The clearest sign that AI-driven transportation has moved from pilot project to daily infrastructure sits in a parking lot outside Mesa, Arizona. Monthly aerial photography analysis, tracked by a Reddit user known as jjricks_ and reported by Electrek on August 12, 2026, found 953 of Waymo’s new Ojai robotaxis staged at the company’s Magna-operated factory, with 684 units, about 72%, already finished and ready for deployment and 269 still awaiting conversion. Alongside them sat 12 finished Hyundai Ioniq 5 robotaxis, the first visible evidence of Waymo’s next vehicle platform entering its fleet.
From single vehicles to an inventory business
The sheer scale of vehicles sitting ready at a single site marks a shift in how Waymo’s AI driving system reaches riders. Rather than adding cars one at a time as demand grows, the company appears to be building inventory ahead of demand, converting base vehicles in batches so it can activate new markets quickly once regulatory and operational approval comes through. Waymo is currently completing roughly 500,000 paid rides a week across about a dozen cities, including San Francisco, Los Angeles, and Phoenix, and has set an internal target of reaching 1 million weekly rides before the end of 2026, a doubling that would require exactly the kind of vehicle stockpile now visible in the Mesa lot.
What changed under the hood
The Ojai vehicle represents Waymo’s sixth-generation Driver system, which the company has simplified considerably compared to earlier generations: it uses 13 cameras and four lidar units, a sensor package Waymo describes as a 42% reduction from its prior hardware generation. That simplification is largely a cost story. The full hardware suite now costs under $20,000 per vehicle, a figure that matters enormously for unit economics at a business built on operating hundreds of thousands of rides weekly rather than selling cars to individual owners. Base vehicles are manufactured by Zeekr in China and then shipped to Mesa, where Magna handles the sensor retrofit and quality conversion before cars are cleared to enter service.
The geography of the next expansion
Waymo’s stated ambitions extend well beyond its current service area. The company has said it intends to lay groundwork for ride-hailing operations in more than 20 additional cities during 2026, a list that reportedly includes Tokyo and London alongside domestic markets. A separate “Up Next” list names Washington, D.C., Denver, Detroit, Las Vegas, Seattle, and San Diego as the next wave of U.S. cities expected to get service. To fund that expansion, Waymo has reportedly raised $16 billion aimed at scaling its robotaxi fleet internationally, capital that would explain why nearly a thousand vehicles are sitting finished and ready rather than trickling into service gradually.
The competitive contrast that frames the story
Electrek’s reporting explicitly frames Waymo’s inventory buildup against Tesla’s much smaller robotaxi footprint, noting that Tesla has been operating what the outlet describes as a couple dozen Model Ys in Austin after more than a year of testing its own autonomous ride-hailing service. That gap, hundreds of finished, sensor-equipped vehicles staged for deployment at Waymo versus a small supervised pilot at Tesla, has become one of the most-cited data points in arguments over which company’s self-driving approach is actually closer to commercial scale, even though the two companies have taken very different technical paths, with Waymo relying on lidar-heavy sensor fusion and Tesla betting on a camera-only vision system.
Reasons for caution in the numbers
It is worth noting that the inventory figures themselves come from independent aerial photography analysis rather than an official Waymo disclosure, meaning the exact count of vehicles at any given moment is an estimate rather than a company-confirmed figure. Waymo has also not publicly detailed how quickly it expects to convert the remaining 269 unfinished units, nor given a firm date for when new international markets like Tokyo or London will actually begin carrying paying riders rather than conducting mapping and safety-validation testing. Regulatory approval, not manufacturing capacity, has historically been the bottleneck slowing Waymo’s expansion into new cities, and that dynamic is unlikely to disappear simply because vehicles are ready and waiting in Arizona.
What the buildup signals going forward
If Waymo does convert its remaining inventory and push weekly rides toward the 1 million mark by year-end, it would represent one of the most concrete real-world demonstrations to date of AI systems handling a physical, safety-critical task at meaningful public scale, millions of individual driving decisions made autonomously every week, not in a lab or a limited pilot, but in ordinary city traffic carrying paying passengers. The Mesa lot, in that sense, functions as a physical inventory of how much confidence one company is willing to put behind its AI driving software ahead of actual demand, a bet that will be tested as the vehicles roll out of Arizona and into cities that have not yet said yes to driverless rides at scale.
