2026-05-15

Robotaxi tracker — May 2026 critical statistics across Waymo, Tesla, Zoox

— views

Cross-operator snapshot of who is actually shipping paid autonomous rides as of mid-May 2026: Waymo at 500K/week across 10 cities, Tesla Robotaxi scaling from <40 vehicles to 70+ Cybercabs at Giga Texas, Zoox at 350K+ riders since 2025 Vegas launch.

A consolidated tracker of who is actually shipping paid autonomous rides as of mid-May 2026 — not pilot announcements, not press tours, but live commercial service. Three operators have meaningful scale: Waymo, Tesla Robotaxi, and Zoox. Cruise (GM) wound down in late 2024 and is not included. All figures below are referenced inline.

Headline numbers

Operator	Paid rides/week	Fleet in service	Live cities	Cumulative autonomous miles
Waymo	~500,000^[1]	3,700+^[2]	10^[3]	200M^[2]
Tesla Robotaxi	Not disclosed; ~700K paid rides cumulative through Q4 2025^[4]	~37-42 in Austin (≤25 unsupervised)^[5] + Dallas/Houston	3^[6]	Not disclosed
Zoox	Not disclosed	Not disclosed	1 (Vegas)^[7]; SF + Miami + Austin testing	~2,000,000^[7]

Waymo — the operational leader

500,000 paid rides per week as of March 27, 2026 (TechCrunch).^[1] Doubled from 250,000/week in April 2025.
3,700+ robotaxis active across the fleet; cumulative 200M fully autonomous miles logged as of March 2026.^[2]
10 cities live: Phoenix, San Francisco, Los Angeles, Austin, Atlanta, Miami, Dallas, Houston, San Antonio, Orlando.^[3]
5 cities planned later in 2026: Washington DC, Detroit, Las Vegas, San Diego, Denver.^[3]
Target: 1 million weekly trips by end of 2026, per Co-CEO Tekedra Mawakana.^[8]
Vehicle base: Jaguar I-Pace (legacy) plus Hyundai IONIQ 5 and Zeekr RT (production fleet roll-in).

Tesla Robotaxi — the production ramp

Austin fleet: roughly 37-42 vehicles total, with ~19-25 running unsupervised as of late April 2026.^[5] Service area expanded multiple times since the June 2025 Austin launch.
3 live cities: Austin (launched June 2025), Dallas (April 2026), Houston (April 2026).^[6]
70+ Cybercabs spotted in outbound lots at Giga Texas on 2026-05-13, up from ~40 a week earlier on 2026-05-08.^[9] First Cybercab off the line: 2026-02-17; volume production began April 2026.^[10]
Cumulative paid rides: ~700,000 across Austin + Bay Area as of Q4 2025; no 2026 update published.^[4]
Operating hours: unsupervised service in Austin extended into evening hours as of May 4, 2026.^[5]
Caveat on fleet size: the vast majority of Austin vehicles still carry safety monitors. Headline counts and unsupervised counts diverge — read both.
Musk on the Q1 2026 earnings call: “I don’t think probably unsupervised FSD or Robotaxi revenue will be super material this year, but I do think it will be material… in a significant way next year.”^[5]

Zoox — the purpose-built bet

~2 million autonomous miles since the 2025 Las Vegas launch; 350,000+ riders carried cumulatively.^[7]
Las Vegas Strip is the only live commercial market; expansions in early 2026 added Sphere, T-Mobile Arena, Las Vegas Convention Center, and most Strip resorts.^[7]
Uber app integration announced for summer 2026 in Las Vegas.^[11]
New cities in build-out: San Francisco, Miami, Austin (employee-passenger testing phase).^[7]
Vehicle: purpose-built, no steering wheel, no pedals, bi-directional, four-seat carriage layout. Differentiates from Waymo (retrofit) and Tesla (Cybercab is also purpose-built).
Operating model: fixed pick-up and drop-off zones (not free-form like Waymo). Trade-off: simpler routing, less last-mile flexibility.

Cross-operator dimensions that matter

Dimension	Waymo	Tesla	Zoox
Hardware	LiDAR + radar + cameras	Cameras only (FSD)	LiDAR + radar + cameras
Vehicle type	Retrofit (Jaguar / Hyundai / Zeekr)	Retrofit (Model Y) → Cybercab purpose-built	Purpose-built
Service model	Free-form pickup	Free-form pickup (small geofence)	Fixed zones
Pricing	Surge / dynamic	Flat $4.20 promo → market	Promotional / flat
Disclosure cadence	Quarterly + blog	Sporadic (X posts + earnings)	Sporadic (PR)
Geographic footprint	10 cities; international planning	3 cities	1 city

What the numbers actually say

Waymo is the unambiguous operational leader by every reportable metric — rides per week, cumulative miles, city count, fleet size, disclosure quality. Tesla has the strongest production capacity story (Giga Texas Cybercab ramp), but its live-service scale is one to two orders of magnitude behind Waymo. Zoox has the most differentiated hardware bet (purpose-built bidirectional vehicle) but the smallest live footprint.

The interesting comparison is rides/fleet density: Waymo at ~500K rides/week ÷ 3,700 vehicles = roughly 135 rides per vehicle per week (~19/day). Tesla’s per-vehicle utilization is not disclosed but with fewer than 25 unsupervised vehicles in Austin, even very high per-vehicle ride counts produce a small absolute throughput.

Watch list for the next 90 days

Tesla weekly Cybercab count at Giga Texas. 40 → 70+ between May 8 and May 13 is steep. Sustaining that doubling rate would put 200+ units in lot by end of June.
Waymo Washington DC launch. First East-Coast capital-city deployment.
Zoox Uber-app go-live in Vegas. Summer 2026 promise.
Tesla unsupervised vehicle count. The ~25 number is the binding constraint on Tesla’s actual service capacity, not the 70+ Cybercabs sitting in lots.
First quarterly disclosure of Tesla Robotaxi paid-rides count if any. The cumulative-Q4-2025 number was published; whether Tesla will continue periodic disclosure is the open question.

Practitioner note

For builders trying to model AV-economics exposure:

Waymo is the only reportable unit-economics story today. Per-ride contribution margin remains opaque, but rides-per-vehicle-per-week × dynamic pricing × declining hardware cost is a quantifiable scaling model.
Tesla’s bull case is the Cybercab production curve, not the current service. Watch the Giga Texas weekly count and the ratio of unsupervised-to-monitor vehicles. The numerator (Cybercabs built) is scaling much faster than the denominator (vehicles approved for unsupervised operation).
Zoox is the most asymmetric bet but with the least visibility. The Uber-app integration is the single test of “Zoox at consumer demand scale”. If utilization in summer 2026 doesn’t materially exceed the 350K-cumulative-riders run rate, the bear case strengthens.
All three are still nowhere close to break-even on a fully-loaded basis. Treat any “robotaxi profitable in 2026” narrative with skepticism; the credible target is 2027-2028 at the operator-cohort level for Waymo, later for the others.

References