An update on Waymo's driverless taxi service Waymo One, six months after its Phoenix rollout: safety drivers remain at the wheel for most rides
Context & Ripple Effects
Waymo One launched commercially in December 2018 for Early Rider members in Phoenix as the first self-driving taxi service (first test-ride impressions), after Ars Technica reported that critics were already warning about thin regulation and transparency ahead of the planned fully driverless launch (pre-launch criticism). Six months in, Forbes reports that safety drivers are still behind the wheel for most rides — a gap between the 'driverless' label and day-to-day operations that the later coverage confirms was structural rather than temporary.
First-order effects
- Early Rider members and paying riders in Phoenix get a commercial service whose default mode still includes a human safety driver, meaning Waymo's per-ride economics carry driver cost on most trips despite the 'driverless' branding.
Second-order effects
- Scaling stays constrained even once drivers come out: by late 2020 the service in Chandler was handling only about 100 rides a week with cars capped at 45 mph (Chandler operations), so removing drivers is the gating step to any meaningful ride volume or pricing competitiveness against conventional taxis and rideshare.
Third-order effects
- The pattern points to autonomous taxi services maturing through narrow, geofenced zones at low speeds before widening — culminating in Waymo's plan for fully driverless rides without safety drivers in San Francisco (San Francisco expansion) — while regulators face pressure to set rules for services marketed as driverless but supervised most of the time.
The trend: Commercial robotaxi deployments are progressing from supervised launches toward genuinely driverless operation city by city, with the removal of safety drivers — not the initial commercial launch — marking the real milestone.