Waymo reportedly mulling a breakup with Uber

Waymo has decided it no longer needs Uber's distribution. That is the only way to read the Financial Times report that the Alphabet unit is hunting an exit from the ride-hail giant's network in Austin and Atlanta. The contract runs until May 2028. Waymo has already told Uber it will launch on its own app in those cities come January 2028 — alongside the Uber listing, for now. That five-month overlap is not a partnership. It is a managed divorce.

The Phoenix breakup earlier this year was the prototype. TechCrunch broke that story first. Now the template repeats in two more markets. Waymo gave Uber a taste of its fleet, learned the operational rhythms of a ride-hail marketplace, and is now moving to own the customer relationship entirely. Why wouldn't it? Every ride booked through Uber is a rider Waymo cannot re-engage, a data point it does not own, a brand impression it cedes to a competitor that is also building its own autonomy stack.

Uber knows this. Its executives have spent months signaling the split before it arrived. Chief Technology Officer Praveen Neppalli posted video of what he called "scary" Waymo behavior. Chief Executive Dara Khosrowshahi used an earnings call to knock robotaxi conduct in school zones and emergency scenarios — without naming Waymo, though nobody missed the target. These were not safety interventions. They were negotiating leverage aired in public. When a partner starts critiquing your product on earnings calls, the partnership has already curdled.

The regulatory fights tell the deeper story. Waymo and Uber have landed on opposite sides of fresh policy battles over robotaxi rules. That is not accidental. Uber wants regulations that favor fleet-agnostic platforms — rules that let it plug in any autonomy provider, including its own. Waymo wants rules that bless its vertically integrated stack. Each needs the other's lobbying muscle, but neither trusts the other's endgame. So they fight in city halls and state capitals while the contract ticks down.

The timeline reveals Waymo's confidence. January 2028 is aggressive. Building a consumer app that reliably matches riders to driverless cars at scale — with support, pricing, rebalancing, all the invisible machinery Uber perfected over a decade — is not a side project. Waymo would not commit to that date unless it believed its product could carry the full weight of demand generation. The "alongside" period is insurance, not dependence.

Uber's position is more fragile than it admits. The company has spent years promising investors that its network is the moat: plug in any autonomy supplier, keep the rider, take the margin. But the suppliers are learning they don't need the moat. Waymo has the capital, the brand, the tech, and now the operational reps to go direct. So does Tesla, whenever its robotaxi materializes. So might Zoox or any well-funded entrant. Uber's moat only holds if the best supply stays exclusive. Waymo just proved it won't.

The financial terms of the Austin-Atlanta deal never leaked. They don't need to. Whatever revenue split they negotiated, Waymo has decided the lifetime value of a direct rider exceeds it. That calculation changes everything. It turns Uber from a channel into a customer-acquisition cost — one Waymo will stop paying the moment its own app proves it can fill the cars.

Uber will spin this as portfolio management. It will note its own autonomy investments, its partnerships with other AV developers, its diversified stack. All true. But diversification is what you trumpet when you lose the flagship. The flagship was Waymo. Losing it in three markets inside a year is not portfolio management. It is a strategic retreat.

The policy fights will intensify. Waymo will push for rules that reward vertical integration — dedicated pickup zones, proprietary data standards, liability frameworks that assume a single-stack operator. Uber will push for open interfaces, mandated interoperability, platform-neutral permitting. Cities will hear both sides and write rules that reflect whichever lobbyist showed up last. The public loses either way: fragmentation disguised as safety regulation.

Waymo's bet is that the robotaxi market matures like smartphones, not like taxis. Apple didn't need carriers to design the iPhone experience; it needed them only for spectrum. Once the device was strong enough, it dictated terms. Waymo believes its stack — hardware, software, ops, support — is now strong enough to dictate terms. The January 2028 date is the moment it stops asking permission.

Uber's counter-bet is that riders are lazy. They open one app. They won't download a second for a marginally better ride. That held for years in human-driven ride-hail. But robotaxis change the product enough — no driver awkwardness, consistent behavior, predictable pricing — that the switching cost drops. Waymo doesn't need to be ten times better. It needs to be noticeably better, and available in the same minutes. If it clears that bar, the habit breaks.

The next signal to watch is not another market exit. It is Waymo's marketing spend. If it starts buying installs, running referral loops, blanket-covering Austin and Atlanta with "download Waymo One" campaigns before January, the breakup is real and the timeline holds. If it stays quiet, the date slips. But the direction is set. Waymo chose the harder path: own the whole loop. Uber chose the easier story: we're a platform. The market will decide who was right. My money is on the company that just fired its biggest distributor.