Uber Technologies, Inc. ($UBER)

Okay. Good afternoon, everyone. Thank you for joining. My name is Nikhil Devnani. I'm Bernstein's U.S. emerging Internet analyst. I cover a range of marketplace businesses for us here at Bernstein, including Uber. It's my great pleasure today to have with me Balaji Krishnamurthy, Uber's newly appointed CFO. Balaji, welcome to the SDC.

Thank you.

Before we get started, I want to remind everyone to please refer to the Investor Relations website from Uber for any disclosures. And if you would like to submit questions for this session, you can do so via the QR codes available to you. Now Uber needs no introduction. I'm sure everyone in this room is a consumer, but just to set the stage a little bit, the business is truly a global scaled platform now doing about $190 billion plus in gross bookings volume, supporting over 200 million monthly active customers. And there's a lot that I want to get into regarding the core business very soon. But I do feel it's most topical to start with the news that hit over the weekend regarding Uber's reported interest in Delivery Hero. Now I know Balaji, there's only so much you can really talk about at this point on this issue. But we've looked at this deal from our point of view as well. Strategically, it's interesting. It's also a relatively large deal. So I would just love your perspective and some commentary on why this might be the right use of capital for Uber really at this point in time because I think there's a lot of demands on the business today. There's organic investments you're making. There's AV investments you're making, you have a capital return program you can lean into as well. So why would a combination like this make sense for Uber today?

Great. So maybe I'll start by talking about our approach to capital allocation, our M&A philosophy, and then I'll touch on Delivery Hero specifically to the extent that I can comment on it. So if you think about the approach to capital allocation that we have taken, it's been consistent, and I'll just remind everyone how we think about this. The first priority for us is reinvestment into the core organic efforts that we have. And we've been in a very, very fortunate position that with the scale that Nikhil just alluded to, we've still been able to find a number of growth opportunities where the returns for us have been very attractive. And just to put a finer point on it, over the last 3 years, our gross bookings have grown 1.7x. and our free cash flow has 10x in that period of time. And so it's just really a proof point for the sort of reinvestment opportunities we have seen, and we continue to see those in front of us. So that will be the first priority. Once we've done that, we still sit with about $10 billion of annual free cash flows that we're generating and then we think about what -- how we reallocate it in the best way. The first priority for that reallocation is our investments in supporting our EV strategy, right? We are going to invest to future-proof ourselves against the opportunity that's in front of us. We think this is going to be a massive TAM expansion for Uber, and we're making appropriate investments against that. Then we think about M&A, and we will look at selective M&A, and I'll spend a minute on the sort of focus areas for us and what kind of parameters we look at when we think about assets sale. And then finally, once we've reinvested for organic and inorganic efforts that meet our ROI hurdles, we then have excess capital that we are returning to shareholders through our buyback program, and we'll continue to do so, right? So maybe let me go back to M&A and talk about the M&A philosophy that we employ. For M&A, when we think about assets, the first bar they have to meet is that they have to fit with the core strategy that we have and it needs to be an asset that can accelerate our organic efforts complements us in some way that we don't have capabilities. Then as we think about what are the core strategic objectives that an asset needs to meet, we are looking for assets that either help us expand our TAM, broadens out our platforms, distribution capabilities or brings us new capabilities from a technical standpoint that we may not have in-house. And then we are going to hold a pretty high bar for any acquisitions that we are undertaking. And I think it's instructive to spend a minute on thinking about how we define that high bar internally. For us, the bar that we are thinking through is, does this asset meet the strategic objectives we are talking about, and it does it meet multiple strategic objectives because that becomes easier for us to underwrite. The second piece that we are thinking through is, do we have a clear plan from offer to regulatory approval to integration? And on the other side of this, we should have high confidence that we can drive a lot of accretion on an adjusted EPS basis because the alternative is that we can buy back our own stock, right? And we have a lot of visibility into our own business, and we can do that with high conviction. So that means that we should be able to say, if we make an offer, do we know what the antitrust regulatory approval process looks like, do we need shareholder approvals? Do we have a path to getting that? And then do we have a clear integration plan that we can put on paper today or through the process and then go execute once the deal closes. So all of that, if they are met, then the last check that we have to do also is, is this asset available at a reasonable price? Again, we don't want to overpay for acquisitions, even if it meets all the other criteria for us, right? So that's kind of the construct with which we operate. So let me come back to Delivery Hero with all that backdrop in mind. So what we have publicly announced is as of yesterday, we have exposure to about 37% of Delivery Hero. And what Delivery Hero has also publicly said is that they're actively reviewing strategic options for the company, right? And with the position we now have, we have a seat at the table for the outcomes that may emerge from the strategic review of Delivery Hero. I think what we have found interesting about Delivery Hero in the past, and we've said some of this in the past is they do have some assets, some markets that are quite attractive where they have executed really well. They have a strong position, they're quite profitable in those markets. And markets that can be quite complementary to our Mobility presence as well. So if you think about the Middle East, for example, where we have a very strong Mobility presence, we don't have a Delivery presence and that can be an interesting combination for us, right? Similarly, they have very strong organic positions in markets like Argentina, Korea, et cetera, that could be interesting as well. Then as you look at the remainder of the sort of framework I laid out, we know that given our global platform, scale and advantages on that front as well as the global tech stack, which is best-in-class that we can replicate for any assets we acquire, we feel pretty confident that the accretion opportunities should we undertake any kind of transaction would be quite meaningful and that is attractive to us. And then finally, as I said, it needs to be at a reasonable price. So for anything to happen, it would have to be at a seller expectation that's reasonable. We have been quite disciplined in building out that 37% exposure that I just talked about, our cost base is in the low 30s, and that's basically what I can say at this point.

And when you just step back and think about the Delivery business globally and really the international portfolio, I think there's a lot of investor comfort and understanding with the U.S. market. It feels like a more stable market. It's getting more profitable. I think international is more of a black box folks because there's less data, there's more competition or at least perceived competition. And so there's a question always about the quality of international growth and exposure versus domestic. When you look at the global lay of the land, how do you think about that difference?

Yes. I would say our Global Delivery business is quite underappreciated. First, we have, through the last few years, built out very attractive positions in some marquee international markets, right? We have a market-leading position in Canada, U.K., France, Australia, Taiwan and Japan, among others. These are all very attractive markets with a lot of untapped TAM still in front of them. And we are continuing to be -- we have that leading position with a pretty high margin output in these markets. Now the interesting thing is even if you look at a market like Australia, which is amongst the most penetrated Delivery markets in the world, where we have a strong #1 position, we're still growing upwards of 30% year-on-year in that market with very high margins. And really, the opportunity that we have discovered is that we continue to expand in these parcel markets, which continue to be quite underpenetrated everywhere. And then on top of that, our grocery and retail efforts are kicking into high gear as well. If you look at markets like Canada, U.K., Japan, all of these markets are growing at roughly 20% or higher and a market like Japan, for instance, has accelerated into that 20% plus growth territory that we're talking about. So quite attractive footprint that we have, and it complements our Mobility position in all of these markets very nicely.

Let's take a step back and talk about Uber in aggregate. A few years ago, you gave an Investor Day framework that pointed to mid- to high teens gross bookings growth on a compounded annual basis. You've been exceeding that. You talked about high [ 30s ] to 40% EBITDA growth, and you're well on track there. As you look forward to the next 3 years for you as a new CFO in the seat, right, what are your strategic priorities? And what does that financial framework look like for Uber going forward?

Yes. So maybe just stepping back. My previous role before taking on this role was running the P&L for both Mobility and Delivery. So -- if you think about the strategy that we've been executing, I have been involved with that through the last few years. And with that sort of a lens, I would think about the path forward as a continued refinement of what we're doing versus something that we need to change, right? And as I said, we are in a good position with that healthy growth with continued operating margin leverage that we have been demonstrating, and we want to execute on that kind of footing going forward as well. To just respond to the sort of 3-year framework that Nikhil just reminded everyone of, for Q1, if you looked at it on a 3-year trailing basis, our top line compounded at about 20% CAGR relative to that mid- to high-teens CAGR that we had laid out for 3 years. Our profits compounded at about 50% CAGR in that period of time relative to the high 30s to 40% CAGR we were talking about. And then the free cash flow conversion was north of 100%. So a very, very healthy position that we sit in as of Q1. And what we are looking for and our outlook for Q2 again was 18% to 22% growth for gross bookings. We are showing continued profit margin expansion against that. And we just want to keep doing more of that. And the goals here are unchanged: invest behind the platform, invest in product innovation, drive operating leverage through disciplined cost management. And then if we are doing that blocking and tackling well, we remain in a good spot to sustain that mid- to high teens kind of top line trend line.

The cross-platform push for the last few periods, I think has been clear to see, and it's all getting tied together now with Uber One. It is somewhat surprising to me, though, that you still only have about 20% of your Mobility and Delivery audience that's actively engaged and eligible to engage across both apps actually doing that. So why do you think that number isn't higher? And what do you think needs to happen to get it to go higher?

Yes. So it's a great question. I think the backdrop for that, the first thing you have to ensure when you're building these kind of businesses that fit together but are distinct categories on their own is that you have to build businesses that are individually really strong in every market that they're operating in, right? So there's a lot of base level work you have to do to ensure that your Delivery offering is as good as your Mobility offering because otherwise, you're cross-selling into a service, that could be inferior to what you have on the other side. And our first -- and vice versa, right, where Delivery may be stronger, Mobility has to be as good. So our first objective over the last few years, as we've thought about our platform is to ensure that we have best-in-class individual services, right? So when you think about that 20% overlap that you just articulated, these cross-platform consumers are quite valuable for us, but I would think about that as largely an output of what we have been doing rather than something that we have driven up in a concerted way, which is something that we have started doing much more actively over the last year to 2 years, right? So meaning that we built really good businesses and naturally consumers converged on to both, and that is roughly a good portion of the 20% overlap that you're seeing there. And now we are building out the motion to actively say, how do we convert this consumer from Mobility to Delivery or vice versa? How do we ensure that the value to the consumers is higher if they're engaging on both sides of the platform? And do we target our consumer acquisition efforts in a different way? So what we have started doing is if you look at our app now, we are starting to roll out our universal search feature, One Search, which effectively says if you're on our Mobility app and you search for a restaurant you may be wanting to go to that restaurant or we can show you that you can also get a delivery from this restaurant in the future. Or if you're searching for a burger on the Eats app maybe you want to go to a burger -- hamburger store, right? So those kind of overlaps are beginning to happen for us. Then the other piece that you have -- we've actively been doing is our Uber One push, where roughly 2/3 of our Delivery gross bookings are coming from Uber One and only 1/3 of our Mobility gross bookings today is coming from Uber One. And roughly 50% of our overall gross bookings is Uber One. That convergence for Mobility as it increases it starts driving increasing overlap of consumers as well. And then I think when we look at the best-in-class markets where we have pushed higher on this metric, we're over 25% in those markets already on the overlap, right? So there's opportunity that is untapped here and the ceiling, we don't yet know what it could be, and we'll keep grinding towards that.

If you think about the business at a cohort level, Uber is no longer a new service. Everyone is broadly familiar with it. So what have you observed in terms of the size of the new cohorts you're acquiring? And now that you have an integrated platform, what happens to the observed frequency and retention of these cohorts with this broader platform approach?

So the good news is, if you think about the 200 million MAPCs that we had in Q1, roughly -- we are still driving a lot of audience growth for our platform. We have talked about 16%, 17% kind of growth on audience for now multiple quarters. And that has been the predominant contributor to a growth algorithm as we come into this year. When you think about new consumers within that 200 million MAPCs base, we added roughly 40 million first-time users in just Q1, right? And that number is 40% higher than it was 3 years ago. So we have continued to not only drive new user acquisition through the sort of markets that we have operated. It's really a product innovation led story where when you think about what has contributed to that first-time user growth, if you look at our Mobility business, roughly 25% of our first-time users in the quarter were coming from our -- the two ends of the barbell that we've talked about, both the new affordable products and the premium products that we've been adding. Similarly, for Delivery, a good contribution is coming from our grocery and retail efforts, which you would think hypothesis would be that you have all this base of online food delivery consumers, and you can sell them grocery and retail offerings Well, it turns out you can actually acquire consumers who come to you for grocery and retail and then convert them to online food delivery as well, right? So the product-led growth that we have has been very meaningful. The second piece is, of course, we're showing up in markets where Uber has been underrepresented. And our sparse markets push that we have talked about for both Mobility and Delivery has been very important there. And then finally, as you look at the new market expansion efforts as well, again, that continues to be a theme. From a retention and engagement standpoint, we are seeing continued strong retention of the cohorts we're acquiring. And again, Uber One tends to play a pretty big role in that retention effort. And then engagement you've seen the trend lines it's marching up steadily as we have gone through the last few years. Our expectation is that until we are adding new consumers, engagement growth will be somewhat subdued because there's a downward pressure from the new consumers that you're adding, but that's healthy. And in the long run, as these cohorts mature, you should see some handoff to frequency lifts as well.

On that barbell approach for the different product suite you now have within Mobility, is that widely available to consumers across your main markets? Or is there because you know you have product market fit in certain things now, right, like Reserve and Wait & Save, you know these products resonate with consumers. Are those already widely available? Or can you land and expand them? And is there a low-hanging fruit in just taking them abroad to other regions?

Yes. So the beauty and difficulty of having a global business is that you're going to have a lot of products that you can scale globally instantly and there will be plenty of products that you can only launch in a local market, right? And that's fine. We -- It means that our teams are always on top of a lot of innovation efforts across the footprint that we operate in. So maybe I'll give you a couple of examples. When you think about a product like Reserve or Teens, right? Those have global resonance. When you offer a reservation offering to consumers, there's no reason why it should be only in the U.S. It can launch in Brazil, it can launch in the U.K., it can launch in Australia. Wherever you operate, you can just go with that product pretty quickly. Teens is a similar product. You have to gear it for local regulations differently, but ultimately, it's a product that works everywhere. When we think about a product like Moto, there is a market adoption that you have doing through as well. And it works very well in Brazil, and it will not work in the developed markets, and it might not even work in some of the emerging markets where there is not a 2-wheeler culture, right? So when we kind of find ourselves with that kind of a really resonant product in one market, but we don't see fit for that in other markets, it forces us to go back to the drawing board and say, "clearly, affordability matters and it drives a ton of new customer acquisition for us. How do we ensure that we have an affordable product available in these markets?" And the answer for the U.S. has been Wait & Save, right? Wait & Save as a product has had nearly double the fit in sparser markets than other products that we have launched, similar to Reserve, by the way. And then secondly, from an affordability standpoint, it has the resonance across the U.S. And in a lot of other markets, it will have the same sort of role to play as well. So I think it's a combination of those things, and we remain nimble to address the opportunities.

And in aggregate, the Mobility business started accelerating in the back half of last year, when you came into 2026, you explicitly called out an expectation that the U.S. business would improve both on a trips and bookings growth basis. So can you maybe elaborate on, I guess, the drivers of that U.S. improvement? And really your conviction level in that playing out as you look at the balance of the year here?

Yes. So again, as a reset for folks who may be newer to the story, the arc for our U.S. business versus our international business over the last 4 to 5 years has been that the U.S. has seen a very, very inflationary insurance environment from -- starting from COVID, but truly from 2022 to 2024. And international markets where that isn't necessarily an issue, we did not have a real inflationary headwind as such, right? So what that meant for consumers was pricing in the U.S. rose as a response to the inflationary headwinds we faced. And international markets, that was not necessarily the case, and consumer pricing was much more stable in that period of time. So coming into 2025, what we saw was our U.S. business had decelerated significantly whereas our business in Europe and LatAm, which have similar or even higher penetration in some cases, was still growing 30% year-on-year. So it just created this really remarkable A/B test for us on pricing, which if you think about it, it's long-term elasticity that we were seeing at that point of time. So in 2025, given our work on policy reform for insurance, we started seeing some relief, and we started -- our philosophy has been consistent, and we've been articulating this consistently to investors as well. When insurance was a headwind, we priced it to consumers, when we find relief against insurance costs, we passed it back to consumers and not really look at it as a profit level for ourselves. So that reform theme has continued to play through 2025 and 2026. And by the back half of 2025, our U.S. business started reaccelerating. And coming into this year, we said that we expect hundreds of millions of dollars of savings from the reform efforts that we have here. And that is a natural tailwind that we expect we can pass on to the consumers. The second thing that we have learned and to your question of why we had more confidence in May versus February. We've also concluded our rate renegotiations in March, which is the annual rate renegotiations we have to do with our carrier partners. And at this point, we have confidence that the rate increases are in the low single-digit percent, which is kind of the most benign we have seen in many years for insurance costs as well. So that has been a big theme, right? Then the second piece is, again, going back to the barbell, the product adoption that we are seeing for our Wait & Save product is quite attractive. We've also recently launched Elite, which is the premium end of the spectrum, which we will continue to roll out across markets. And then U4B continues to grow at twice the rate of our overall Mobility business. So that's a big driver. And then finally, the sparse markets effort in the U.S. as well, where we are seeing double the growth of our dense markets. All of those together continue to give us quite a lot of conviction in seeing trip and gross bookings acceleration in '25 versus 20 -- sorry, '26 versus '25.

I'll have to say Bernstein will let me try the Elite product as it rolls out.

They should.

I should.

You earned it.

Look, I feel like every -- I want to talk about AVs, which is a huge theme in the space, obviously, and every, I think, difficult question in AVs comes back to this idea of will the technology fragment out over time or not? Will there be many software providers in this space or not? When you look at the subset of partners you're working with, both in the U.S. and abroad, what are the data points that give you conviction that, that is how this market will evolve over time. And there is not, in fact, a big, big gap that exists between Waymo and what everyone else is doing?

Yes. So I think the way we see AV, and I said at the outset, we think about AVs as a huge opportunity. Uber has, throughout its history, been a supply-constrained business, right? And AVs give us an ability to tap into a completely new sort of supply pool that will expand over time. And we remain excited about the intersection of autonomous vehicles with humans, and it should hopefully put us in a place where human drivers, jobs become easier and better and AVs can come in and complement the sort of rides that human drivers are providing today. What we expect will happen and what we are working on ensuring is the outcome here is multiple AV software partners that get over the finish line of L4 commercialization, both in the U.S. and in international markets. And what we are seeing right now is today, there are two scale players, truly scaled players and the scale is still relatively small, but Waymo at about 0.5 million weekly trips in the U.S. and Baidu at about 350,000 weekly trips in China. And then WeRide and Pony are the two other players who are fast catching up to that. And I know Pony recently said they will have about 3,000 cars on the road by the end of this year as well. For the Chinese players, they are looking at expansion out of China, and they're beginning to expand into a number of international markets. And we've already launched with WeRide, for example, in Abu Dhabi and Dubai, and we have announced expansion with both Pony and Baidu as well in other markets, right? So there's a lot of activity coming from those players in the Middle East, Europe, et cetera, in the next year. So for international markets, our thesis already exists to a large extent and what we are looking for and what we expect is every jurisdiction is going to look for some local winners. And we fully expect that players like Wayve will get over that finish line as well and start deploying in markets like London, Tokyo, et cetera, in the near term. So international markets, I think there's plenty of proof points. That's where the world is going. For the U.S., right now, we are working with a number of partners who are at various stages of getting to that commercialization effort. And I would say the few -- just to name a few, in the near term, what you would see from us is the first 1 is going to be our launch with Zoox, which for those who haven't followed this space closely is backed by Amazon. Zoox is live with their L4 deployments in both Vegas and San Francisco right now with a relatively limited number of cars. And as they continue to expand out, they're going to partner with us in Vegas first and then in a market like Los Angeles next year, and they will be driver out by definition from the start. They don't even have steering wheels, right? The second partner that we are looking at making good progress with is Nuro, which we have said that by the end of this year, we expect to be live with them in Bay Area with the commercial deployment. We are already doing early rider testing with them in the market. So we feel pretty good about that. And then Wayve, which I mentioned for London and Tokyo, we continue to explore whether they could be launching in the U.S. as well. And then finally, NVIDIA has been quite clear that they're targeting this category quite actively. And their approach here is interesting, and they're going to be building into this space. To augment all of those partners -- and by the way, there's a number of others who are coming out to market on a longer timeline, but we're equally excited about them. But to augment all of their efforts, we're building a suite of services we're calling Uber Autonomous Solutions. The most notable of the offerings there is our data collection effort. If you think about what is a big bottleneck to software development and the training of these models, it's rich data that can show the model all of the edge cases it will encounter in the real world. And guess what, our trips at about 40 million trips a day. We see every edge-case that can exist multiple times a day around the world. So we are beginning to deploy sensor-kitted vehicles on a fleet of Uber-driven vehicles. And these will be deployed into regular Uber trips generating revenues for Uber and collecting up to 2 million miles of data by the end of this year on a monthly basis and scaling from there, right? And then in addition to that data collect effort, we have multiple additional services offerings from customer support, insurance and everything else that you need to be able to deploy a commercial service as well. So I think there's a lot of activity that's going to come, which will help us prove out the sort of strategy that we've been articulating and we feel confident that will be the case. The final thing I'll say here, though, is -- it is important to remember that physical AI, as in terms of deployment in the real world is not as frictionless as virtual AI deployment in the digital world, right? And what I mean by that is you're going to have multiple bottlenecks that will dictate how quickly AVs can roll out across the markets. Software, which looks mature, will still run into issues, whether it's related to weather or stalled traffic lights in a city, and you will see these defleeting events that we have been seeing from various players. So it's not a straight line. There will be fits and starts. Then you've got bottlenecks on how quickly your fleet ops can scale out, how quickly you can build out the ground operations. And then OEM production is another bottleneck, you will be kind of dependent on how quickly supply hits the markets. And finally, regulations, again, are not a straight line, and you're going to have to address legitimate concerns from cities and states on labor substitution, congestion, safety, et cetera. Some of those questions very much were also there when Uber was scaling, right? So this is not a new theme. This will continue to be the case for AVs as they roll out across the globe.

Given what you just described there, what do you think a reasonable range of outcomes is in terms of how significant AVs will be for the Uber network as you look at 5 years or 10 years?

So 5 years, if you take that sort of a time frame, I think you have to start -- I'll let everyone build their own models. My view is it's still going to be relatively immaterial relative to Uber's global volumes, right? And even Uber's U.S. volumes, it's going to be somewhat immaterial in the sense that it will be a small percentage of the volumes that we're talking about. And the reason I say that is because we are currently operating with a roughly 15 billion trip volume. Our global driver base is about 10 million. Both of those numbers have been growing about 20% year-on-year, right? So you're talking about a base business where we are adding roughly 3 billion trips year-on-year, whereas the global autonomous ecosystem right now is at about 50 million trips altogether. And whatever sort of exponential curve you expect on top of that in a 5-year period, it's going to be relatively immaterial, especially when you consider the bottlenecks that exist for AV scale-out. In a 10-year period, we do think that this becomes quite interesting because at that point of time, you would have addressed a lot of the hardware scaling challenges. Hopefully, regulations have been resolved in a favorable way for the category as well, and the build-out that you need across our footprint, again, will be materializing at that stage. So I think at that point of time is between that 5- and 10-year mark is where we really see that TAM expansion potential coming because the price point needs to start coming down for TAM expansion to materialize, right? So that's where you start seeing real scale out.

And how do you see the AV ecosystem looking in that future? What are the different layers to the stack there?

Yes. So I think there's going to be five clear layers to the stack with various kinds of permutations for how they are set up with different partners. But the stack in our mind is the marketplace, which faces the consumer. Then you've got the AV software players, which are developing software, you will have OEMs who have to make the cars. Then you have fleet and asset owners, right? So somebody needs to operate the fleets on the ground. They may or may not own the cars. And then you will also have asset owners who have to own the charging stations, maintenance yards, et cetera, on the ground. So that's the fourth layer. And then the fifth is third-party financing that will be the capital that supports the build-out of that fourth layer, right? So that's the setup we see. I would say there's going to be permutations of -- there will be players who want to go vertically integrated and do everything on their own. There will be others who are going to try and do some combination of that. But our view is fairly straightforward. We'll engage with partners across that stack, and we believe we have a right to win with the marketplace itself, and that's a focus area that we will spend most of our energy on with enabling efforts on everything else.

One of the questions we frequently get is the extent to which you are willing to or want to use your balance sheet to help this ecosystem come together. What's your perspective on what that looks like now and then what that might look like longer term as well?

Yes. So I think that's a good frame that you've put it in because there is a 2-phase approach to this. The first phase is what we need to do to get the ecosystem going, and then once the ecosystem is running on its own, how do we enable it on the other side of it. So for the first phase where we are kickstarting both the software development effort, ensuring that the OEMs are coming to market and ensuring that the on-ground infrastructure, the fleet ecosystem is coming to life. We are taking a pretty active approach to handholding and supporting that development effort with our own capital. So what you've seen us do is a few things. We have stepped in with equity financing for some of our software partners. There we've gone in and supported that sort of initiative. The two markets that we have tried to solve for are: One, can we do this with a milestone-based timetable where the capital is tied to sort of our partners meeting their milestones on a predetermined timeline. And two, does our capital enable these partners to raise external capital. And that has so far been a thumb rule of for every dollar that Uber has invested, our partners have been able to raise $2.5 to $3 of external capital against that, right? So it's turbocharged their ability to raise funding. Then the second piece of investment that we have talked about is work with OEMs. And for OEMs, we have stepped in with some level of, again, similar equity investments in some cases. But more importantly, we have stepped in with offtake agreements, where OEMs know that the cars that they're producing, Uber is backing those cars rolling off their assembly lines. And the titles are transferable. So it doesn't necessarily mean that Uber owns all the cars because we can get our fleet partners to step in and own those cars as well. And ultimately, our expectation is that, that category will get fully financialized once you know what kind of revenue opportunities, lifetime -- useful lifetime and then residual value of those cars looks like. The third piece that is the sort of investment that we have talked about here is really -- and this is going to be -- have the shortest timeline in my mind. It's the on-ground infrastructure investments that we're making today. And that's really designed to support the launches we have this year, next year and maybe start some part of 2028. Because there is a lead time to building out all that infrastructure and while our belief is that it's going to be owned by third parties over time and it likely will be owned in REIT-like structures, in the initial part, we need speed to market. And as we've said, by the end of this year, we want to be live in up to 15 markets globally, we're solving for speed here versus perfection on how it's set up, right? So those are the three variations of major investments we're making. And then the fourth one is the investments we're making in Uber Autonomous Solutions, whether it's the data collection efforts or other things, which, again, are services that all our partners are going to need and over time can be potentially monetizable as well. So that's Phase 1. Phase 2 on the other side of this, we do think the vehicle -- as I said, vehicles will be financialized. The assets on the ground will be third parties financing it. And really, when you think about our software partners, they should be viable entities that do not need Uber support once they have working L4 solutions as well. So we expect to remain quite asset-light on the other side of this transition.

You talked earlier about cost curves coming down, prices coming down, that being an opportunity. One of the questions we wrestle with is, as that happens, does that put downward pressure risk on your economics in these given markets. So what's your response to the bear case that this risk is becoming a bit of a race to the bottom on pricing and that pressure is your unit economics?

Yes so I would say this is -- the question is a little bit upside down. And I know you're articulating the bear case, I've heard it many times. The scenario in which prices come down is a good scenario. And the reason prices are coming down is because software and hardware has become cheap enough and the scale of deployments has become large enough that your fixed cost burden on a per-trip basis is quite low. And when that happens, it means that consumer pricing is going down, right? And I would say, if you go back to the history of Uber's original expansion, really. It was a wrong comparison to think about ridesharing TAM as what the New York Taxi Medallion system implied, and it ended up being multiples larger than that. The reason it happened is because, one, the service was better but also it was cheaper than taxis out of the gate, right? So it does make a big difference in how the TAM looks when you come out on the other side of this. So as long as we have AV supply on our network and the prices are going down for the reason of the cost structure compressing, that's a great outcome because that means that the end market is much larger and multiples larger than what we have in the counterfactual scenario. And ultimately, what you would see then is even if the margin is lower, which I'm not sure that is -- I'm not conceding that, that's the case. But even if it is lower, the gross profit pool is significantly larger than what you would have had in the other counterfactual world of no AVs existing, right? So I think that's an important consideration to keep in mind. I would also caution here that this is a relatively long-term question because as you think about the next few years, the hardware cost structure and the scale of deployments is not large enough for the price point to be below human drivers. And we've said in the U.S., the average Uber X trip is -- it costs about $2 per mile. So AVs have to be below that $2 per mile mark to be able to sustainably offer that price point to consumers, which is lower than Uber, right? So I think that's a world in which they're going to be matching Uber's pricing for the most part, maybe plus or minus 10%. And in that sort of a world, I would think about this as a question of what kind of economics are we setting up with our partners. And generally, we now feel pretty good that the partnerships that we've announced if they are able to deliver on the time lines that we're talking about, we should be in a pretty constructive spot where both our partners and Uber's economics are healthy and sustainable for a broader scale out from there.

I want to ask you about AI workloads, which are growing within Uber and growing within all of these tech companies. When you look at the compute needs and the cost of that compute relative to maybe some of the benefits you might be getting longer term with how you think about headcount more broadly, how do those 2 things net out in your mind?

Yes. So again, this is another theme that's been quite exciting, right? If you rewind over maybe an 18-month period, which is the time line in which AI adoption at Uber has inflected quite significantly. A year ago, if you'd ask me what kind of returns we're getting for the investments we're making, it would have been pretty hard to give you a clear answer on that because it was so early that we saw the potential from the capabilities that were being presented but we didn't really have any clear proof points on what we could get there. Where we are right now is that since at least December, we've seen an exponential increase in the usage of AI tools in our developer force. At this point, 95% of our engineers are using AI tools on an ongoing basis. And that we are seeing that over 10% of the code going into production for us at this stage is autonomously written, meaning engineers are coming in only at the review stage before it goes into the production line, right? So we are a culture, we're a company where the culture is. We want to have precision. We want to be able to say, for a dollar, we are investing, what is the return we are getting, and that's how we've run all the billions of dollars of incentives that we deploy. I would say we're not quite there on AI investments for developer productivity, but directionally, it's a clear signal that what we are seeing is that for the investments we're making, we're getting a productivity lift. And if you ask me again in a year, we're likely going to be able to even quantify what kind of returns we're getting, right? So that's on the developer side, what we've been seeing. The second area where AI is having a pretty big impact is on our customer-facing efforts. So customer support is the first area where we are making reasonable investments here. For us, we are doing about 15 billion trips. Every trip has at least two humans involved. So the surface area for customer support tickets is very, very large. And we -- as you can imagine, with the heterogeneity of our business across the globe and the sort of products we sell, we have a lot of policies for customer appeasement when it comes up. I would say AI in this specific org is going to have a pretty meaningful impact and we're already beginning to see some cost savings and better quality of customer support outcomes for our consumers. And then finally, on consumer-facing Agentic users, we are beginning to work on first-party Agentic solutions on our apps, and we are also engaging with the third-party services there. So maybe just the bottom line, though, is where we are today, we have enough conviction that the investments we are making here are driving productivity. And so what we have done is we have tempered the pace of hiring, and we -- and this is broadly across the company, but specifically from an engineering standpoint, the hiring ramp we have for the remainder of the year is significantly lower than what we thought it would be when we came into this year.

So we've got tempered hiring. We've got hopefully more fixed cost leverage coming through related to that. The Uber story over the last few years has been one of profit expansion quite substantially. When you think about reinvestment opportunities and OI flow through to the bookings growth you're putting up, how do you think about that balance between the two? And how should we be thinking about your margin and your margin ramp from here?

Yes. So I think, as I said earlier, I don't think there's a big change in our philosophy. We have, over the last few years, delivered really attractive top line growth. And we have done that while delivering a lot of margin expansion against that, right? And it's been very much driven by operating margin expansion, driven by fixed cost items, but also just discipline on how we're thinking about all the discretionary levers that we have. As we look forward, we have certainly tempered the pace of margin expansion relative to the historical norms. And the context there is important, right? We were coming from a world where Uber had been quite unprofitable, then it became barely profitable, and now we're getting to a spot where we are at a decent level of profitability. So that second derivative certainly is changing the pace at which we're expanding margins is slower than we have done in the past. But it's not a function of our inability to drive more margin expansion, but rather a function of the opportunities we continue to find, both in Mobility and Delivery. And so as long as we can sustain this mid- to high teens kind of growth engine, and it's coming from volume-based growth rather than from price. That's a good place to be as long as we can sustain that with healthy margin expansion beyond that growth, right? That means that we can compound our bottom line at a very attractive rate. And then, of course, we continue to do more optimization below the line as well.

Coming over the last couple of minutes here, we've talked about a lot of different themes across the business. How would you just generally sum up the direction that Uber is going in? What's the message you want to leave investors with? And what they should come to expect of you going forward over the next few years?

Yes. I would say, again, this is a business that has a meaningful impact on consumers and earners around the world. And we take that very, very seriously. It means that we have to be continuing to innovate and earn our place for that consideration that both our consumers and earners and merchants are giving us. And innovation is going to be top of mind for us to serve those audiences. As we do that, we also have a high bar for the capital allocation that we need to do here. I talked about the priorities and the sort of inflection in our opportunity with AVs that we have in front of us. So we will invest behind that. But we are in a place where we can do that in a disciplined manner and continue to drive accretion over time and really reducing our share count remains a focus area for us as we go forward from there as well. So what I would -- if I had to describe it in a sentence, I would say this should be a company, this will be a company that continues to be innovative remains disciplined and really drives a disciplined kind of financial model for our investors as well.

Maybe a quick final question there. Should we think about this as an "and" strategy between these different capital-allocation priorities?

Yes. I think I would very simply think about it as where we are and where we're going. So if you're looking at our free cash flows right now at about $10 billion and you project out the next 3 years, our free cash flow is growing at a healthy rate. So you can see the sort of free cash flow output that we are generating here. We've got -- we are a solidly investment-grade company. So we have plenty of firepower from our investment-grade rating as well. And we've articulated our financial policy clearly that we will not exceed 2x leverage, but that means where we are sitting with 1x leverage. We have plenty of room there. And then finally, we've still got $8 billion or $9 billion of equity stakes on our balance sheet, which are we've been pretty clear. Most of them are financial investments for us, not strategic. So we'll look to monetize them when it makes sense. So all of that put together puts us in this place where it's an "and". We can make investments in AVs, we can pursue some disciplined M&A, and we can continue to buy back our stock. So there's not really a meaningful trade-off in those from my mind. And what we're solving for is the best financial output that we can drive for the company.

Great. That's a great place to leave it. Balaji, thank you so much.

Thank you.