AINEWS
CEO Dara Khosrowshahi frames Uber's autonomous vehicle push as a reaction to AI industry backlash. The seven to twenty eight city ramp in two years tells a faster story.
Uber is committing more than $10 billion of capital to autonomous vehicles and planning a rapid expansion from 7 live cities today to as many as 15 by year's end and 28 by the end of 2028, according to CEO Dara Khosrowshahi's Q2 2026 prepared remarks. On the same call, he also pledged that the rollout will be "more deliberate" than the AI industry's recent sprint, a line that sounds like caution but functions as a licensing strategy.
Robotaxis, for readers who have not seen the term, are driverless ride-hail vehicles summoned through an app, with no human behind the wheel. Uber is not building them. It is building the platform that deploys fleets operated by more than 30 partner companies, including Wayve, which secured a permit in July to begin a joint robotaxi service in London.
The "more deliberate" framing is Khosrowshahi's response to a specific pressure. He named it directly: the public backlash that has hit AI labs and data-center operators over the past two years, including the friction over nondisclosure agreements that has followed high-profile incidents. Uber, he argued on the earnings call recap, does not want to inherit that reception. The data-center comparison is the key. Khosrowshahi is treating AV city permits the way utility-scale AI projects now treat community engagement: as a precondition for the next permit, not an afterthought.
The roadmap makes the "deliberate" claim hard to read at face value. Going from 7 cities to 15 by year-end and 28 by the end of 2028 is roughly a fourfold expansion in under 30 months, faster than Uber's ride-hail footprint grew in any comparable window. A truly cautious rollout would not triple and then nearly double its live count in two and a half years. Khosrowshahi's positioning is that the expansion is not slow; it is sequenced to stay ahead of the backlash curve.
Waymo is the public robotaxi incumbent, and industry analysis treats its lead as durable for the next 18 months. Uber's alternative is not to out-build Waymo; Waymo builds its own cars and runs its own service. Uber's alternative is to host every other AV partner on a single platform, taking a cut of rides and spreading the regulatory load across dozens of operators. As one analysis frames it, Uber is winning the orchestration layer even when it loses the headline city.
According to TechCrunch's reporting, Uber has been pressing state and federal regulators for rules that would favor platform-based AV deployment over vertically integrated fleets, putting the company on a collision course with Waymo. The "more deliberate" line is consistent with that posture. Slow, well-publicized rollouts in friendly cities build a record of operational compliance that platform-hostile regulators cannot easily discount.
What the dual pressure looks like in practice: Khosrowshahi is borrowing the language of the AI backlash to license his catch-up against Waymo. The license is operational goodwill, city by city, in jurisdictions that have watched data-center fights and decided they do not want the same fight over their streets. Uber is offering to be the operator that did not happen with the last wave, and the $10 billion is the deposit.
The test is whether the "deliberate" language survives contact with Waymo's actual pace. If Waymo stumbles on a city permit or a service incident, Uber's hand is obvious: accelerate. If Waymo keeps its lead clean, the sequencing will hold. The first checkpoint is the 15-city target by December 2026, with the prepared remarks putting 28 cities on the clock for the end of 2028.