Uber Stock Analysis: Record Cash, a 28.6% Slide and a Moat Robotaxis and AI Agents Can Bypass

Uber is a better business than it has ever been. It generated $9.8 billion of free cash flow in 2025, its gross bookings have grown more than 20% for four quarters in a row, and 208 million people used it in an average month of the second quarter of 2026. None of that has helped the share price. On 12 October 2026 Uber closed at $71.50, 28.6% below its all-time-high close and roughly where it first traded two and a half years ago.

The gap between the business and the stock has a name, and Uber's management uses it themselves: autonomy. Robotaxis threaten the one relationship Uber's model is built on, the one with the human driver. AI assistants threaten the other one, the relationship with the rider. This article follows the video's argument from the numbers, through the founder story and the divestments that shaped today's Uber, to the robotaxi strategy and the question of what Uber's moat actually is. Prices are anchored to the 12 October 2026 close unless another date is stated.

Hovering in Limbo for Two and a Half Years

Uber's highest close was $100.10 on 6 October 2025. On 12 October 2026 it closed at $71.50, a fall of 28.6%. The stock had already ended the first quarter of 2024 at $76.99, so a shareholder who bought then has lost money over two and a half years of record results. In between, the price touched $60.32 at the end of 2024 and then ran up through 2025 before giving the entire move back.^[1]

Earnings days tell the same story. Since the peak, Uber has reported four times, and the stock fell about 5% on three of those four days. The reported reasons were small misses or soft guidance on profit while the company cut prices and stepped up spending on autonomous vehicles, and in August 2026 a robotaxi spending plan that overshadowed otherwise strong results.^[1]

The P/E Is Tiny, and the Real Multiple Fell Even Further

Uber had no meaningful price-to-earnings ratio until the third quarter of 2023, because trailing earnings were negative. When profits arrived they were small, so the multiple started high: 94 in Q3 2023 and 124 in Q1 2024. By the end of 2024 it had collapsed to 13, and in October 2026 it stood at 15.7, based on trailing diluted earnings of $4.55 per share.^[2]

That headline figure flatters Uber. GAAP earnings since late 2024 include large one-off items. In Q4 2024, net income was $6,883 million against operating income of $770 million, helped by a $6.4 billion tax valuation release. In Q3 2025, net income was $6,626 million against operating income of $1,113 million, helped by a $4.9 billion tax release and a $1.5 billion revaluation gain. Q2 2026 included another $1.6 billion revaluation gain.^[2]

A cleaner measure divides the share price by trailing operating income per share, which ignores tax releases and investment revaluations. Because operating income is pre-tax, the level is not comparable with a normal P/E, but the trend is unambiguous. It fell from 104 in Q1 2024 to 22 in October 2026.

The Good: Cash, Growth and Members

Financial Standing

Uber's free cash flow was $390 million in 2022, $3.36 billion in 2023, $6.90 billion in 2024 and $9.76 billion in 2025. The 2025 figure is 42% higher than 2024 and 25 times the 2022 level. In the second quarter of 2026, trailing twelve-month free cash flow passed $10 billion for the first time.^[2][3]

The balance sheet changed direction in 2026. Net debt, meaning total debt minus cash and short-term investments, was $2.89 billion at the end of 2025. By the end of June 2026 it was $7.33 billion, 2.5 times higher. The reason is Delivery Hero: Uber deployed about $4 billion in the second quarter, mostly on market purchases of Delivery Hero stock, and paused its buybacks. Net debt does not count the $8.8 billion of non-current investments Uber still holds, including its stakes in Aurora and Lime.^[3]

Demand underneath the cash flow keeps compounding. Uber had 202 million monthly active platform consumers in Q4 2025, 18% more than the 171 million a year earlier, and 208 million in Q2 2026. Uber handled 3.87 billion trips in that quarter alone; it counts both rides and deliveries as trips.^[3]

Revenue Growth

Revenue was $52.0 billion in 2025, 18% more than the $44.0 billion of 2024 and four times the $13.0 billion of 2019 on the restated basis Uber now uses. Gross bookings, the total value of rides, deliveries and freight on the platform, reached $193.5 billion, 19% more than the $162.8 billion of 2024.^[2]

Growth is not fading at this size. Q2 2026 gross bookings were $58.0 billion, 24% higher than Q2 2025 as reported and 22% at constant currency, the fourth quarter in a row above 20%. Uber reported this while the stock was falling.^[3]

Business Units

Mobility, the rides business, produced $29.7 billion of revenue in 2025, 57% of the total and 18% more than 2024. Delivery, which includes Uber Eats and grocery, produced $17.2 billion, 33% of the total, growing 25%. Freight, Uber's trucking marketplace, produced $5.1 billion, 10% of the total, and was the only division that shrank, by 1%. Delivery's share of revenue rose from 31% to 33% in one year while freight's fell from 12% to 10%.^[2]

Delivery is not a side business. Uber Eats was doing under $1 billion of gross bookings when Dara Khosrowshahi joined in 2017; by his account it does over $100 billion now. Uber's President Andrew Macdonald puts it more sharply: Uber has "basically an equally sized food delivery business that's in market leading positions in most of our markets around the world. We get almost zero credit for that."^[5][6]

Uber One

Uber One is the paid membership across rides, delivery and grocery. It launched in the US in November 2021 at $9.99 a month and bundles discounts, member pricing, priority service and perks. It had 30 million members at the end of 2024, 46 million at the end of 2025 (55% more than a year earlier, by Uber's own count) and passed 50 million in Q1 2026. Members now drive more than half of gross bookings and about two-thirds of delivery bookings.^[4]

The app itself is becoming a profit pool. Uber's advertising business generated "well over $2 billion in annualized revenue, up over 50% YoY" at the end of 2025, and delivery advertising crossed 2% of delivery gross bookings. Cross-selling works too: the CFO says nearly $15 billion of run-rate delivery bookings come from the mobility app.^[2][4]

The Bad: A Low-Margin Business

Margins are growing, but this is still a thin-margin business. Dara Khosrowshahi says so himself: "$10 billion in free cash flow, which is great, but it's on well over 10 billion trips a year. So these are, we are not a high margin business."^[5]

Operating margin, income from operations divided by revenue, was 3.0% in 2023, 6.4% in 2024 and 10.7% in 2025. Quarterly margins reached the low teens in 2026, but part of that rise is accounting: since January 2026 a UK VAT change means certain payments to drivers reduce revenue instead of counting as costs, which lowered reported revenue by $1.1 billion in Q2 2026 alone without changing profit.^[2][3]

Uber reached its first full year of operating profit in 2023, with $1.11 billion, after a $1.83 billion operating loss in 2022. Keeping that profitability in a competitive market, while investing in a technology that does not make money yet, is the hard part. A thin margin leaves less room for mistakes in exactly the area where Uber now has to spend.^[2]

The Threat: Robotaxis

Margins are already low, and robotaxis add pressure on top. Management does not hide it. "I expect that AV won't be profitable for a few years going forward," Dara said on the Q3 2025 call. On the Q2 2026 call, CFO Balaji Krishnamurthy added that as Uber gets closer to deployment and scale-out, "there will be a P&L impact." Neither has put a number on how much margin robotaxis will cost.^[4]

The Word That Took Over the Earnings Calls

Counting "AV", "autonomous", "autonomy", "self-driving" and "robotaxi" across Uber's last six earnings calls shows how the topic took over. The Q1 2025 call had 18 mentions. The Q4 2025 call had 63, more than three times as many, and it is the call where Dara opened by acknowledging that "there are reasonable questions being asked and a debate being had about what autonomy means for Uber, both in the short term and the long term."^[4]

The Biggest Investor Concern

The market's fear is easiest to see in one comparison. Waymo, Alphabet's robotaxi company, raised $16 billion in February 2026 at a $126 billion post-money valuation. Uber's market value at $71.50 is about $146 billion. Waymo does about 500,000 paid rides a week; Uber handles about 300 million trips a week, including deliveries. That is roughly 600 times the volume, valued at only about 16% more. A private funding round and a public market price are not identical measurements, but the signal is clear: investors pay for owning the driver, not for owning the network.^[7]

Uber's own president frames the stakes in the strongest terms available. Asked about autonomy, Andrew Macdonald said Uber is investing to bring it to its platform, "but if we didn't then it certainly would be existential for our core business."^[6]

Robotaxis Threaten Uber's Future

If Uber cannot take part in robotaxis, it risks being left behind. The pressure arrives from every direction. Tesla launched Cybercab rides in Austin in early September 2026, after a launch event on 3 September, and had more than 300 Cybercabs operating there by October. In the two weeks after the launch, Uber's share price fell 6.6%, from $76.45 to $71.43, according to market coverage at the time.^[8]

China is further along than most investors assume. Baidu's Apollo Go has operated in 28 cities and logged over 350 million autonomous kilometres, more than 240 million of them fully driverless. Since 20 August 2026 it also runs on Uber in Dubai, which shows the double edge: the same companies that threaten Uber are, for now, also its suppliers. Dara himself says that "the Chinese capabilities in terms of manufacturing both in terms of quality and cost at this point is unrivaled."^[9][5]

The Travis Affair: Losing a Visionary and Control

To understand why Uber approaches robotaxis the way it does, you have to go back to 2017. Travis Kalanick co-founded Uber with Garrett Camp in 2009 and ran it until 2017. On 20 June 2017 he resigned as CEO after five major investors, including Benchmark, wrote a letter titled "Moving Uber Forward" recommending that he step aside immediately. Benchmark, an early investor, then sued him.^[10]

Early investor Shervin Pishevar later alleged that Benchmark's Bill Gurley "hired some private investigators, had weekly meetings about how do we get rid of Travis starting in January of 2017." That is Pishevar's account, not independently verified history. The episode was turned into a TV series, Showtime's Super Pumped: The Battle for Uber, in 2022.^[10]

Dara Khosrowshahi took over as CEO in 2017. Asked about his first days, he describes "complete chaos," and says "the board was fighting for control." His job was to stabilize a company that, in his words, had "lost trust with stakeholders, with regulators, with the public." He did that. But the insiders from the founder era see a cost. Emil Michael, Uber's former Chief Business Officer and Kalanick's right-hand man, said in 2022 that "what Dara does at Uber, what a company like Verizon does, they're all financially engineered, they're not innovating." Shervin Pishevar calls Dara "the great stabilizer."^[5][11][10]

Control changed structurally too. Before 2017, Uber's Class B shares carried 10 votes each, and Kalanick owned about a third of them. The board approved one share, one vote in October 2017 as a condition of SoftBank's investment, and the change took effect when that deal closed on 18 January 2018. Since then no Uber CEO has had founder-style control. The argument here is that this matters most in exactly the moment Uber is in now: a founder with voting control, as Mark Zuckerberg has at Meta, can make a bet the market punishes. A hired CEO answering to one share, one vote finds that much harder, and disruption is when such bets are needed.^[10]

Is Uber an E-Hailing or an Everything Company?

So far Uber is largely a road business: rides and delivery are its core. It has begun to step outside that, most visibly with hotel bookings through Expedia, launched in April 2026, but these are early moves.^[24]

The delivery business is massive, but in Uber's home market it is the challenger. "We are not number one in the US," Andrew Macdonald says of food delivery, where DoorDash leads, and he describes Uber as playing "the challenger role, which we relish."^[6]

Acquisitions: Mostly Delivery, Never Self-Driving

Uber's acquisitions since its 2019 IPO show where it chose to place its bets. It bought Careem, the Middle East ride-hailing company, for $3.1 billion (closed January 2020); a controlling stake in Cornershop, a grocery delivery company (2020); Postmates, US delivery, for about $2.65 billion in stock (2020); Drizly, alcohol delivery, for about $1.1 billion (2021); and Transplace, freight logistics, for about $2.25 billion (2021).^[12]

The pattern continued. A $950 million deal for Foodpanda Taiwan was blocked by Taiwan's Fair Trade Commission in December 2024 and terminated in March 2025. Uber bought 85% of Trendyol GO in Türkiye for about $700 million (2025), Getir's food delivery business plus a minority stake in its grocery arm for about $465 million (closed July 2026), SpotHero, a parking app, for $617 million (2026), and agreed to buy ezCater, business catering, for $2.3 billion (announced 6 October 2026). The largest by far is Delivery Hero, at about $14.8 billion in equity value (€41.50 per share, or $13.7 billion net of the stake Uber already holds), still pending with closing expected in the second half of 2027.^[12]

None of these is a self-driving company.

Divestments: Selling the Future

Uber's sales tell the other half of the story. In 2020 it sold Uber Eats India to Zomato and its European freight business to sennder. In May 2020 it folded JUMP, its e-bike and e-scooter business, into Lime in exchange for about 16% of Lime, invested an $85 million convertible note and booked a $27 million loss. Lime went public in June 2026, and Uber owns 22.87% of it.^[13]

Elevate, Uber's flying-taxi unit, went to Joby Aviation in a deal that closed in January 2021. Uber now holds under 5% of Joby, and "Uber Air powered by Joby" is expected to start carrying passengers in Dubai in 2026. Most importantly, Uber sold ATG, the Advanced Technologies Group, its own self-driving unit, to Aurora. The deal closed in January 2021. Uber booked a $1.6 billion gain, took about 26% of Aurora on a fully diluted basis and invested another $400 million in cash.^[13]

Uber has since been turning that stake into cash. In three block sales in June, August and September 2026, it sold 168.9 million Aurora shares for about $1.13 billion. It still holds 157.1 million shares, about 9.2% of Aurora's Class A stock.^[13]

Why Sell the Self-Driving Group? A Fatal Crash and a Cash Crisis

On 18 March 2018 an Uber self-driving test vehicle, a modified Volvo XC90, struck and killed a pedestrian in Tempe, Arizona. The US National Transportation Safety Board found that the safety operator had been distracted by her phone, and also cited Uber's inadequate safety culture. Asked years later whether the death changed his mind, Dara said: "I think we were leaning that way," and called it "certainly a very, very important factor in our determination."^[14]

Money mattered as much. "My job is to make sure the company survives," Dara told Stratechery; Uber had gone "from losing $2 billion a year to losing $4 billion a year, which is pretty scary." His conclusion: "Hardware was not our superpower." Andrew Macdonald adds the pandemic: when ATG was sold, "our mobility business had lost 84% of our topline in 3 weeks." And on the technology: "we did not believe we were leading in autonomy at the time."^[14][6]

Out of necessity, then: bad at hardware and burning cash it could not afford. Andrew does not call it a regret. Asked to imagine that ATG had become "one of the leading autonomous players globally and we completely control our destiny," he answered: "Yeah, I think that would be a good thing for us," while warning that it would be "rose-colored glasses" to assume staying in would have made Uber the leader.^[6]

Uber's overriding concern at the time was profitability, which is why it divested so much and pushed for market-share acquisitions instead. CFO Nelson Chai summed up 2020: "We made some big moves this year, acquiring businesses like Cornershop and Postmates while divesting others like ATG and Jump. These decisions have resulted in a much more focused and ultimately stronger company."^[13]

Robotaxis Are All the Rage Now

Today robotaxis are a rounding error inside Uber. In August 2026 Dara said autonomous vehicles complete "hundreds of thousands of trips per week" against about 300 million in total: "It's even less than 0.5% of our overall trip volume."^[4]

Uber has two answers to why that is not a problem. First, utilization: Uber says AVs on its platform complete 30% more trips per vehicle per day than on standalone first-party platforms, based on publicly available data. Second, geography: CFO Balaji Krishnamurthy says "nearly 75% of our US profits" come from markets outside the largest metros, which are "unlikely to be addressed by AVs for a long time to come." Sixty percent of mobility gross bookings are outside the US altogether. Dara's own timeline is long: "You can imagine the majority of our trips being fulfilled by robots of some kind. Probably not 10 years from now, but you go 15 to 20 years from now, you're going to start getting there."^[4][15]

Uber's Robotaxi Investments: Better Late Than Never

Uber expects to commit over $10 billion to autonomy over the coming years. The money comes in two forms. The first is equity in self-driving software companies, tied to milestones; Uber says that for every dollar it invested, its partners raised another $2.50 from other investors. The second is balance-sheet support: fleet operations, real estate such as depots and charging, and purchase guarantees that let carmakers ramp production. Uber works with more than 30 autonomous partners across mobility and delivery.^[3][4]

The supply side is measured in vehicles. Partners have committed about 120,000 vehicles to the Uber network over the coming years; the CFO calls them "120,000 vehicle commitments that we are looking to deliver over the next few years." They are partners' vehicles, not 120,000 cars Uber buys, although Uber backs some with purchase guarantees. Uber also put $100 million into Atoms, Travis Kalanick's new company, which describes itself as "an industrial software company" with "no plans to enter the saturated robotaxi market," even though the Financial Times reports it is building robotaxi technology and has held preliminary talks with Uber.^[4][21]

Loose Partnerships: Mostly Non-Exclusive

Uber's partnerships are many, and most of them are loose. NVIDIA plans to scale robotaxis built on its platform on Uber to 100,000 vehicles over time starting in 2027, with Los Angeles and San Francisco first and 28 cities by 2028. But Lyft, Bolt and Grab also build on NVIDIA's Hyperion platform. Stellantis committed an initial 5,000 vehicles on NVIDIA's platform.^[16]

Waymo is the clearest warning. Its service in Austin and Atlanta is exclusive to Uber only until January 2028, when Waymo plans to launch its own app in both cities. In Phoenix, Waymo and Uber parted ways on 29 June 2026; Uber called it "an intentionally limited deployment" of "just over a dozen vehicles."^[17]

The firmer commitments come with Uber's money attached. Rivian will supply 10,000 robotaxis in a first phase and at least 40,000 more if second-phase pricing is agreed, and cannot sell Level 4 vehicles to Uber's direct competitors during an exclusivity period. Lucid, running Nuro's driving software, has committed at least 35,000 vehicles. Waabi's first 25,000 passenger vehicles are exclusive to Uber.^[18][4]

Inside Uber: The AV Lab

Uber launched AV Labs in January 2026, led by VP Danny Guo, to collect rideshare driving data with robotaxi-grade sensors. Uber planned to put up to 500 data-collection vehicles on the road during 2026; its Autonomous Solutions announcement refers to a broader data fleet of "thousands of specially-equipped vehicles across dozens of cities," so public counts differ. Uber plans to run partners' driving software in shadow mode in these cars and flag every point where the software would have acted differently from the human driver.^[19][20]

The lab also publishes research that goes beyond data. GeoWAM, a 2026 paper listing Uber AV Labs as an affiliation, outputs the trajectory the car should drive, which is the core of a driving policy. Its authors also wrote UniDrive-WM and VLGA, which do the same. An Uber job posting for AV Labs asks a senior engineer to "lead technical direction for end-to-end AV foundation models."^[19]

Research papers are not a deployed robotaxi, and nothing in the public record shows a production Uber driving system. But the work exists, and Uber does not present it as a driver.

Uber Autonomous Solutions: The Waymo Approach Without the Driver

In February 2026 Uber launched Uber Autonomous Solutions, its package for running partners' robotaxis. It includes an in-car screen and controls built to work across hardware setups, first used in the Nuro-Lucid robotaxi; AV Mission Control, "a real-time view of every vehicle" with remote assistance; a "command authority system that strictly arbitrates who can issue commands" to the car; routing and pickup guidance from Uber's trip data; field teams for cleaning, towing and lost items; and depots, insurance and financing. Dara's stated goal is "to become the world's leading commercialization platform for autonomous vehicles."^[20][3]

This is Waymo's operating approach. Waymo riders get help through an in-car screen, Waymo's fleet response agents answer the car's questions remotely, and Roadside Assistance retrieves stuck vehicles. What Uber does not copy is the part that makes Waymo valuable: Waymo designs its driver "from the silicon up," while Uber's package runs on someone else's driving software.^[20]

The Great Hedge

Uber is in a double bind. If it goes all in on its own robotaxis, it loses its partners. Dara explains why Uber left self-driving in the first place: AV developers "didn't want to partner up with a platform that was also a competitor." If it skips robotaxis, it threatens its core business. "Either we had to go vertical or we had to have the platform play," he says. "We couldn't have our cake and eat it too."^[14]

So Uber does both, carefully, hedging every bet. It develops its AV lab, collecting data and training models. It lets partners test their systems in shadow mode in its cars. It holds stakes in Aurora, Lucid, Waabi and Atoms, and works with more than 30 partners. What it does not do is commit to either option. The video's verdict is that this is the risk: too cautious to commit to either path, Uber does neither properly, and many small hedges are not one large bet.

All Hail the Moat

What is Uber's unique moat? Management answers with distribution. "Our distribution, our brand, can win on a long-term basis," Dara said on the Q2 2026 call. Andrew Macdonald: "I think in the end distribution wins." Management also frames the moat as the user experience, and Andrew puts the goal plainly: "I want people to start at the Uber app for the Uber experience."^[4][6]

Distribution is critical in the age of AI. But it is not a moat if a competitor already has more of it. ChatGPT had 1.2 billion weekly users by Sam Altman's count at OpenAI DevDay on 29 September 2026. Uber had 208 million monthly users. That is 5.8 times as many people, and the real gap is larger, because a weekly count is always smaller than a monthly one. If AI apps add booking interfaces for robotaxi fleets, Uber's core business can be threatened almost overnight.^[22][3]

So distribution cannot be Uber's moat.

Holding the Agents at Arm's Length

For now, Uber keeps AI assistants at arm's length to retain control. "We have a great market position, so we're able to kind of often dictate the terms of trade in those discussions," Dara says of talks with third-party agents. In practice, Uber's app inside ChatGPT lets US riders "plan a trip within the ChatGPT app and then complete the ride request and payment process in the Uber app." Claude has had an Uber connector since April 2026 that also hands booking back to Uber's app.^[4][22]

But there are cracks already. On the Galaxy S26 and Pixel 10, Google's Gemini can operate rideshare apps including Uber directly on the screen, like a human would, in a US and Korea beta. It stops only at the final tap: the user still presses the order button. No connector, no terms of trade.^[22]

Uber says it avoids aggregation, but aggregation looks inevitable. Andrew is explicit about what he does not want: "I don't want to be aggregated on price." Yet an agent that can operate apps can open Uber, Lyft and Waymo one after another and compare prices without anyone's consent. That is aggregation, whether Uber agrees to it or not.^[6]

Drowning in Its Own Moat

Uber seems uncertain about its moat, and the moat appears to be eroding quickly. Uber has its own AI features, and they are real: Cart Assistant builds a grocery basket from a photo, an AI assistant for drivers built on OpenAI's models advises on earnings, and voice booking on OpenAI's Realtime API is rolling out. None of them stops aggregation, and management knows it. "I think the inbound interaction is going to move more and more away from apps," Dara says, even if people keep using apps.^[5][22]

Uber's moat today rests on three things: driver relationships, a quality app and distribution. The driver relationships are large: 10 million drivers and couriers. They are also the part robotaxis remove. Humans need an interface; robots can connect without one. A robotaxi has no loyalty and no habit. Where its fleet contract allows it, it can be dispatched into ChatGPT or any other agent through an API, no app needed, and switch to wherever it earns the most. People will not use only Uber's app either; they will use ChatGPT or similar tools when that is more convenient.^[4]

Zuckerberg's story is the template. He "didn't really use Instacart that much before," and now his weekend baking ingredients arrive through Instacart because his agent orders them. He even asks whether that account is "really for me, or is it for the Muse?" Instacart gained a regular customer, but the relationship now runs through an agent Instacart does not control. Uber will have to go the same way, and it will need a new moat when it does.^[23]

Headless access will not be a choice; it will be the default. Uber's answer is Dara's travel analogy. In travel there were fears of "metasearch and this layer above the travel companies," but as the industry consolidated, the value went to "the Expedias, the Airbnbs and the Booking.com." "We've kinda seen this movie before," he says, and he expects "the majority of our transactions come direct." Andrew is more cautious: "that sort of worst fear hasn't played out yet. And I'm not saying it won't."^[4][6]

Uber's Future Strategy, or What It Should Be

There are two threats, and they need different answers.

For robotaxis, Uber has to do what ChatGPT cannot. It must move from a relationship with drivers to a relationship with robotaxis, which means controlling the self-driving experience, and it has to go further and become a player in self-driving itself. The video names three parts of that moat.

Driving data. Uber can lead here. It facilitated about 14.8 billion trips in the last four reported quarters. Those are ride and delivery transactions, not sensor recordings, but no other company has comparable coverage of real-world trips across more than 70 countries, and with sensors on enough cars that coverage could become driving data.^[3]

Self-driving software. The video's argument is that owning the driving model is the only way to own the experience and to use that data fully. Without it, the data improves someone else's car.

The harness. The implementation layer that defines the ride: routing, best practices, orchestration and tools. It is the app, but also how Uber orchestrates autonomous vehicles, its own or others', so that every car performs best on Uber. Just as an AI model performs differently depending on the harness around it, Uber can build the leading harness for robotaxi fleets. If AVs perform best under Uber's orchestration, Uber wins.

But Uber Is Giving It Away

Uber is not leaning into this. It says it does not build the driver, and its dedicated data fleet is small next to its 10 million drivers and couriers: up to 500 cars planned for 2026 in AV Labs, against millions of driver cars that Uber's CTO hopes to fit with sensors only eventually. It sold its self-driving unit and bets on partners to solve the problem for it. And it shares the data. Partners receive Uber's processed driving data, free for now. Uber's CTO describes the aim as to "democratize this data."^[19]

Outlook

Uber reports its third-quarter 2026 results on 4 November 2026. The robotaxi effort it reports on has fewer leaders than at the start of the year. Guy Peterson, head of business development for autonomous partnerships, left for Rivian's spinoff Also; Lando Juarez, head of legal for the autonomous unit, left for Bedrock Robotics; and Sarfraz Maredia, global head of autonomous mobility and delivery for more than 12 years, leaves on 16 November 2026. His job is being split between Samarth Kejriwal for driverless operations and Aaron Emrich for autonomous delivery rather than filled.^[25]

The next milestones are still a year away: NVIDIA-based robotaxis on Uber in Los Angeles and San Francisco in 2027, and Delivery Hero's closing in the second half of 2027. Insiders and large investors are buying. Dara bought $10 million of Uber stock on 10 September 2026, 141,000 shares at a weighted average of $70.96, and Pershing Square holds 34.3 million shares, its largest position. Analysts' consensus target of about $101 to $106 is 41% to 48% above the 12 October close.^[26][16]

Uber has time. The business is growing, and the more it grows, the more it can invest in the future. Nothing here suggests an immediate danger: robotaxis are under 0.5% of trips, and Dara imagines a majority of robotic trips only in 15 to 20 years. But time is not the same as certainty.