Uber's AI agent numbers from one of its biggest weeks: 320 million trips and orders, 4 rides
Four. Not four thousand. Uber said on its Q3 earnings call that in Halloween week, one of its largest at about 320 million trips and orders across Uber and Uber Eats, AI agents fulfilled "a grand total of 4 trips" on the Rides side. Uber Eats did better, with around 500 orders. CEO Dara Khosrowshahi's summary: "very, very early and very limited traction".
This is rare hard data in a category that runs almost entirely on demos. DoorDash said last week it has received only a few hundred agent orders so far. Compare that with the money: Instinct raised at a $10B valuation in September, and Meta says Muse tops 10 million users with shopping among its top activities.
Khosrowshahi offered the most useful explanation without quite meaning to. Booking a ride takes 50 seconds in the app, and reordering a favorite meal takes 90. When the existing flow is already that good, the agent has no time to save, and you still have to trust it with your card. The agents that win consumers won't be replacing a 50-second tap. They'll go after the jobs that take an afternoon. Uber can presumably only count what it can attribute to agents, so an agent driving a phone or browser may look like a normal customer here. Treat 4 as a floor. It's still a tiny floor.
Khosrowshahi still expects a net positive for demand and says "every single AI system company is talking to us". And the same week Anthropic made a telling choice in the other direction. Personal Finance in Claude, in US beta for Pro, Max and Ultra, connects bank and card accounts via Plaid with read-only access and "can't move, manage, or invest your money". Up to two years of transactions, chats in the hub not used for training, and Claude tells you to check important figures with your bank. So a finance agent, where mistakes cost the most, shipped with the action button removed.
That's where consumer agents actually stand in late 2026. Plenty of reading and advising, almost no doing. The people betting on an agentic economy might want those two facts side by side: the capability curve is steep, and the Uber number is 4. The gap between them is trust, payment rails and habit, none of which a better model fixes on its own. Sources: https://www.fool.com/earnings/call-transcripts/2026/11/04/uber-uber-q3-2026-earnings-call-transcript/ https://techcrunch.com/2026/11/04/almost-no-one-uses-ai-agents-to-book-their-uber-company-says/ https://claude.com/resources/articles/introducing-personal-finance-in-claude
← Back to all articles
This is rare hard data in a category that runs almost entirely on demos. DoorDash said last week it has received only a few hundred agent orders so far. Compare that with the money: Instinct raised at a $10B valuation in September, and Meta says Muse tops 10 million users with shopping among its top activities.
Khosrowshahi offered the most useful explanation without quite meaning to. Booking a ride takes 50 seconds in the app, and reordering a favorite meal takes 90. When the existing flow is already that good, the agent has no time to save, and you still have to trust it with your card. The agents that win consumers won't be replacing a 50-second tap. They'll go after the jobs that take an afternoon. Uber can presumably only count what it can attribute to agents, so an agent driving a phone or browser may look like a normal customer here. Treat 4 as a floor. It's still a tiny floor.
Khosrowshahi still expects a net positive for demand and says "every single AI system company is talking to us". And the same week Anthropic made a telling choice in the other direction. Personal Finance in Claude, in US beta for Pro, Max and Ultra, connects bank and card accounts via Plaid with read-only access and "can't move, manage, or invest your money". Up to two years of transactions, chats in the hub not used for training, and Claude tells you to check important figures with your bank. So a finance agent, where mistakes cost the most, shipped with the action button removed.
That's where consumer agents actually stand in late 2026. Plenty of reading and advising, almost no doing. The people betting on an agentic economy might want those two facts side by side: the capability curve is steep, and the Uber number is 4. The gap between them is trust, payment rails and habit, none of which a better model fixes on its own. Sources: https://www.fool.com/earnings/call-transcripts/2026/11/04/uber-uber-q3-2026-earnings-call-transcript/ https://techcrunch.com/2026/11/04/almost-no-one-uses-ai-agents-to-book-their-uber-company-says/ https://claude.com/resources/articles/introducing-personal-finance-in-claude
Comments