Uber (UBER.US) Q3 conference call: Brazil and India are dragging down the “driver shortage”, and supply has now improved in October
The Zhitong Finance App learned that Uber (UBER.US) announced the results for the third quarter of 2026 before the US stock market on November 4. Revenue for the quarter was US$14.766 billion, up 10% year over year, higher than FactSet's forecast of US$14.66 billion; non-GAAP earnings per share were US$0.88, up 35%, higher than the forecast of US$0.86; total bookings were US$59.819 billion, up 20%, slightly lower than StreetAccount's forecast of US$59.832 billion.
Travel revenue fell 2% year over year to US$7.520 billion, lower than the forecast of US$7.71 billion, and the growth rate of bookings fell to 18%, becoming the core topic of the conference. CEO Dara Kosrosahi said that the travel business “grew healthily, but fell short of company expectations because the supply of drivers in several international markets failed to keep up with strong consumer demand.” Boosted by accelerated profits and progress in autonomous driving, Uber's stock price closed up 4.91% to $69.87 on Wednesday, recovering half of the increase after an intraday high of $72.50. Kosrosahi said that October has seen “encouraging improvements,” and the company expects total bookings to increase by 18% to 22% in the fourth quarter, the same as in the third quarter.
Focus and Response
What received the most attention at this conference was the supply bottleneck in the travel business. Travel bookings for the third quarter were US$29.512 billion, up 18% year on year. The company said the growth rate was 1 to 2 percentage points lower than its own expectations; travel revenue fell 2% to US$7.520 billion, lower than the StreetAccount statistics of US$7.71 billion, and the sector's operating profit increased 27% to US$2.362 billion. Management attributed the gap to supply rather than demand: Chief Financial Officer Balaji Krishnamurti said that the escalation of the Middle East war is driving up global fuel costs, and the impact on driver supply in price-sensitive emerging markets such as Brazil and India is obvious. The combination of tightening immigration policies and peak season demand has widened the gap, but he stressed that “we are still experiencing supply constraints, not demand constraints.”
The company said it has made targeted investments in driver acquisition, retention and participation. The increase in global supply hours accelerated in October, and Brazil and India improved markedly. Analysts asked three times about the sustainability of supply issues and the extent of investment, and were concerned about whether the 8% profit margin in the travel sector could be maintained. Krishnamurti responded that it would not indicate any substantial changes in the profit margin trajectory. Supply investment in the fourth quarter mainly affected the yield rate, not the profit margin.
Delivery becomes the main engine of growth
Delivery bookings amounted to US$28.539 billion, up 22% year on year, fixed exchange rate increased 23%, organic growth accelerated; revenue of US$5.488 billion, up 23%; sector operating profit of US$1,057 billion, up 30%. Kosrosahi said that the number of US merchants increased by about 12% year on year, members currently contribute 70% of delivery bookings, and Uber One members are about 55 million. Freight bookings reached US$1,768 million, an increase of 35%, and operating losses narrowed to US$33 million.
Grocery retail and dark stores
The annual operating rate of the grocery and retail business was US$15 billion, an increase of 40% over the previous year. The company said it has reached market openers nationwide and added partners such as Target, Decathlon and Priceline in Australia, Tesco (Tesco) in the UK, and Rewe in Germany. Krishnamurti said that the hidden store is a “plan adapted to local conditions” and will not be fully rolled out; market conditions and asset return requirements must be met before entering.
Autonomous driving shifts to commercial execution
Kosrosahi said that the focus of autonomous driving “is shifting from cooperation and testing to driverless commercial services.” Zoox will launch Uber in Las Vegas later this month, and Motional will then switch to fully driverless; the platform will have about 700 vehicles by the end of the year, more than 2,000 vehicles by the end of 2027, and tens of thousands of vehicles in 2028 and beyond. Krishnamurti said that it has long hoped that third party capital rather than Uber's balance sheet will cover fleet financing, and the company is satisfied with the returns that can be obtained by third parties.
Fourth quarter guidance
The company expects total bookings of US$63.25 billion to US$65.25 billion in the fourth quarter, up 18% to 22% year on year at a fixed exchange rate, with a median value of US$64.25 billion, slightly lower than the StreetAccount statistics of US$64.26 billion; non-GAAP earnings per share were 0.95 to 0.99 US dollars, up 34% to 39% year over year, and the median value was in line with FactSet's expectations of $0.97.
Capital allocation shifts to integration
Common shares were repurchased in the third quarter of US$750 million, with a cumulative total of US$4.279 billion in the first nine months. Kosrosahi said that the company will invest more resources in organic growth and the autonomous driving ecosystem and “shrink mergers and acquisitions”. The current focus is on integrating the announced Delivery Hero and ezCater deals. He said that the high-end business grew 35% year over year, and corporate mobility (U4B) grew by nearly 50%, which is one of the two main lines of mergers and acquisitions.
The contribution of AI agents is still minimal
According to Kosrosahi, 75% of trips are completed through “one-click ordering”. The average time to place an order is 50 seconds, and the average time it takes for diners to repurchase is 90 seconds. In the week of Halloween, Uber and Uber Eats completed a total of about 320 million trips. Of these, there were about 500 Uber Eats orders from AI agents, and only 4 travel orders. The company said the transaction from the agent was “very, very early” but more than 10 AI assistants are already negotiating access terms.
Competition and personnel
Referring to the potential merger between DoorDash and InstaCart, Kosrosahi said, “Honestly, I don't think this will have much impact on us,” adding that existing merchants are actively engaging with Uber and hope to keep the second platform option after the merger is completed. Additionally, President and Chief Corporate Affairs Officer Jill Hazelbeck left his job after 10 years and will join OpenAI.
Full meeting minutes:
I. Executive speeches
Dara Kosrosahi, CEO
Thanks, Alex. The third quarter was yet another quarter of strong growth and continued expansion of profitability for Uber, with record levels of users and engagement on the platform. We have reached 215 million monthly active platform consumers who use Uber more often than ever before. This helped us achieve our fifth consecutive quarter of an increase of over 20% in total bookings, a 35% increase in non-GAAP earnings per share, and a record free cash flow of $10.4 billion over the past 12 months.
The mobility business continued to grow at a healthy rate, but fell short of our expectations as driver supply in multiple international markets failed to keep up with strong consumer demand. We are responding with targeted investments to improve driver availability and market balance, and we saw encouraging improvements in October.
Our focus is to drive continued growth while maintaining strong profitability in the business. The delivery business experienced another excellent season. Total bookings increased 23% year over year, and organic growth accelerated. Our investment in choice, affordability, and reliability makes Uber Eats more useful in everyday situations, from restaurant meals to weekly groceries to last-minute essentials. With approximately 55 million Uber One members, we are deepening consumer relationships in the two major businesses of travel and delivery.
These direct consumer relationships are at the core of our AI strategy. We're using AI to make everyday transactions easier, make discoveries more personalized, and selectively explore new distribution channels that may bring incremental demand.
In terms of autonomous driving, our focus is shifting from collaboration and testing to driverless commercial services. Zoox will launch Uber in Las Vegas later this month, and Motional will soon switch to fully driverless operations. We expect the Uber platform to have around 700 fully driverless vehicles by the end of the year, and more than 2,000 by the end of 2027. Equally important, we are focused on demonstrating the utilization and economics required to scale up.
Entering the fourth quarter, our priorities are clear: improving mobility supply, strengthening delivery momentum, meeting autonomous driving milestones, and continuing to grow profits.
Before handing over the conference to the moderator, I would also like to give special thanks to our President and Chief Corporate Affairs Officer, Jill Hazelbeck. After working at Uber for 10 years, she is going to OpenAI. It's a huge gain for them, and for us, losing someone like Jill is a loss. She has always been an important part of the company culture. Fortunately, she has built an excellent team, many of whom will advance to higher positions, and many work directly with me. So I want to personally thank Jill for the past 10 years, and she has helped me in countless ways, both professionally and personally.
She's an excellent consultant. This is a huge win for OpenAI. I'm really happy that the team she's built is ready to take on greater responsibilities and they are totally worth it in the company. Thanks so much Jill. Having said that, I'll leave the meeting to the moderator to ask questions.
II. Q&A session
Goldman Sachs Analyst Eric Sheridan: Capital Arrangements and Return Calibration for Electric Vehicle Fleet Financing
Q: Thank you very much for taking the questions. I'd like to go back to the shareholder letter and prepared statement on electric fleet financing. Can you talk more deeply about how we should consider balance sheets and incremental free cash flow, and how much money would you like to allocate behind these investments? How should we view return rates, or how do we align economics with business opportunities? Thank you very much.
A (Balaji Krishnamurti): Thank you, Eric, I will answer this question. I don't think it's convenient to talk too much about the specific structure being designed yet. But I'd like to ask you to take a look at the $1 billion financing arrangement we set up with Santander earlier this year for our traditional fleet partners. This could serve as a template for the arrangements we make in the next few months on our autonomous vehicle fleet. We are actively promoting such arrangements. Specifically, you should expect us to make announcements on both infrastructure financing and vehicle financing.
Our long-term expectation is that this portion should be funded by third party capital rather than Uber's balance sheet. In the short term, as you have seen in the press release, there are about 700 driverless vehicles on our network this year, reaching 2,000 by the end of next year. Capital investment in these vehicles is relatively limited. But as we begin to expand beyond 10,000 vehicles in 2028 and beyond, we hope to bring in third-party capital. We're happy with the return on investment that third parties can get.
A (Dara Kosrosahi): Eric, I'll add one more thing. We think the financing for this type of vehicle would be a bit unusual. Looking at the financing of car rental fleets, etc. is all based on the residual value of the vehicle, and I don't think the residual value of autonomous vehicles has yet been established. So in the early stages, we'll put these vehicles on our balance sheet. We have a very strong balance sheet and are capable of doing that. I think a lot of fleet financing will be based on the revenue potential of these fleets.
What we can prove is that vehicles on our network are being utilized at a very high rate, actually higher than vehicles not on our network. We showed a 30% utilization premium on the network. We are very flexible in doing this. So if a fleet fails operationally for some reason, we can take that autonomous vehicle into separate ownership, and it can still make money on the Uber network. Therefore, we believe that financing on the Uber network has a real advantage over alternatives in terms of capital costs. This is the business model we're building with our partners, and their response has actually been quite positive, and they want to spread the risk. So we're pretty optimistic about this, but we have to establish a business model, prove utilization advantages, and show these to banks before we can move forward.
Bernstein Analyst Nikhil Devnani: International Market Supply Gap and Investment Intensity
Q: Thank you for taking the questions. I'd like to follow up on the international market supply developments you mentioned. Want to check if anything changed during the quarter? I remember last time we talked about Brazil, but now it seems like there are more markets reflecting this situation. It would be helpful if you could elaborate on what you saw during the quarter. Also, looking ahead, how should we consider the required supply-side investments? Because some variables, such as immigration, feel more enduring. So it's helpful to understand the strength and durability of these investments, as you're trying to speed up your business again. Thank you.
A (Balaji Krishnamurti): OK, thank you Nikhil. What I'm saying is that many of the situations are a continuation of the various dynamics we've seen this year. If you look at the macro environment, with the escalation of the Middle East war, the impact on fuel costs has already been felt on a global scale. For us, the most obvious ones are emerging markets, where both consumers and drivers are price sensitive. What we talked about in Brazil last quarter was the impact of distribution competition in the market on motorcycle supply.
However, in the past quarter, we saw a rise in fuel prices in markets such as Brazil and India, which had quite an obvious impact on the supply situation. Similarly, immigration dynamics exist throughout the year. However, as we entered the second half of the year, we saw a rise in demand brought about by the holiday season, and the demand side showed quite strong performance. When supply was in short supply, the impact of supply on the market was amplified. This is also the situation we mentioned in Canada. Now, on our scale, this is a good annoyance.
In operation, we are still facing supply constraints rather than demand constraints. We know how to meet these kinds of challenges, and history has proven it. We are making targeted supply investments, focusing on driver acquisition, retention, and engagement. As this investment progresses, we are already encouraged by the early signs of supply seen in October. Global supply hours growth accelerated in October, including significant improvements in Brazil and India. So, taken together, you can see that our outlook is stable in the 18 to 22% growth range, the same as in previous quarters.
So we feel good about the current situation; the growth runway ahead is very long. Over the past 5 years, our penetration rate of eligible welfare population in the top ten markets has doubled, yet it is still only 10%. So we feel very good about the future space.
A (Dara Kosrosahi): I think we're dealing with an additional factor: we're clearly facing year-over-year supply pressure, but we're also facing seasonal supply pressure. September, October, November, and December are our peak travel and delivery seasons. In some markets, such as Brazil, travel and delivery rely on two wheels of capacity, which is one of the types of supply we need. So these factors are intertwined. This is a seasonal issue as well as a year-over-year issue, and both occur at the same time. But we have a good sense of it.
We are very optimistic about the October trend and the trend going into the fourth quarter. Also, this isn't the first time we've dealt with this situation. Next question.
Morgan Stanley Analyst Brian Novak: Grocery Retails' Self-Employability and Clandestine Store Strategies
Q: OK, thanks for taking the question. I'd like to ask a question about groceries. From a strategic perspective, what investments do you need to make in the grocery and retail business in the US market to further accelerate and compete in this field? What do you think about the need for stronger self-employment (1P) capabilities, that is, own secret stores or own contract operations?
A (Dara Kosrosahi): OK, Brian. We're excited about developments in the grocery and retail business. Our current annualized operating rate is $15 billion, an increase of 40% over the previous year. We think this is a $10 trillion addressable market, a huge market for us. What we've seen over and over from consumers is that they want more, faster delivery, and they want everything faster. When Amazon introduced overnight delivery, it was magic; now when hourly delivery was magic, soon people would expect everything to be done in 15 minutes. We're right in this field.
Our investment areas are really about choice, affordability, and consumer experience. In terms of choices, we're very excited to expand Costco across the country. We've also added many new international players, including Target, Decathlon, and Priceline in Australia, Tesco (Tesco) in the UK, and Rewe in Germany. So we will continue to invest in product selection. As product selection increases and more consumers go online, the conversion rate also increases. We continue to invest in affordability, whether it's the Tuesday Fresh Sale or the members' free shipping on orders over $60, our members love it.
We then continue to invest in product innovation to ensure picking and packaging is as accurate as possible. We invest in shopper quality, pay more to high quality shoppers, and introduce technology, such as having shoppers take pictures of empty shelves to verify that the item they're looking for isn't actually on the shelf. It also helps consumers not get angry when they see empty shelves, it's not the shoppers' fault, it's just that the item they want isn't there. This helps build trust between the two parties in the market.
So for us, it's all about execution. When it comes to dark stores, our top priority has always been to work with our partners to help them grow their business. We're excited about Delivery Hero, who have built a sizable instant retail and dark store business. The benefits of hidden stores are that you have better control over your inventory, the success rate of picking and packaging is higher, and the out-of-stock rate is lower, so quality is improved due to vertical integration. This is a very exciting part of our potential acquisition of Delivery Hero.
That team has been building the instant retail business for over 7 to 8 years, has achieved break-even, and is growing at a very attractive rate, and we're really looking forward to including it in the portfolio after completing the Delivery Hero deal. One more point about secret stores.
A (Balaji Krishnamurti): For us, this is a tailor-made solution. We don't want to get into this business everywhere, but there are markets where it makes sense. As Dara mentioned, Delivery Hero has done a very good job with its dark store layout in the Middle East. We are also in a position to examine the Turkish business, which is close to profit, both through our presence in this market and through the GATHER deals we have completed. Furthermore, we also tried organically in Taiwan.
So at the end of the day, it must be a business that can operate well, must meet market conditions, and must have a good return on assets when entering this field. We'll be selective where we participate. Otherwise, we'll work with grocery partners who are experts in their field.
Citibank Analyst Ron Josie: The Geo-Rhythm of Launching Driverless Vehicles
Q: OK, thanks for taking the question. I'd like to ask more questions about the promotion of fully driverless cars, from 700 cars this year to 2000 cars next year. Can you help us understand the geographical pace of promotion? I remember you said that the US and international markets each account for 50%, how did you consider that. Thank you.
A (Dara Kosrosahi): Yes, of course. We're so excited. We already have a few live projects. Obviously, we have cooperated with Waymo (Waymo) in the US market, and have also launched driverless services with WeRide (WeRide) and Baidu in several international markets. We expect to launch Zoox and Motional in Las Vegas this year, as we mentioned. Next year, we expect to add Nuro in California and then in Houston, Texas, at least early 2027. These are all completely driverless commercial passenger services. The AV Ride will also go live in Texas in mid-year. Internationally, we expect to add Pony (My Little Pony) and Momenta to our list of driverless car operators.
At the end of 2027, I do think we will test passengers and vehicle operators with NVIDIA (NVIDIA), Rivian, and hopefully Waabi, and Wayve outside the US. So, the partner strategy and combination of all these partners will really begin to accumulate. We are quite confident that we will reach 2,000 vehicles by 2027, but the real goal is to reach tens of thousands in 2028 and beyond, truly expanding the ecosystem. Also, we have operations in these markets, we have partnerships with cities, we operate in those markets, and we don't need to build demand from scratch.
The demand is already there. So I think we can launch services for partners very quickly and with a very high utilization rate, which makes us more optimistic about the opportunities here.
Wolfe Research Analyst Shweta Kajulia: Autonomous Driving Unit Economics and Capital Allocation
Q: Thank you for taking the questions. I'd like to ask two questions. The first one concerns the unit economy of autonomous driving. In the example you gave, paid trips increased by 30% and fixed costs decreased by 23%. Even if partners pay Uber a transaction fee or commission, it's still incremental for them. So my question is, how long will this 30% increase in paid trips last? How long will this advantage last in the long run? The second question concerns capital allocation strategies. You mentioned this in your shareholder letter, and there are no changes in this regard. As far as priorities are concerned, you don't expect further mergers and acquisitions before or after the end of the year. But I want to ask what is ezCater's logic? What do you think of future capital allocation for mobility, delivery, and autonomous driving? Thank you very much.
A (Balaji Krishnamurti): OK, Shveta. I'll answer the first question, Dara added the second one. As for increased utilization, I'd like to start by saying that we think this is a very lasting advantage for Uber. To provide context, when we talk about a 30% increase, we're comparing Uber's hybrid deployment in the US to a mature self-operated deployment. These mature self-operated deployments have structural advantages because they include San Francisco, which is currently particularly beneficial for autonomous driving deployments. As the scope of deployment expands, we expect the utilization rate of homegrown deployments may be worse than what they saw in San Francisco.
In fact, this is exactly what we saw on the ground today. As far as Weimo and Tesla are concerned, the number of daily trips per car in the new market they have entered this year is in the low range of about ten, and what we have seen on these mixed market networks is a high range of 20 or so. In fact, in these markets, our utilization rate is over 100%. I think another aspect is that we might also have an advantage in other parts of the unit economy. First, vehicle costs will drop over time, which is true for many participants.
But overall, what we're trying to ensure is that when we work with OEMs, we can see preferential pricing due to size. The second area where we will also have an advantage is sharing charging, maintenance, and fleet infrastructure, as these can be shared among multiple EV partners. I tend to compare it to the advantage that hyperscale cloud vendors get over local data centers serving a single enterprise. So I think these two points, plus the utilization rate, will support long-term economy. Our goal is for autonomous driving trips to provide profit margins comparable to human driving trips over time. Once above that level, we can share the savings with consumers and truly expand the addressable market.
A (Dara Kosrosahi): OK, with regard to capital allocation, etc., our priorities have not changed. Clearly, the first priority is to drive organic growth. Investing in the autonomous driving ecosystem, we talked about the $10 billion commitment based on certain deliverables, including investments and partnerships, and autonomous vehicle fleets, which is another priority. Then there are mergers and acquisitions and buybacks, and we have sufficient capital for both. We're moving to the first two and scaling back on mergers and acquisitions, mainly because we're really focused on consolidating deals that have already been announced, and overall they've performed well.
But we need to make sure we get the right cost of capital in mergers and acquisitions. If you look at our historical M&A ideas and future directions, it mainly focuses on a few categories. First, expand the geographical scope of transactions. If they allow us to provide services in more countries, Delivery Hero is an example, and Trendyol is an example.
Not only are we entering a new geographic region that is growing very fast, but it also allows us to provide platform advantages, that is, cross-promote travel and delivery, and cross-promote membership plans to be implemented in more countries. We have operations in many of Delivery Hero's marketplaces; this is just a geographic extension of our platform. Our second reason for concern is that we can build specialized capabilities that we believe are truly profitable. We talked about the growth of the high-end business; for example, it grew 35% year over year, and our corporate mobility (U4B) business grew nearly 50% year over year.
Therefore, high-end is our theme, and corporate business is our theme. This is the direction we want to double our investment. For example, this is the acquisition of Blacklane and ezCater, which is what you asked about. Blacklane has provided us with more professional high-quality transportation, while ezCater has professional capabilities, and they only work with merchant partners who focus on food and beverage. Often, delivery partners have to go into the office to arrange everything, which is a different level of service.
Being able to provide Blacklane and ezCater to our corporate travel audience is a huge advantage, and I think we now have half of the Fortune 500 companies as customers. So these are mature businesses, and we can introduce different competencies to the corporate travel audience and connect more closely with them. The third is a category of new products that can grow with our audience. Spot Hero is an example. We want our products to be more relevant to commuters; commuting is one of our biggest categories.
Being able to give commuters an Uber to Work pass to avoid peak pricing and provide parking when they want to drive themselves is a great example for us to build a service inventory for our top customer base. So these are the topics we're looking for. We're very excited about the book acquisition. We are very excited about ezCater. Overall, as an independent business, they are operating at a higher level than our trading assumptions, and now we need to ensure that the integration is in place and the team is treated well.
If all goes well, that's why next year I think the focus will be more on organic growth, autonomous driving, and then opportunistic share buybacks.
A (Balaji Krishnamurti): I would like to add that from a process perspective, we are always evaluating the “buy” logic of the transaction for any transaction that you see us doing. We're also comparing value-added math; what it would be like to buy back our own stocks directly. In this specific case, since you specifically asked EzCater, the math of self-built and acquired couldn't be more obvious. We're talking about buying a business that's already profitable now, and we have more synergy effects to come. Compared to self-built businesses, we expect self-built solutions to reduce profits by hundreds of millions of dollars.
So I think we feel really good about this logic and are now focusing on execution.
MoffettNathanson Analyst Michael Morton: Opportunities and Risks of DoorDash's Merger with InstaCart
Q: Good morning, thank you for the question. I'd like to talk about the potential merger between DoorDash and InstaCart. Uber has talked many times about its cross-platform experience from Uber Eats to the mobility business. So I was wondering if you could talk about the opportunities or risks this merger might create. Thank you.
A (Dara Kosrosahi): Yes, of course. Honestly, Michael, I don't think this will affect us too much. Our grocery and retail business is growing, with an annual operating rate of $15 billion and a 40% increase. Our focus is really on continuing to improve products, and products still have a lot of room for improvement in product selection, quality, punctuality rate, and out of stock rate. When we talk to our merchant partners, they want choices, they want to develop direct channels, and they want many partners to work with, including us, DoorDash, and InstaCart.
So this potential deal took a partner from the table. We have received active contact from our partners. They are very worried about this potential deal and ask us what we think; secondly, they want to actually cooperate more with us and build a business with us, so if DoorDash and InstaCart merge into a business, they won't rely too much on a single platform, which will account for a very large share of the online market. But from our perspective, the potential of this market is 10 trillion dollars, which will create a competitor with a large share of the online market.
But we think we'll continue to invest in Uber. These types of acquisitions can be disruptive. Currently, we have received contacts from merchants who want to connect more deeply with us and want to build a business with us, so they have alternatives. As to whether one alternative or two is better, it is up to the government to decide.
Wells Fargo Analyst Ken Gavlersky: US Travel Trends and Organic Acceleration of Delivery
Q: Thank you. If so, I'd like to ask two questions. First, can you talk about the trend of domestic travel in the US? What did you see in the third quarter? You mentioned that travel growth is improving. I'd like to see if I could talk a little bit about how the US situation compares to the international market you mentioned earlier. In particular, the market is concerned about the continuing situation in California, and even some Texas markets where autonomous driving penetration is higher. Second question, you mentioned the acceleration of organic growth in delivery. Can you talk about the drivers of restaurants and groceries, respectively? Are you seeing the core restaurant business accelerate in the overall organic acceleration? Thank you very much.
A (Balaji Krishnamurti): OK, thank you Ken. Regarding US travel, we saw America's relatively resilient growth, compared to the international trends we talked about earlier in the conference. Growth slowed in the third quarter, but was not as severe as in international markets. The rise in global gasoline prices is a problem, and we are also seeing some supply pressure in the US, but we are dealing with the targeted supply investment we mentioned earlier. I think in addition to the supply problem, what we continue to see in the US is a steady rise in insurance tailwind.
We talked about a marked acceleration in travel growth in California, which continues to outperform the rest of the country. Second, our product barbell strategy continues to be very healthy in the US. American barbell products tripled their overall business in the third quarter, driven by strong innovations around growth, including Wait and Save, Comfort, Reserve, and Black, as well as preferred products for young people and women, all targeting new consumer groups and creating more choices for our customer base. Finally, from the perspective of a sparse market, we see too much room for growth in the US.
In sparse markets, less than 10% of eligible consumers have used Uber in the past 12 months, compared to 50% in dense markets. So we continue to improve supply and reliability in these markets, which should have a very clear long-term impact. Dara, would you like to answer the second one?
A (Dara Kosrosahi): OK. We are quite satisfied with US delivery, which has accelerated very well. This is the second consecutive quarter of organic growth and delivery as well as US delivery. There are many reasons. First, mathematically speaking, our grocery retail business and direct sales business are growing faster than our takeout delivery business, and are becoming a larger part of the portfolio, so a larger, faster growing part is getting bigger. Our core restaurant business is also accelerating. This is indeed due to merchant choices, and the number of new businesses we added in the US increased by about 12% year over year. As we add more options, everything helps increase conversion rates, and we get new customers online. Membership is definitely our best. Members now account for 70% of total delivery bookings, and members are deepening their interaction with consumers. Of course, there are also sparse markets.
As Balaji said, the sparse market is a strategy we implement globally, but it is definitely a strategy we implement in the US for delivery. This helped. Last but not least, the deal. We focus on merchant funded promotions, in addition to the membership discounts we give to consumers. If you look at our advertising business and commercial-funded promotions, they grew by over $10 billion year over year. This is the money we give to consumers, and the return to merchants is new consumers and a higher penetration rate in the market. So all of these are working right now.
If there's one problem, it's that it's just like travel, and we continue to invest in delivery and couriers to ensure that the growth of couriers keeps up with the consumer demand we're seeing in catering. This is our biggest challenge right now. But we think the team is fully capable of handling it.
Evercore analyst Mark Mahaney: Sustainability of profit margins in the mobility sector
Q: Hey, I just want to ask a question about the profitability of the mobility sector. The profit margin last quarter was 8%. As you launch supply plans in international markets, does that mean you can maintain your current profit performance? Or would there be a detour? Is this structural or just a temporary detour? Or will there be a detour in terms of profit margins? Thank you.
A (Balaji Krishnamurti): OK, Mark. I think I'm still going back to our old philosophy, which is that we want to drive the company's overall profit margin to expand year over year. Ideally, we would also like to do the same thing for all sectors. However, on a quarterly basis, there are some fluctuations in the business, regardless of the supply investments we usually make during periods such as the holiday season. So I won't.
I'm not going to point out any substantial changes in the profit margin trajectory of the business, and we will provide some explanation on the impact of travel supply investment on profit margins in the fourth quarter, but from a profit margin perspective, this is not necessarily a substantial change. Look, I think the last point here about profitability is to look at our outlook. We delivered a quarter of moderate earnings per share growth of 30%, and the fourth-quarter outlook we just provided was the midpoint of a 37% year-over-year increase in non-GAAP earnings per share. So if you look at it, you'll see profit growth expanding, which puts us in a pretty good position in 2027.
A (Dara Kosrosahi): Mark, add one more point so you can understand the supply and demand situation. We have basic pricing for consumer pricing, as well as basic pricing for driver and courier pay. And then we have motivation. There are rider-side incentives, driver-side incentives, diner-side incentives, and courier incentives. These are more flexible funds. We'll ensure drivers show up for specific times, such as Halloween; or for specific locations, if we think there's a game, we'll provide incentives for drivers to show up at the game at the right time.
But the cool thing is that we're just rebalancing incentives between passengers and drivers, as in this case. Last year, more incentives went to passengers to lower prices, and this year we reduced passenger incentives and used this money to fund driver incentives. This does mean that sometimes passengers are more expensive, which is something we are concerned about. We'll weigh profitability, but that's what we need to deal with. Having a flexible incentive pool, which we allocate to supply or demand based on market balance, is part of the art we master.
Then there's a lot of science, algorithm science, such as if you had a dollar, what would be the maximum predicted return of that dollar? Should I invest this dollar for drivers in Brazil or passengers in India? Algorithms are making these choices, and they're structured to make these choices based on billions of trips of training, and can distribute these investments in ways humans can't.
J.P. Morgan Analyst Doug Ammuth: AI Agents' Direct Connect Strategy and Third-Party Distribution
Q: Thank you so much for asking. Would love to hear your latest thoughts on AI agents. Can you talk about some of the goals and differentiation you can provide by building agency technology directly within the Uber platform? Then what do you think of the opportunities on third party platforms like Mistral. Thank you.
A (Dara Kosrosahi): Yes, of course. I think the focus is on making sure we invest in providing the best experience through direct connections. Price, delivery time, reliability, choice. Obviously, we have membership benefits, and members now account for over 50% of our bookings. It's just the best experience to build a direct connection. We're really focused on making sure the direct-connect app experience is already frictionless.
So, the average trip order takes less than a minute, actually 50 seconds. This is the time people spend on our app, and 75% of our trips are based on a personalized single click, based on your destination and the data we have. If you look at the median time for diners to reorder their favorite food, it's only 90 seconds. So we've made sure that accessing Uber doesn't require a lot of fuss; you can come in and out.
Because of the data we have, where are you going, when, etc., we think a large number of people will come directly to our platform. Listen, I used to work in the travel industry, we are dependent on Google, and I don't want to go back there. Therefore, I think the truly first-class market technology, which we are building and operating, will bring more and more people directly to us. We're building our own AI experience.
For example, we have a shopping cart assistant, and we are very excited about it. Early signals suggest that shopping baskets built using shopping cart assistants are about twice the average shopping basket. We are also studying various types of assistants, driver assistants, etc., to ensure that we have the best AI system in terms of AI, which allows you to enter and exit our service very quickly. That being said, we'll be discussing it with a third-party AI agent. We see this as another opportunity to reach our platform, and we think it will be a net positive aspect of ecological needs.
You can guess based on our service that every AI systems company is talking to us, and we're actually negotiating the terms of the deal right now. The good news is that in most markets, you have two big travel providers and two big delivery providers. I think I've counted them now. There are 8 different assistants, and I think there will be more than 10. So I think that as a content provider in our field, in the physical space, we are the performing party. The services we have built are in high demand.
I think the demand for our services will be very, very high. So what we're really excited about right now is still very small. I think you're telling me what is the current trading volume of AI?
A (Balaji Krishnamurti): OK, I want to use numbers to explain. Last week was one of our biggest weeks because it was Halloween and we completed around 320 million trips on Uber and Uber Eats. The number of Uber Eats orders from AI agents is about 500, and the total number of orders from agents in the travel business is 4.
A (Dara Kosrosahi): So it's very, very early, and we're currently seeing very limited traction. Listen, we're hoping there will be more traction. We do want consumers to use the service in an enjoyable and simple way, and we think the direct connection service will be the most enjoyable and simple, but we work with third parties. It's just very, very early. We want to make sure we don't rely on any single player. I think we're in a very good position in this environment. OK, that's all I think. Thank you all for participating.
Thank you to the Uber team for everything they've done to deliver an amazing quarter. We're excited to end the year with a very, very strong quarter, and very grateful to our investors and shareholders for putting their money and trust in us. More exciting is yet to come. Thank you.