TL;DR:
Between May 2025 and May 2026, Waymo’s recall-scope count of autonomous cabs rose from 1,212 vehicles to 3,791 vehicles.
That is more than a 3x increase in the Waymo fleet size in roughly one year.
We set out to measure Tesla’s robotaxi progress. Along the way, I accidentally tripped over something just as interesting, but in a very different direction: Waymo’s fleet footprint, as reflected in NHTSA recall filings.
NHTSA recall entries include a total number of potentially involved vehicles, and Waymo has been generating a steady stream of recalls since 2024. That’s not a sworn headcount of every Waymo vehicle on Earth, but it’s a solid proxy for fleet scale in the U.S., which is especially neat when autonomous taxi providers are cagey about providing exact fleet numbers.
Waymo Fleet Growth, as Seen Through NHTSA Recall Counts
Click a point
This panel updates when you click a datapoint.
Tip: Hover still shows a quick tooltip, but click locks the details here.
Note: The 1,212 and 3,067 datapoints are marked with an asterisk because they are partial-fleet proxies tied to 5th-gen ADS, not guaranteed whole-fleet counts.
Sources (NHTSA recall fleet proxy readings)
| Vehicle count | Recall Report Submission Date | NHTSA link | Notes |
|---|---|---|---|
| 444 | Feb 13, 2024 | 24E013000 | Software issue after low-speed contact involving a towed pickup in Phoenix. Broad proxy. |
| 672 | Jun 10, 2024 | 24E049000 | Utility pole collision in Phoenix. Often treated as whole-fleet proxy for that period. |
| 1,212* | May 12, 2025 | 25E034000 | * Partial-fleet proxy. Tied to vehicles operating 5th-gen ADS scope (barriers like gates and chains). It is unclear how many 6th-gen vehicles exist at this point, if any. |
| 3,067* | Dec 10, 2025 | 25E084000 | * Partial-fleet proxy. Tied to 5th-gen ADS scope (behavior around stopped school buses). |
| 3,791 | May 6, 2026 | 26E026000 | Explicitly covers vehicles equipped with 5th and 6th gen ADS scope (flooded or potentially untraversable roadway conditions). Strong whole-fleet proxy reading. |
Update: An eagle-eyed reader and autonomous recall subject matter expert points out, the dates reported above are the dates that the recall was submitted to the NHTSA by Waymo. The actual polling date of the fleet size occurred earlier. For precision’s sake, you could model this data by extrapolating the actual poll date from the federal government reporting or from internal manufacturer docs (that I don’t have access to). I prefer the submission date just for consistency and easy auditability. Article dates are updated and tagged to reflect this!
The Caveats
Two of these datapoints are tied specifically to 5th-gen ADS scope, which means they may undercount anything operating under a different generation or configuration at the time. That’s why they should be treated as partial-fleet proxy readings, not an absolute census. Luckily, those iffy datapoints are book-ended by whole-fleet recalls.
By May 2026, the recall scope explicitly references 5th and 6th gen ADS, which makes that figure feel closer to a full-fleet proxy. Still, every number here is a recall population, and recall populations can be narrower or broader depending on what exactly is being patched.
They’re Adding *Waymo* Taxis
Even with this acceleration, Waymo is still not “big” in the way normal people picture a national transportation service. The largest datapoint we have is still under 4,000 vehicles.
But if you’re asking who is actually deploying autonomous vehicles at meaningful scale in the U.S., Waymo is very likely the leader right now. And compared to Tesla’s robotaxi rollout, these NHTSA numbers read like a different league entirely.
