Waymo Added Three Cities and Crossed 4,000 Vehicles

Waymo opened public robotaxi service in Denver, San Diego, and Tampa on September 1, bringing the Alphabet subsidiary to 14 US metropolitan areas. The expansion pushed its fleet of Jaguar I-Pace hatchbacks and the new Ojai minivan past 4,000 vehicles, with roughly 300 of those being Ojais.

Separately, the Nevada Transportation Authority voted unanimously last month to permit Waymo to operate up to 1,000 autonomous vehicles over the next year in Clark County — Las Vegas.

The pace is the story. Waymo entered its first Uber-partnered market in Austin in March 2025. It has added eleven metros since.

Scaling stopped being the hard part

For most of the last decade, the binding constraint on robotaxis was capability. Could the vehicle handle an unprotected left, an occluded pedestrian, a construction zone, rain? That question consumed a decade and something like $30 billion across the industry.

It is now largely answered for Waymo's operating domain, and the constraints that replaced it are ordinary business constraints: vehicle supply, depot buildout, remote assistance staffing, insurance, municipal negotiation, and state permitting. Those are hard, but they are hard in the way logistics is hard, not in the way autonomy was hard.

The Ojai is the clearest evidence of the shift. Adding a minivan is not a capability decision; it is a unit-economics and use-case decision — more seats, more luggage, airport runs, families. You do not diversify your vehicle platform until the driving problem is stable enough that the vehicle is a product question.

The Nevada permit points the same direction. A regulator authorizing up to 1,000 vehicles in a single county is not conducting a safety experiment. It is granting operating scale, and Clark County — dense, tourist-heavy, grid-simple, with enormous point-to-point demand between a strip of hotels and an airport — is close to the ideal robotaxi geography.

The competitive picture just got crowded

For years the robotaxi race was Waymo and a long gap. The gap closed noticeably this quarter.

Tesla now charges for rides in driverless Model Y vehicles in Austin, Dallas, Houston, Miami, Orlando, and Tampa — six markets — and has begun paid Cybercab service in Austin using a two-seat vehicle with no steering wheel, no pedals, and no mirrors. That vehicle drew an NHTSA audit query within hours of launch over how Tesla self-certified it against federal motor vehicle safety standards.

Zoox is expanding into more US markets on its own purpose-built vehicle.

The three companies are running genuinely different strategies, and Tampa is now the city where two of them overlap directly.

Waymo's approach is expensive sensing — lidar, radar, cameras — with heavy mapping and a methodical per-city validation process. High cost per vehicle, high confidence, slow entry, and a regulatory record clean enough that a state authority will approve 1,000 vehicles at once.

Tesla's approach is camera-only on a vehicle it manufactures at enormous scale, with the bet that the marginal cost of adding a city approaches zero once the general driving policy is good enough. Faster expansion, cheaper hardware, and a regulatory posture that generates federal inquiries.

Both bets are now generating revenue in the same country, which means the comparison is finally empirical rather than rhetorical.

What the numbers do and do not say

Fleet size is a weak proxy and worth treating carefully. 4,000 vehicles across 14 metros averages under 300 per market, which in a city the size of Denver or San Diego is a thin service — long waits, restricted geofences, limited hours. "Launched in a city" and "meaningfully serving a city" are different claims, and the gap between them is where most robotaxi coverage goes wrong.

The Nevada permit is the more informative number precisely because it is a ceiling rather than a deployment. 1,000 vehicles authorized in one county is roughly a quarter of Waymo's entire current fleet, which tells you what density Waymo thinks it needs for a market to work economically — and how far its current per-city presence is from that.

Waymo has also said it is planning launches in London and Munich. European entry is a materially harder problem: different regulatory frameworks per country, denser and older street geometry, weaker precedent for autonomous permitting, and none of the state-by-state pathway that has made US expansion tractable.

Why this is the year it compounds

Robotaxi expansion has a reinforcing structure that most technology rollouts do not. Every additional city produces more miles, which produce more edge cases, which improve the driving policy that all cities share. Every additional vehicle amortizes the fixed cost of the software over a larger base. Every clean regulatory record makes the next state's approval easier — which is visibly what happened in Nevada.

That is why the cadence went from one city a year to three cities in a day. It is also why the competitive window for anyone not currently operating is closing: the compounding accrues to whoever is already driving.

What to watch

Utilization, not city count. Waymo reaching meaningful ride density in the metros it has already entered matters more than metro fifteen.

How Tampa resolves. It is the first market where Waymo and Tesla compete head-to-head on price and availability with different sensor philosophies. That comparison is worth more than any benchmark.

Whether the NHTSA Cybercab query goes anywhere. Waymo's expansion runs on regulatory goodwill. A federal action against a competitor's self-certification would reshape the permitting environment for everyone.

London and Munich. If Waymo can port its permitting playbook to Europe, the addressable market roughly doubles. If it cannot, US saturation is the ceiling.