Waymo Hit 500,000 Rides Per Week and Autonomous Finally Crossed Into Revenue

Waymo posted a milestone on Saturday. The company crossed 500,000 paid rides per week across its markets, which include Phoenix, San Francisco, Los Angeles, and Austin. Alphabet's 2026 target for Waymo was one million rides per week by end of year. Waymo hit half of that goal less than three months into the year.

Growth is doubling year over year on the paid-rides metric. A year ago, Waymo was at roughly 250,000 weekly rides. Eighteen months ago, it was below 50,000. Whatever you believed about autonomous vehicles as a business case a year ago, the revenue line is making those beliefs increasingly hard to hold.

Zoox announced the same weekend that it was expanding to Austin and Miami, quadrupling its San Francisco service area. Uber announced plans to launch Europe's first robotaxi service with Pony AI and Verne in Zagreb. The autonomous-vehicle category as a whole just had the most active operational weekend in its history.

Why does this matter for operators outside transportation? Because the Waymo milestone is a working template for how any physical-AI category matures. You start with a narrow geographic footprint, a highly instrumented operating area, and a team of humans monitoring remotely. You build trust one neighborhood at a time. You iterate on edge cases until the failure rate is lower than a human driver's. Then growth compounds.

That template applies to a lot of industries that are not transportation. Healthcare AI is in the highly-instrumented-narrow-footprint phase now. Legal AI is in the pre-commercial-demo phase. Finance AI is in the compliance-gated early-commercial phase. Every one of those categories will eventually hit a point where the quality threshold is met, the monitoring overhead drops, and growth switches from linear to compounding. The Waymo curve is what that transition looks like.

Why aren't we talking about this more? Because AVs have been 'two years away' for so long that most operators have stopped paying attention. That is a mistake. The category moved from demo to revenue sometime in the past eighteen months, and most people who write about AI for a living still frame it as speculative. It is not. Half a million paid rides per week is a real business.

The practical takeaway for an operator is about what this tells you about the timeline for every other physical-AI category. If Waymo went from 'ambitious demo' to 'scaling business' in roughly four years of serious operations, the same timeline probably applies to other physical-AI categories once they cross a similar quality threshold. Humanoid robotics, surgical AI, dexterous-manipulation robotics in warehouses. Those are all currently in the narrow-demo phase. The transition to revenue could arrive faster than most industry maps show.

There is also a practical takeaway about defensibility. Waymo's edge over competitors is partly the model. It is mostly the data accumulated from years of real operation in real cities with real customers. Every ride produces telemetry the model can learn from. That data moat compounds faster than any competitor can reproduce it.

The same pattern applies to any AI category where real-world operation generates training data. The companies that get into production early, even at a small scale, accumulate data that becomes hard to catch up to. If your company has a unique operational vantage point, the cost of using it to build a defensible AI product is much lower than it was two years ago. That window is also getting narrower, because your competitors are probably running the same calculation.

The Waymo milestone is not just a Waymo story. It is the whole autonomous-vehicle industry starting to look like a category winner is going to emerge within the next two years, with Waymo positioned as the most likely one. It is also a signal for every other physical-AI category about what maturity looks like and how fast it can arrive once the capability bar is met.