A look at the humans employed to rescue robotaxis; Waymo uses the Honk app to pay $20 to $24 to manually close robotaxi doors and $60 to $80 to tow its cars
A recent power outage in San Francisco shows how vulnerable self-driving vehicles can be when problems arise.
Context & Ripple Effects
Waymo’s San Francisco service evolved from an early permit to charge riders into a higher-volume operation, with the company reporting more than 100,000 paid weekly robotaxi rides across its operating cities. Its local fleet growth had also positioned it as an early leader in the city’s robotaxi contest.
That scaling has long carried operational scrutiny: earlier coverage documented robotaxis obstructing traffic and emergency vehicles after San Francisco opened to round-the-clock service. The outage shows that edge-case recovery remains an operational, not purely software, problem.
First-order effects
- Waymo incurs a defined on-demand recovery cost when vehicles need physical intervention, while Honk-dispatched workers become part of the service’s live operating model.
- During a power disruption, supervisors can be overwhelmed by vehicle confirmation requests, limiting the effectiveness of centralized human oversight precisely when many cars need help at once.
Second-order effects
- The economics of robotaxi service must account for field recovery, dispatch coordination, and supervisor capacity alongside vehicle utilization; recurring incidents can raise the cost per completed trip.
- Rivals seeking to match Waymo’s scale must demonstrate not only autonomous driving performance but also credible incident-response networks and graceful degradation during infrastructure failures.
Third-order effects
- Robotaxis are likely to be judged as managed mobility systems rather than self-contained vehicles: reliability will depend on the integration of autonomy, remote operations, local labor, and city infrastructure.
- If recovery work remains common at scale, the industry’s automation claims will increasingly be tested against the full cost and responsiveness of human fallback operations.
The trend: Robotaxi deployment is shifting from proving autonomous driving on ordinary trips to proving resilient, economically viable operations when vehicles and infrastructure fail.