Waymo Treats Uber Partnership Tactically

The Phoenix split shows that robotaxi aggregators are not durable moats, they are distribution contracts that can be repriced, narrowed, or ended market by market. Waymo uses Uber where the extra rider flow is worth sharing economics and control, but Phoenix shows it can also route demand through Waymo One and keep the rider relationship itself. That makes Uber useful for launch speed in Austin and Atlanta, not indispensable to Waymo’s core network.1 2 3 4 5

• Phoenix is now a direct Waymo market, while Uber remains the exclusive booking channel for Waymo rides in Austin and Atlanta. Waymo’s own partner page now lists Uber only for those two cities, and its Phoenix page points riders to Waymo service directly. That city by city split makes the partnership look tactical, not permanent.1 4 5
• Uber’s strategy is to be the demand layer for many AV fleets, not just Waymo. Internal research shows Uber has signed 25 plus AV partnerships and takes a commission on rides regardless of whose vehicle shows up. If one AV operator leaves a city, Uber can try to plug in another, which is why the Phoenix loss matters less to Uber than to a fleet operator that depends on aggregator volume.2 6
• Waymo has enough scale to walk away where it already has density. Waymo was doing about 500,000 weekly paid rides across 11 cities by April 2026, with Phoenix as its oldest public market and about 100,000 weekly rides in earlier 2026 research. In mature zones, Waymo can justify owning pickup, pricing, and rider data instead of paying an aggregator toll.2 3 6

Going forward, aggregator deals will likely become launch tools rather than long term anchors. Waymo should keep using Uber to fill cars fast in new cities, then push more demand onto Waymo One once a market is dense enough. The winners in robotaxi will be the operators that can switch between aggregator demand and direct demand without stalling utilization.2 3 4 6