Uber is sorting the world into robotaxi cities and everything else

On September 2, Uber's app went dark in Nigeria and Uganda mid-trip, and the next day its first commercial robotaxis began picking up passengers in London ^[1][2]. The two events look like a coincidence of timing. They are not — they are two faces of one logic, and Uber's chief executive has already said what that logic is. A few months earlier, Dara Khosrowshahi conceded that robotaxis may not be economically viable in India for at least five years ^[3]. That admission is the sorting mechanism. If the economics of self-driving don't work in India, they don't work in Lagos or Kampala either. The London launch the same day is the other face of the same decision: capital is flowing to cities where autonomous vehicles can plausibly operate, and away from the places where they can't. The pattern runs across the map. London, Las Vegas, Dubai, Zagreb, Munich — wealthy cities with the regulatory frameworks and road networks for robotaxis — are getting the investment ^[4][5][6]. Uber has committed more than $10 billion to autonomous fleets and equity stakes, including 35,000 self-driving cars through Lucid Motors ^[7]. The company says the savings from the exits will be redirected toward ride-sharing, delivery, and robotaxis ^[8]. Khosrowshahi made the framing explicit.

We are removing layers, simplifying team structures, refining our global location strategy, and focusing our people and investments against the biggest opportunities ahead of us. — Dara Khosrowshahi

But the honest version has a wrinkle. Uber did not sacrifice a thriving Nigerian business for its robotaxi future. Bolt had already passed Uber as the most-downloaded mobility app in Nigeria, and Uber's contribution to the economy — N34 billion in 2023 — was real but not a position of strength ^[9]. The exit is co-caused: a market Uber was losing to a local competitor, and a market with no self-driving future. The AV logic explains why Uber didn't fight to keep it; it doesn't explain why it was losing in the first place. Nor is this a retreat from global scale. In July, Uber agreed to buy Delivery Hero for $14.8 billion, nearly doubling the number of markets where it offers delivery ^[10]. In May it announced its first India data centre and two tech centers ^[3]. It is still expanding its footprint — just not in places where its role as a platform intermediary has no autonomous future. The sorting is about what kind of business Uber is in each market, not whether it is in the market at all. That is where the bet gets precarious. The partners Uber needs to survive the transition are the same companies building the capability to cut it out. Waymo has told Uber it will launch its own app in Austin and Atlanta by 2028, running alongside Uber until the contract expires ^[11]. Tesla runs its own robotaxi service in Austin, Dallas, and Houston ^[12]. Baidu's Apollo Go is expanding to Europe and the UAE ^[13]. Each operates its own app and fleet — the very thing Uber's aggregator model was supposed to make unnecessary. The fractures are already showing beyond Waymo. Uber divested its stake in Serve Robotics in August amid operational clashes ^[14]. And the spending meant to buy Uber a place in this future is running hot: the company burned through its entire 2026 AI budget in four months ^[15]. Its chief operating officer was blunt about what that money bought.

On AI, we are very early, but what we are seeing is that we are able to cost-efficiently deliver some productivity lifts with developers. — Balaji Krishnamurthy

The geographic sorting answers one question — where Uber will be — and leaves the harder one open. Even in the cities it is keeping, the partners it needs are the ones preparing to replace it. An analyst put the demand plainly after the Waymo news ^[16].

A clean mobility print would stabilize sentiment, but a durable recovery requires Uber to address the structural question head-on. — Dave Mazza

Uber has drawn its line through the map. What it hasn't shown is that it can hold the cities on the right side of it.