Waymo Robotaxi Fleet Expansion Targets 20,000 Vehicles by 2027

A fivefold bet on scale

Waymo is reportedly preparing the largest fleet buildout in the history of driverless ride-hailing. According to Semafor's reporting, the Alphabet subsidiary wants to grow from about 4,000 robotaxis today to 8,000 by the end of 2026 and 20,000 by the end of 2027.21 Hitting the first target would mean doubling the fleet in less than three months.21 By InsideEVs' count, reaching the second would require roughly a fivefold expansion over about 14 months.22

The outlets that picked up the story agree on these numbers. They differ on how official the numbers are. Several say Waymo declined to comment to Semafor and that the targets came from people familiar with the plans.2129 Another account says Waymo itself told Semafor about the plan.23 South Korea's Digital Today notes that Waymo has not publicly committed to any specific fleet figure.24 The safer reading is that these are internal goals that leaked, not public promises. That matters, because a company can quietly revise a leaked target without having broken its word.

The money behind the plan

The report comes shortly after Waymo closed its first debt deal, a $5 billion loan from lenders including PIMCO, Blackstone and Sixth Street.21 It sits on top of a $16 billion equity round earlier this year,25 which valued the company at about $126 billion.26 One account describes the loan as private credit with no guarantee from Alphabet. Waymo borrows against its own operations, and the deal terms reportedly allow fleet leases to be securitized later.23 Waymo has said the money is meant to fund growth both in the US and abroad.23

This shift in financing may be the real story. Equity funding pays for technology risk. Debt pays for assets that lenders expect to earn steady returns. If outside credit firms are willing to lend billions against a robotaxi fleet without a parent guarantee, they are treating driverless cars less like a research project and more like rolling stock. Waymo's own operating numbers help explain that confidence. Trips quadrupled in 2025,13 and weekly paid rides reached 500,000 across 10 cities by late March 2026.17 The company has repeatedly said it aims for one million paid rides a week by the end of this year.1812

The vehicle: Chinese-built, Arizona-finished

Most of the new fleet is expected to be the Ojai. It is built on an electric vehicle from Zeekr, a brand of China's Geely, and Waymo installs its self-driving system at a factory in Mesa, Arizona.21 The Ojai replaces the Jaguar I-Pace, which Jaguar stopped making in 2024.21 Citing research firm MoffettNathanson, the New York Post reported that Waymo is on pace to import about 5,100 Ojai vehicles by the end of this year.21 The van is already the only option for public riders in Denver and San Diego.15

This is where the regulatory angle gets sharp. The Commerce Department's ban on Chinese-made connected-vehicle and self-driving software takes effect from model year 2027. Waymo says it fits its own sensors and software in Arizona and shares no rider data with the carmakers.29 Its argument is that the Chinese content is only the vehicle body and drivetrain, not the brains. Still, the risk has not gone away. One analysis points to a bill in Congress that could remove the vans from US roads by 2030.29 Another openly asks whether the Zeekr supply chain survives if restrictions on Chinese-built vehicles tighten before the 2027 target.23

My reading is that this is the biggest single point of failure in the plan. Waymo can raise more money, hire more depot staff and build more chargers. It cannot quickly replace its main vehicle platform if Washington decides that Chinese-built hardware is off limits. A Hyundai Ioniq 5-based platform has been discussed as part of the fleet mix,16 but nothing in the coverage suggests it could absorb a sudden shortfall at the needed scale by late 2027.

Safety record meets federal scrutiny

Waymo enters this expansion with the strongest safety data in the industry, and with an active list of federal investigations. Across 271 million rider-only miles, the company's own analysis reports 95% fewer crashes causing serious injury or death than a human benchmark.29 An independent study by the Insurance Institute for Highway Safety looked at about 50 million driverless miles from 2021 to 2024. It found 81% fewer injury crashes and a 68% lower rate of police-reportable crashes.29 Two different methods point the same way.

The IIHS also warned about scale. Crash reporting is required, but reporting miles driven is voluntary, and only Waymo publishes its driverless mileage. IIHS President David Harkey said the current system is not good enough to continuously monitor a large expansion.29 So a fleet five times larger would be monitored by a federal data system that the leading independent safety researchers already call inadequate.

The incidents behind the investigations show what can go wrong at scale. In December 2025, NHTSA opened a preliminary evaluation after Waymo vehicles were seen passing stopped school buses.9 The Austin Independent School District documented at least 19 such cases.3 Districts said the violations continued even after a software update that covered more than 3,000 vehicles.4 In January 2026, a Waymo struck a 9-year-old child near a Santa Monica school, which led to NHTSA and NTSB reviews of how the system handles school zones.25 Waymo said its vehicle braked hard from about 17 mph to under 6 mph before contact.5

In May, Waymo filed a recall covering 3,791 automated driving systems because the software could let vehicles drive into standing water on higher-speed roads. As an interim fix, it tightened its weather-related operating limits.1 The flooding problem led the company to pause service in five cities.21 Freeway service was still being restored in the summer, with Phoenix first on July 29.15 The recall's unit count shows how quickly problems now spread: one software flaw touched nearly the entire fleet at the time. At 20,000 vehicles, the same kind of defect would mean a recall more than five times as large.

A regulator pulling in two directions

Federal policy is easing and tightening at once. At NHTSA's first National AV Safety Forum, Transportation Secretary Sean Duffy pressed to modernize Federal Motor Vehicle Safety Standards, which still assume steering wheels and pedals.6 The agency has launched three rulemakings aimed at a single national standard. Since January 2025, it has also opened or upgraded more than 30 investigations, reportedly with about 25% fewer staff.5 Approval also does not carry over from one city to the next, so each new market brings its own local approvals and conditions.26

Waymo's growth so far suggests the patchwork can be managed. The company won California approval to operate across 18 counties in August.12 Nevada has permitted up to 1,000 of its vehicles in Las Vegas.15 Abroad, Waymo said it will begin testing in the Madrid region and aims to carry riders there in 2027. WeRide and Uber already hold Spain's first permit for driverless cars.21 A Toronto service is planned for 2028.26

Reading the numbers

The coverage does not agree on how many cities Waymo serves. Waymo's own blog counted 14 public US metros as of September 1.12 Other reports say 15.21 Some count around 20 locations once testing markets and international sites are included.11 The differences come from whether employee-only and waitlist markets are counted, and they show how much of Waymo's map is still in rollout.

The gap with competitors is not in dispute. In Texas, Waymo runs hundreds of vehicles, while state records showed Tesla with just 42 robotaxis.21 Some analyses also point out that buying vehicles is the easy part. Each one needs sensors and computers installed, plus remote operators, charging, cleaning and maintenance before it can carry paying riders.2226 Going from 8,000 to 20,000 would take a 150% increase within a single year.26

My view is that the first target is far more credible than the second. The Ojai imports already in progress could plausibly support 8,000 vehicles by December, at least on paper. The 2027 figure depends on three things Waymo does not fully control: trade policy toward Chinese-built vehicles, the outcome of federal investigations, and whether safety data stays strong as miles pile up in new conditions like Denver snow and Las Vegas traffic. The debt financing shows that investors believe Waymo can scale. The next 14 months will show whether regulators and road conditions allow it.