Uber’s AV Story Grows Up—Just as Its AI Agent Story Shrinks to Four Rides
Q3 delivered growth and cash, but the real news is a pivot from AV land-grab to fleet economics, a driver-supply pinch in Brazil and India, and a reality check on agentic commerce.
UBER · Earnings Call · 2026-11-04
The quarter that didn’t match the tape
Uber’s Q3 26 call had the familiar top line: 215 million monthly active platform consumers, a fifth consecutive quarter of 20%+ gross bookings growth, 35% non-GAAP EPS growth, and record trailing 12-month free cash flow of $10.4 billion. Delivery gross bookings rose 23%, and grocery/retail reached a $15 billion run rate growing 40%. Operating margin was 13.3%, up 1.9pp year over year. Yet the stock sits roughly 30% below its October 2025 peak and has lost 3.2% over the last 90 trading days. At 18x trailing free cash flow, the gap between operating momentum and share price is the story: the market is no longer paying for growth alone. It wants proof that the next leg—AVs, grocery, AI distribution—can be financed without eroding the marketplace.The supply pinch is new
The first crack is Mobility. Dara Khosrowshahi was unusually direct: “Mobility continued to see healthy growth but came in below our expectations as a result of driver supply not keeping pace with strong consumer demand, in several international markets.” — Dara Khosrowshahi, Chief Executive Officer · 2026-11-04 That is a different problem from the one management described in Q2 26, when the emphasis was U.S. Mobility acceleration, insurance savings, and sparse-market expansion. Now the constraint is supply, not demand. Balaji Krishnamurthy pointed to “elevated fuel price impact in markets like Brazil and India in particular, quite meaningfully impacting the supply situation.” — Balaji Krishnamurthy, Chief Financial Officer · 2026-11-04 Immigration dynamics and peak-season demand are compounding it. Uber is reallocating rider incentives toward driver incentives—a targeted supply investments program that is already showing October improvement. But it is a take-rate risk: “we would have some color on just the Q4 supply investments for mobility that would have an impact on the take rate, but not necessarily from a margin standpoint, anything meaningful.” — Balaji Krishnamurthy, Chief Financial Officer · 2026-11-04 The market may accept that distinction for one quarter. It will not accept it for four.AV: from “15 cities” to “1P deployments” and cost of capital
For several quarters Uber sold AV breadth: partner count, city count, a $10 billion commitment. In Q2 26 Dara said, “we're live in 7 cities, and we're on track to be live in 15 cities by year-end.” — Dara Khosrowshahi, CEO · 2026-08-05 Q3 still lists the partners—Waymo, WeRide, Baidu, Zoox, Motional, Nuro, Pony, Momenta, NVIDIA, Rivian, Wayve—but the language has shifted to utilization and financing. high utilization, 1P deployments, fleet financing, and cost of capital now dominate the company’s own keyword profile. Dara explained why: “residual value of the AV vehicles is not established yet... in the early days, we are gonna take these vehicles on our balance sheets.” — Dara Khosrowshahi, Chief Executive Officer · 2026-11-04 Balaji framed the end state: “our long run expectation is that this should be financed by third party capital, not by Uber's balance sheet.” — Balaji Krishnamurthy, Chief Financial Officer · 2026-11-04 The AV story is growing up—from land-grab to unit economics. The claimed 30% utilization premium versus mature 1P deployments is the core proof point. But taking vehicles onto the balance sheet is why effective net cash fell to $3.1 billion, down 70% year over year. Uber can afford it; the question is whether third-party capital arrives before the balance sheet does too much of the work.Grocery and the dark-store temptation
Delivery is the offset. Grocery and retail are now a $15 billion run rate growing 40%, inside what Dara calls a $10 trillion TAM. The new top company keyword is dark store, and it is not boilerplate. Uber historically positioned itself as partner-first in grocery; now, via Delivery Hero, Getir, and organic experiments in Taiwan, it is selectively moving toward vertical integration. Dara’s rationale: “the benefit of dark stores is that you have a much better handle on inventory... your pick-and-pack success percentages... found rates are much higher.” — Dara Khosrowshahi, Chief Executive Officer · 2026-11-04 Balaji called it “fit-for-purpose,” with a return-on-assets test. That is a real strategic shift in quick commerce, even if management insists it will not own dark stores everywhere. The potential DoorDash-Instacart combination is a backdrop: Dara says it will not affect Uber much, but merchants are already reaching out, worried about dependence on a single platform. Uber’s answer is more selection, more merchant funded promotions—now over $10 billion annualized—and more membership. With roughly 55 million Uber One members, 70% of delivery gross bookings, the flywheel is real.The AI agent reality check
The Q&A produced the most revealing moment. Dara repeated the direct-app defense from Q2, when he said, “we've kind of seen this movie before... we think the majority of those experiences are going to come direct to us.” — Dara Khosrowshahi, Chief Executive Officer · 2026-05-06 On Q3 he added: “I used to work in the travel business, and we were dependent on the Googles of the world. I do not wanna go back there.” — Dara Khosrowshahi, Chief Executive Officer · 2026-11-04 He touted a 50-second mean ride-booking time, 75% single-tap rides, and a cart assistant that roughly doubles basket size. Then Balaji gave actual agent volumes from Uber’s largest week of the year:Dara conceded, “it is very, very early and very limited traction that we are seeing so far.” — Dara Khosrowshahi, Chief Executive Officer · 2026-11-04 That is the reality check on agentic commerce. AI agents for Uber Eats may be strategically important, but today they are a rounding error against 320 million trips. It cuts both ways: less near-term threat to Uber’s direct relationship, but also less near-term upside from the agent channel. The AI hype cycle is not yet in Uber’s P&L.we did just around 320 million trips on Uber and Uber Eats. The volumes coming from AI agents for Uber Eats were around 500. 500 orders total. And for our Rides business, it was a grand total of 4 trips that we have fulfilled from agents.