Waymo in 6 cities and the autonomous driving

Atlanta got Waymo last month. Miami a few weeks before that. That’s six cities now with fully autonomous robotaxis. San Francisco. Phoenix. Los Angeles. Austin. Miami. Atlanta.

The gap between city launches is getting shorter. Phoenix to LA took over a year. LA to Austin, about nine months. Austin to Miami, six. Miami to Atlanta, weeks.

The network effect nobody talks about

Every new city generates driving data that makes the model better everywhere. A construction detour in Atlanta teaches the system something about cones and barriers that helps every Waymo in San Francisco. A rainstorm in Miami generates training data that improves performance in Austin.

More cities, more data. More data, better driving. Better driving, more cities. It’s a flywheel, and it’s spinning faster with each rotation.

This is the same pattern that made Google Maps dominant. More users meant more traffic data meant better routing meant more users. Waymo’s version: more cities mean more driving scenarios mean better models mean more cities.

The quiet part

Six cities. Zero fatal crashes. Millions of autonomous miles. The safety data is compelling. Waymo cars are involved in fewer incidents per mile than human drivers. By a significant margin.

At some point, and I think we’re approaching it, the moral calculation flips. Right now, the question is “is it safe enough to let a robot drive?” Soon the question will be “is it ethical to let a human drive when the robot is statistically safer?”

That’s an uncomfortable question. I know people who love driving. I know people whose livelihood depends on driving. The answer to “should we let safer machines replace less safe humans” has implications far beyond traffic safety.

But the data is the data. And the data is accumulating in six cities simultaneously now.

What’s next

I’m watching for the announcement of city seven. And eight. And ten. The pace of expansion tells you more about the state of the technology than any benchmark or demo video. When a company expands into a new city, they’re betting their reputation and their liability on the technology working. That’s a signal you can trust.

Ten cities by the end of 2025 feels possible. Twenty by the end of 2026 feels ambitious but not unreasonable. And at some number, the question stops being “is Waymo in my city?” and starts being “why isn’t Waymo in my city?”

I think that number is closer than most people expect.

Related thinking: