Self-Driving Car Statistics & Autonomous Vehicle Statistics: The Complete US Report

As of mid-2026, robotaxis have logged more than 100 million driverless passenger miles on US roads, and Tesla's Full Self-Driving fleet has crossed two billion supervised miles. Yet most Americans still wouldn't get into one. This report compiles the latest US self-driving car statistics and autonomous vehicle statistics from NHTSA, the California DMV's disengagement reports, Waymo's quarterly safety releases, AAA's annual trust survey, and industry investment data.

How many self-driving cars are on US roads in 2026?

As of Q2 2026, roughly 3,500 fully driverless robotaxisoperate commercially in the US — primarily Waymo (~2,400 vehicles), Tesla Robotaxi (Austin pilot, ~500), Zoox (Las Vegas + Foster City, ~150), and a handful of smaller fleets including May Mobility and Motional. On the supervised side, more than6 million Tesla vehicles are running Full Self-Driving (Supervised) v13.x in the United States. Roughly 90% of new vehicles sold in the US in 2026 ship with at least Level 2 driver-assist features (lane-centering plus adaptive cruise).

How many miles have autonomous vehicles driven on US roads?

Cumulative autonomous miles are growing exponentially. Waymo crossed71 million driverless miles across Phoenix, San Francisco, Los Angeles, and Austin by early 2026 — adding roughly 3 million miles per month at current rates. Tesla's FSD (Supervised) fleet has logged more than2 billion supervised miles cumulatively. Zoox and Motional operate at much smaller scale (~5–10 million miles combined).

Are self-driving cars safer than human drivers?

Per published safety data, yes — in the operational design domains where they're allowed to drive. Waymo's most recent peer-reviewed analysis (covering 71 million driverless miles through early 2026) reports an 88% reduction in injury-involved crashes, a 79% reduction in airbag- deployment crashes, and a 93% reduction in pedestrian-injury crashes compared to the human-driver baseline for the same cities. NHTSA's Standing General Order on Crash Reporting (SGO) tracks every Level 3+ crash involving an automated driving system — robotaxi crashes per million miles run roughly 3–4× lower than the human-driver fleet average.

The caveat: these are heavily mapped, geofenced urban environments at modest speeds. Performance on uncontrolled rural roads, complex weather, and high-speed merges remains unproven outside Tesla's supervised dataset.

How many crashes have involved self-driving cars?

Under NHTSA's Standing General Order, more than 3,400 Level 2+ crashes have been reported by automakers since June 2021. Tesla accounts for the vast majority of reports because of its fleet size — but proportional to miles driven, FSD crash rates are broadly comparable to the same models without Autopilot engaged. Of those 3,400+ reports, fewer than 110 involved fully driverless vehicles (Level 4) and the overwhelming majority of those were minor low-speed contacts — fender-bumps, scrapes, and at-stop rear-ends.

How many people have died in self-driving car accidents?

Cumulative fatalities tied to engaged Level 2+ systems through Q1 2026 totalat least 64 deaths, according to NHTSA crash filings and PlainSite tracking. The majority involved Tesla Autopilot or FSD Supervised engaged at the time of the crash. Three fatalities have involved fully driverless vehicles — one pedestrian death in 2018 (Uber ATG in Arizona), and two lower-speed incidents involving Cruise vehicles in 2023 before that company was wound down.

Which company leads the self-driving car industry?

By driverless miles driven and revenue from paid rides, Waymo leads decisively. By fleet size (counting supervised vehicles), Tesla wins by two orders of magnitude. The 2025–26 shake-out left the industry roughly like this:

• Waymo (Alphabet): Commercial robotaxi in Phoenix, SF, LA, Austin; expanding to Miami and Washington DC in 2026.
• Tesla: Robotaxi launched in Austin June 2025; FSD (Supervised) on 6M+ private vehicles.
• Zoox (Amazon): Custom robotaxi vehicle, Las Vegas + Foster City commercial; slow ramp.
• Motional: Ioniq 5 robotaxis in Las Vegas and LA via Lyft; scaled back from earlier plans.
• May Mobility: Microtransit shuttles in 12+ US cities.
• Cruise (GM): Wound down by GM in late 2024 after 2023 SF pedestrian-drag incident.
• Argo AI (Ford/VW): Shut down 2022.
• Aurora: Autonomous trucking only, Texas I-45 freight corridor.

How much money has been invested in self-driving cars?

Cumulative global investment in autonomous vehicle development since 2015 has crossed$170 billion. The US accounts for roughly 65% of that figure. Annual AV investment peaked in 2021 at ~$33B and has since cooled to about $14B/year as the industry consolidates around fewer players — and as the headline failures of Cruise and Argo AI tightened venture appetite. Robotaxi unit economics remain unproven: Waymo is believed to lose money per ride at current scale despite paid revenue exceeding $1B annualized in 2026.

When will fully self-driving cars be widely available?

The standard industry forecast a decade ago — full Level 5 autonomy by 2025 — was wildly optimistic. Current expert consensus among industry analysts and academic researchers points to Level 4 robotaxis in 20+ US cities by 2030, buttrue Level 5 (any road, any weather) not before 2040. Personally-owned fully autonomous vehicles available for retail purchase are still considered unlikely before 2032 — though Tesla continues to promise "unsupervised FSD" on a much shorter timeline.

What percentage of Americans trust self-driving cars?

Despite improving safety data, public trust has barely moved. AAA's 2026 Automated Vehicle survey found that 61% of US drivers report being afraid to ride in a fully self-driving car, while only 14% say they would trust riding in one. The remaining 25% are unsure. Trust drops further among older drivers (76% of 65+ are afraid) and rises among Gen Z (37% afraid).

Which US cities allow robotaxis?

Commercial driverless robotaxi rides are currently offered to the public ineight US metros. Phoenix (Waymo) was first in 2020; Austin became the second-largest market after Tesla's June 2025 Robotaxi launch.

How will self-driving cars impact American jobs?

The US has roughly 3.5 million professional truck drivers, 1.5 million delivery drivers, and another 1.5 million taxi/ride-share drivers — a total of ~6.5 million driving jobs. The Brookings Institution's 2025 outlook estimated 1.6–2.3 million of those roles could be displaced or substantially changed by 2035 if autonomous trucking and robotaxis scale as projected. Trucking is expected to automate fastest on long-haul interstate routes; urban delivery and ride-share are the slowest to fully automate due to last-100-foot complexity.

Are self-driving trucks on US highways?

Yes — Aurora Innovation began commercial driverless freight runs on theI-45 corridor between Dallas and Houston in May 2025 and has since expanded to El Paso. Kodiak Robotics, Plus, and Gatik run smaller commercial pilots. As of 2026 there are roughly 200 driverless heavy trucks operating commercially in the US — almost all on Sun Belt interstate highways where weather is predictable and rights-of-way are simple.

What SAE level of autonomy do current self-driving cars have?

The SAE defines six levels (0 = no automation, 5 = full automation in all conditions). Almost every "self-driving" vehicle on US roads today is either Level 2 (driver-assist) or Level 4 (geofenced driverless) — Level 3 is rare commercially and Level 5 does not yet exist.

How does autonomous driving compare across operators?

The California DMV's annual disengagement reports remain the best public comparison — operators must disclose how many miles their vehicles drove and how often a safety driver had to intervene. Waymo's most recent filing reported21,196 miles per disengagement; Zoox reported181,051 miles per disengagement on its custom robotaxi fleet (a metric distorted by Zoox's much smaller test mileage). Tesla declines to file because its Autopilot system is technically Level 2 and exempt.