Waymo has started giving a limited number of riders access to its next-generation robotaxi, and there is an interesting wrinkle in where it comes from. The vehicle, a minivan-style ride Waymo calls the Ojai, is built on a chassis from the Chinese maker Zeekr. As of 28 May it is accepting select riders in Los Angeles, Phoenix and San Francisco.

What is happening

For now, Waymo is offering a small group of customers free rides in the Ojai to gather feedback and refine the experience before a wider rollout. Unusually for the industry, Waymo is blunt about the vehicle's purpose: the Ojai is built to make money. After years of robotaxis as a costly science project, the conversation has shifted to unit economics and whether each ride can turn a profit.

The Chinese-built part

A Chinese-made vehicle carrying passengers on American streets is the detail that will draw attention beyond the tech press. Waymo first began deploying the Ojai earlier in the year, per CNBC, and the choice reflects a simple reality: China makes electric vehicles well and cheaply. With US-China trade tension a constant backdrop, expect the sourcing to be a talking point, even as Waymo runs its own self-driving software on top.

The bigger picture

This comes as Waymo widens its map. In mid-May it grew its service area to more than 1,400 square miles across 11 US cities, per Electrek. A cheaper, purpose-built vehicle is the next lever, since the cost of the car is a big part of what has kept robotaxis from scaling. For riders, the near-term reality is simpler: in a handful of cities, a driverless minivan might now pull up, and the badge on it was made in China.

Three months on, the Ojai carries everybody, in volume

In May this was a handful of free rides for selected riders. In the week of 19 August, Waymo opened the Ojai to all riders in San Francisco, Phoenix and Los Angeles.

The scale of the rollout is significant. More than 3,200 Ojai vehicles have been imported into the United States — over 2,600 in 2026 alone — and Waymo is reportedly on pace for 5,000 by the end of the year. The vehicle is built by Zeekr, Geely's electric-vehicle brand, at Ningbo, and appears in customs records as the CM1e.

That is a different scale of commitment from the pilot described above, and it was made before the vehicle had carried a paying passenger.

The sourcing became a talking point, and the answer was structural

A Chinese-built vehicle carrying American passengers was expected to draw attention. It did, on two fronts, and Waymo had already designed around both.

The tariff is the blunter one. Reporting puts the rate Waymo pays on these imports at between 102.5 and 127.5 per cent depending on how the categories are counted. A company importing thousands of vehicles at roughly double their price, rather than sourcing domestically, is making a clear statement about what the alternatives cost.

The connected-vehicle rules are the more interesting one. Zeekr ships the base vehicle from Ningbo with no connectivity hardware and no sensors on it. The connectivity, the compute and Waymo's own sensors are added afterwards from a vetted supplier list and are designed in the United States. The Chinese content stops at the shell, which is why a rule aimed at Chinese vehicle software does not reach this vehicle.

Reading the badge misses the point. What matters under these rules is not where a car was assembled, but which supplier owns the parts that connect to a network.

Cheaper vehicles were the next lever. Deployment capacity turned out to be the constraint

While the cost of the vehicle has been a barrier to scaling robotaxi fleets, recent announcements suggest it is no longer the binding constraint.

The Nevada Transportation Authority granted robotaxi permits in August for up to 5,000 Tesla vehicles, 1,000 Waymo and 1,000 Uber across Clark County over twelve months. At the same hearing, Tesla's own Cybercab chief engineer said he did not think the company would be in a position to deploy 5,000 by then. A permit ceiling and a fleet are different objects, and the gap between them is where this industry currently lives.

The same pattern shows up in expansion announcements. Pony.ai and Uber said they would put more than 2,000 robotaxis across five European cities, four of which were not named and none of which carried a date.

Against that, Waymo's 3,200 imported vehicles are unusual precisely because they are countable. A customs record is a harder number than a permit ceiling or a memorandum, and it is the reason the service-area expansion noted above translated into rides rather than into a press release.

What is still open

Built to make money was the framing, and the accounts have not been shown.

Waymo has not published per-ride economics, and a fleet imported at a tariff above 100 per cent starts from a capital cost that a domestically built vehicle would not carry. Whether the Ojai closes the gap depends on utilisation and on how long each vehicle lasts in service, neither of which is public. The scale is now real enough that the arithmetic could be shown. It has not been.