and this is, you know ... And, and AVs continue to get better month-over-month. So, you know, a couple- if we sit down and do this conversation next year or the year after, we might be looking at, like, a 10X improvement over humans, and I think that's, that's just really astounding.
Yeah. Well, we look forward to seeing how that progresses going forward. You're in five markets now, operating, testing in 10 others. When you started out at Cruise, you had the thesis that developing the AV technology in a dense environment like San Francisco, where there's so much complexity, that learning in that sort of a scenario would allow you to scale to other geographies relatively quickly. How is that thesis playing out?
Yeah, that's the question. Another important question is: how do you take this technology, and knowing that it's maturing over time, how do you take that and apply it to commercial markets? In what order, what sequencing? The path we chose was go after a commercial market like San Francisco, which is a billion-dollar ride-hail market, make it work here, knowing that if it works here, it's gonna work in many other markets. That was the theory. If instead you start with, like, sort of a tier two or three market and then try to, like, grow into those bigger markets, you might run the risk that the technology is not just not ready or you have to redo it, which is much harder to do, you know, when you have a service operating.
So anyways, that seems to be working, because what happened is, 2 years ago, we were in 1 city, just San Francisco. End of last year, we were in 3 markets. Now we're in 5 and ramping up operations in 10, so you can kind of predict where we'll be around the end of the year. And the reason that's possible is the core technology, the AV system, you know, the sensors, the compute system, the AI that drives it, has proven to be generalizable, meaning, you know, we build it to work in San Francisco, when we go to a new city, like in Phoenix or Austin, there are some minor differences, but we have the AVs sort of scout and harvest those differences, retrain the neural networks, and now they've become more adaptable and more generalizable.
After you do that a few times, by the time we're in, you know, five markets and got 10 more spun up, the incremental differences that we see in city number six or seven are tiny or almost nonexistent. The question becomes: what's your playbook? Can you lay, lay the infrastructure on the ground? There is some mapping and surveying. Can we do that cost-effectively and quickly, and then, obviously engage with the community so they're ready for us when, when we choose to scale there? You're seeing that happen. You're seeing that happen: one, three, 15. You know, we're on a, on a good trajectory there.
Well, let's talk about the vehicles. Right now, the fleet is comprised of Chevy Bolts. The next vehicle that the company has announced is the Cruise Origin. Talk a little bit more, if you could, please, about the timeline for deployment and perhaps why the Origin could be more cost efficient than the Chevy Bolts that are on the road today.
Okay. So yeah, timeline and cost. So, to begin with, the Origin, it's our new vehicle, and it's the first time an OEM, a car maker, General Motors in this case, has made a car with no steering wheel. And that's just not- it's not just, like, deleting the steering wheel, it's reimagining a car that would be optimal for ride-hailing. So you have bench seats that face each other, tons of legroom. They got a great sound system, lots of cool lighting. It can be a party on wheels if you want it to be that, or it can be your zen oasis, you know, between meetings or on your way to work, where you can just focus and get things done.
It's really a shape-shifter in what it can do, and it's far better than being scrunched into the back seat of someone else's, you know, sedan. So amazing vehicle. In terms of timeline, we have completed the engineering work on this vehicle. We have worked closely with GM to take the, you know, advanced AV software and AI that we have on the streets now, merge it with this new vehicle form factor, and integrate it all together. We've tested and validated it. We actually have run the Origin on public roads, both in Austin and San Francisco. We're testing it, and we are, from what we've heard from NHTSA, just days away from the last regulatory approval, which would let us start production and almost immediately start putting these vehicles on the road.
Okay.
And then lastly, you asked about cost. I do wanna touch on that. This is a situation where we can have our cake and eat it, too. We're actually building a fundamentally better product that's gonna amplify all those usage metrics I spoke about before, like retention of the product, the five-star reviews. All that's gonna get better in this vehicle, but it also costs us less to build, and that's because we've taken a rev on all the sensors, the compute systems, even the software is slightly simplified to lower the upfront cost of the vehicle. And then working closely with GM, we've done a lot of work to bring down or to increase the lifespan of this vehicle.
So it'll last, you know, maybe an average car, 150-200,000 miles, something in that range. The Origin is designed to last 1 million. So you can build, like, 1 Origin for every 5 cars you'd have to burn through in 1 million miles. And so you put that long lifetime, lower upfront cost together, that's a dramatic reduction in the cost per mile to operate these vehicles, which is a key unlock for profitability.
Very exciting. It is a unique vehicle and perhaps a little different than things GM has built before. So talk a little bit more about how quickly you think GM may be able to produce them and get them on the roads.
Well, I mean, that's, that's the exciting thing here. Like, no one else in the AV industry is actually building, like, having an OEM build AVs in the factory, right? So this is a ton of work to figure out how to actually do that, like, how to work through the supply base that's not used to building this kind of technology. Even small things, like in an assembly plant, at the very end of the line, there's an expectation that someone can hop in the car and drive it out to the parking lot, so it can be put on a truck and shipped somewhere. You can't do that if there's no steering wheel, you know? So we had to build new mechanisms to move vehicles through the plant, and there's joystick systems and other things to make that happen.
But the net result is, once we go into production, we can scale very, very quickly and get OEM-level scale in this AV business, which will be the first time that's done, and I think, frankly, GM is, and, and Cruise are years ahead of the competition when it comes to mass producing AVs at scale.
In the past, you've laid out some pretty ambitious cost targets about where Cruise can get to and eventually reaching $1 in cost per mile. Is rolling out the Origin and getting scale enough to get to that point, or are other things going to be needed to eventually reach that kind of a cost target?
Yeah. So $1 per mile is a nice, nice round number, fun to talk about. The reason $1 per mile matters is if you look at what an average American spends on their car, if you look at the cost of the vehicle, insurance, everything like that, on average, about half of the car owners in the U.S., actually, their cost is around $1 per mile or more. So if you can get below a $1 per mile, you have this massive opportunity of people who are gonna say, like, "Hey, maybe I should actually start using robotaxis to get around instead of owning my own car." And so ride-hail is already a big market.
I think it's gonna explode once you hit that inflection point in cost per mile. And in terms of our ability to actually hit that, we have really good line of sight now. Last year, you know, we had some plans. Where we are now, the hardware... So the Origin is about to come out. A couple of years after that, another version comes out that uses our custom chips and silicon we've designed in-house for the first time. So this takes a ton of cost and complexity out of the sensors and compute system. That technology is sourced.
It's going through the supply base, so this is these we have a great deal of precision around both the final cost of this and the timing, which means that in 2025, we the hardware we build then will be capable of reaching those unit economics. So that's that's not that far out from this massive inflection point in in cost structure.
Well, let's stick on the topic of economics. You have a target of $1 billion in revenue by 2025. Talk a little bit more on how you build up to those sorts of economics. How many vehicles should we think about operating? What would you charge, and so forth?
Yeah, so $1 billion you know, in revenue, it's a very, another nice round number, but we're on track for that. And, you know, you're seeing it this year in the evidence that we can scale. We've gone, we've increased our, you know, number of miles over 10x since this time last year, the peak size of the fleet over 10x, since last year. And so we're building some confidence among investors and others that we are on a glide path to achieve the necessary scale to hit $1 billion. And so in terms of AVs and pricing, you know, look, all the ingredients are there. We're talking about OEM scale, so it takes in the thousands of vehicles, to get to, you know, that, that order of magnitude of revenue.
And for an assembly plant like GM, that's nothing. That's, that's nothing. And so it's really about whether we can lay out the infrastructure and, and ramp up the fleet to, to meet that size. And we've done a substantial amount of de-risking to get there just this last year, so, on track for the revenue. Pricing is an interesting thing. You know, on one hand, what we're, what we're hearing and seeing from consumers is evidence that, you know, it's, it's the way they would treat a premium product, right? Like, off-the-charts retention, five-star reviews, vast preference towards the status quo. Typically, when something like that happens, you charge more for something.
On the other hand, the cost structure is, you know, on that timeline, is gonna be such that it costs less than the very expensive process of having a human drive a car, like an Uber or Lyft ride. So you have this, like, premium product pricing potential, very low cost structure.... That huge amount of flexibility between those two, is a huge asset for us because it lets us decide if we want to pull forward profitability, or if we wanna, capture and defend market share in what we believe is a multi-trillion dollar, market. So huge opportunity there, and I think we're gonna be well-positioned and have lots of flexibility to choose how we go about that.
Okay. Well, it'll be interesting for us all to see how that progresses. I do wanna talk about scaling, and it's something that I speak about quite a bit with investors, and so far you've been vertically integrated, right? And you've talked about the close collaboration you have with GM and Cruise. Are there other partners you would consider working with, you know, perhaps fleet partners who would own the vehicles, you know, integrating with other rideshare apps, which perhaps could, you know, help you to scale faster or more cost-efficiently in terms of customer acquisition costs? So talk a little bit more about your willingness to partner, if at all.
Sure, yeah. I heard one of those folks was just on the stage. I saw him walking out. So, Our philosophy is simple. It's customer experience first. So what that means in ride-hail is, today, we are benefiting greatly and delivering a great customer experience by having this tight vertical integration. From the moment you decide you want to go somewhere to when you when the vehicle is pulling up to you, and you use your phone to unlock it and get in, to when you get inside the car and interact with the in-car screens and experience, to when you get out, that's all one cohesive set of apps and flows that people are really resonating with. And so right now, we think that is the best possible experience to have in a ride-hail application.
But of course, the other thing we know is that this industry is evolving really rapidly, so it's hard to rule out any future partnerships. I just know today that's, that's what our customers want. When you look at the delivery business, we also have a really healthy delivery business. We partnered with Walmart, you know, largest retailer in the world, and they. In that business, the best customer experience is actually integrated into, you know, for example, Walmart's existing checkout flow. As a customer, you just, when you buy something from the app or whatever, you wanna be able to select AV Delivery and not have to think about a separate app.
So we're doing that with Walmart 'cause that's the best thing for the customer, doing our own app 'cause that's the best thing for the customer for Ridehail, as of today. And then you also asked about, fleet infrastructure, and partnerships there. Very interested in that. We do partner when it comes to things like, charging infrastructure, parking our vehicles, cleaning and maintaining them. Very open to partnerships there. That's not something that it necessarily needs to be our core competency, and we prefer to keep the business, you know, you know, capital light or CapEx light when possible, and so partnerships are very attractive there.
That’s all very interesting. At CES in 2022, Mary spoke about GM and Cruise’s collaboration and the economies of scale that it created. How is that process going?
I mean, it's been awesome. It's been really cool to see, not just GM, but GM, Honda, and Cruise come together to build a vehicle like the Origin. First of its kind, totally different, you know, backgrounds and, you know, sort of loci of genius between these different companies, like, each brings something to the table. A lot of hard work and effort to get to integrate those things effectively, but the result is pretty amazing. And like I said, like I. You know, this is some of the behind-the-scenes work that hasn't really come to light yet because we haven't put the Origin out on public roads yet. But mass producing a driverless vehicle to be used in robotaxi applications is going to create a huge, like, a massive business opportunity for GM.
GM's the only company in an industry positioned to do this because of their, you know, early commitment in Cruise and investment, not just in taking the technology from R&D and into early commercialization, but to actually simultaneously develop a production program and have the foresight and risk-taking nature to actually spin up years in advance. Like, we've been working on this program for 5 years. All the CapEx and other investment required to build this vehicle in a production plant, that's not a small activity. And to have made that bet at the right moment in time when, you know, there are promising signs, but it wasn't proven yet, so that now we'd be in this amazing position to actually capitalize on that, whereas others are just getting started spinning up these activities.
It's really amazing to see this collaboration, and as you said, economies of scale kick in.
And one of the other opportunities that's been discussed has been the potential for AV capabilities on consumer cars. I mean, maybe talk about the technology and how portable it could be in the longer term onto passenger cars.
Well, I think that's an attractive concept for obvious reasons. Like, if you've got clear safety benefits by taking a human-driven mile and converting it to an autonomous or driverless mile, that's a good thing, and we should want to amplify that, you know, for the benefit of society as much as we possibly can and as quickly as we can. So we've kind of alluded to this in the past, and you know, we've definitely been thinking more about how we can collaborate with GM on that front and laying some of the foundational technologies that would enable something like that, as an example, lower cost sensing and compute.
I didn't really mention this before, but one of the key constraints to our business today is this notion of a geofence, where the cars only operate in a certain part of a city. We've been, I guess you could say, solving that in two different ways. One is we're making it, you know, to support the robotaxi business, we've been making it less costly and more efficient and faster to map in that way, and so going from tens of sq mi to hundreds and thousands pretty quickly. We're also working at it from the other direction, which is building a new technology where AVs no longer rely on these complicated and somewhat expensive maps...
And instead consume readily available map data and rely more on their sensors to kind of create a map of the environment as they go. And so that's gonna enable us to do a lot on the, on the robotaxi business in terms of expansion, but does also open the door for, you know, that kind of thing in the future.
I'm gonna give the audience an opportunity to ask some questions as well, and we're fortunate to have Kyle here. But one thing you said on the last earnings call that I've been looking forward to following up on was a comment you made about something the Cruise team has been working on for the last couple of years that you think will be highly disruptive, and perhaps you'd be able to share more on what you were talking about on the last earnings call.
Do you want to know what it is?
I do.
Too bad. But we'll wait until GM's gonna do a big investor day in November, and we'll talk more about it then.
Okay. Well, the cliffhanger is,
It continues
... is gonna be ongoing. Well, we've got a jam-packed audience, and we're fortunate to have Kyle here, talking in so much depth. So if you have a question, put your hand up, and we'll get a mic over. Yeah, I think we got one in the back to start.
Yeah. Thank you for the detailed introduction. As you mentioned about the inflection point in 2025, like, we could be break even time. Just, can you provide more color on that side? It's kind of how many years? Because we have the depreciation period of our car and the revenue accumulate in those period, right? Just more colors on this inflection point. Thank you very much.
So we'll go into a little more detail on the business model in November. But what I would say is the things that you know get to top-line revenue, like that $1 billion, are you know lower cost of the vehicle, which includes a longer lifetime higher utilization of the fleet, so it's gonna be operating in basically more areas more of the time. So basically, in a 24-hour period, it's operating more hours and serving more rides. And the last is volume. But starting in 2025, the hardware produced then is capable of those very attractive unit economics, which you know depending on pricing, could be you know unit profitable at that point. But at that point, we'll start producing that vehicle.
So we'll have a mix of vehicles we have now that are slightly more expensive and a mix of lower-cost vehicles, and so it'll take some time for that fleet to migrate to the lower, lower cost structure before you see, you know, things flow through to the profit side. But I think, you know, taking a step back, still looking at what that business looks like at that point in time, you know, the revenue growth, if you know, if we're successful there, and we're on track to do that, will be extremely good, you know, like, record-breaking for a consumer products consumer product like this. And I think, you know, the pathways to future scale and the risks associated with that will be significantly reduced.
I think it'll be a really good business, at that point in time, regardless of how aggressive we choose to be towards profitability.
I think there's a hand up there. Yeah, in the front. Oh, sorry. Yeah, maybe start over there on the side.
Thank you. A very helpful discussion. Could you tell us a little bit from a regulatory point of view, where we are and where we're going in the context of the products you talked about, in terms of the government rating system, the level one, two, three, level five? My understanding was we were at L2 for the Teslas of the world, et cetera. In the context of what you're doing, what is the range that's going from here to there?
Good question. So I think there's two pieces. I'll start with the second one on the SAE levels. Frankly, I don't care about those. The distinction to me is when you can get in a car, and there's no driver in it, you can sit in the back seat, fall asleep, do work, whatever, or whether you have to do hover hands and kind of monitor it. We're in the former camp. These are true driverless cars. You're never asked to do any part of the driving. You're not even allowed to. And that's a huge difference in terms of value to a customer. Like, you know, you can if you've had a couple extra drinks, you know, you're out and about and you want to go home, you can use a robotaxi.
You can't do that in a driver assistance system. You know, if you have, like, an elderly parent that's not capable of driving theirselves, they can still go in a robotaxi. They can't do it in a driver assistance system. Huge difference between those two. And I think we're seeing that in the customer engagement numbers and retention, other things. This is a really sticky product. People have been wanting it for a long time. It's finally here. On the regulatory side, there's 2 buckets, sort of at the federal level, there's NHTSA, who regulates safety in the U.S. We have, there's no explicit approval required to do the Bolt AVs we have on the road now.
The Origin does require a special approval because it doesn't have a steering wheel, so it's really hard to be in compliance with all the safety regulations around steering wheels if you don't have one. So we've asked for an exception or an exemption from this, and that's what we're expecting news on any day, which would be an endorsement of our approach to safety for that vehicle. At the state level, in California, we've gone through 10 permit applications with various regulators, the DMV and the California Public Utilities Commission. We are now cleared to operate commercially in all of San Francisco, 24/7, and collect fares. So that was a big win for us. First in the industry, we're one of only 2 companies who can do that at all. And then outside California, it's a very different story.
It took us three years to get all the necessary permits to commercialize in California. It took us three days in Texas, for example. And there are many other states that are similar in terms of arms wide open and, in fact, calling us off the hook, asking when we can be in their city next. So the regulatory situation looks, looks very good. You know, that's something we've been working on for years, and we don't expect that to be a impediment in the business going forward.
Yeah. Oh, yeah, right in front.
A two-part question. What do you think about international expansion? I mean, basically, will you be in 50 American cities before you will be here, before you will be in one international city? And also, how difficult for you do you think it's gonna be to go international, talking about, like, other road signs and stuff?
Another great question. So right now we're testing in or operating in 15 cities, five without drivers, and testing in 10 others. Two of those are international cities, so we have Dubai and in Japan, a smaller city in Japan. And the reason we're doing that is, again, to ensure that as we develop the technology, it is generalizable and works in those markets. And so in Japan, in particular, the AVs are driving on the other side of the road. And so the big question for us is, can we take a whole bunch of training data from Japan, add it to these ML models, and get that correct behavior to emerge? So far, the results are very promising, even though it's a substantially different type of driving.
We do obviously have to adapt to, like, you know, street, different street signs and other things there, but that's relatively straightforward compared to some of these other tasks. In terms of commercialization, you know, we are interested in engaging in Robotaxi business in Japan for sure. Like, there are, like, over a dozen markets larger than Manhattan in Japan, to put that into perspective. So very attractive business opportunity, and that's why we're testing there now in partnership with Honda. And then in Dubai, similar opportunity, and, you know, that could be a nice springboard for us into other parts of Europe, especially to make progress on the regulatory front. So we're preparing for that, but our focus in the near term is definitely scaling up commercial operations in the US.
We've got time for one more question, either from the audience or else, I'll go ahead and ask it. Yeah, right over there, please.
Thank you. Somewhat related to an earlier question, could you talk a little bit about how does the cost per mile compare today relative to Uber, taxi, other public transport, and where do you see it go?
So right now, I guess we talked about a dollar per mile as being like a, a, a target for Cruise. I think in San Francisco, I don't have the latest numbers, but call it $3-$4 a mile, so pretty big difference. Of that $3-$4 a mile, call it 75%-80% goes to the driver, and so, you know, the rest is overhead for these ride-hail companies. That's a huge opportunity if you, once you sort of take that driver portion that you have to pay out, out of the equation, the cost falls through the floor, and what you're adding back is whatever the cost, the amortized cost of the AV technology is back on top of that.
And so it takes a substantial investment and significant scale to do that. But because the cost of a human wage and a human driver is high and only going to get higher over time, you know, there's a really big opportunity to create a fundamental advantage on cost. And that's before we get into any of the just straight customer preferences for AVs. The fact that the driving is consistent and smooth, you get the same thing every time. You don't have a driver that you don't know in the car. It's not someone else's space.
And lastly, we hear, you know, from some of our customers who are women, that they just don't want to get into a car with a stranger that they don't know, especially if it's late at night or something. So we will have this fundamental cost advantage just by nature of the fact there is no human driver necessary, and then coupled with a really compelling product.
Great. Well, unfortunately, we have run out of time. Kyle, thanks so much for coming.
Thank you. Thanks for having me.