Uber’s proposed hybrid robotaxi model was reported on September 23, 2026: autonomous vehicles would serve steady ride demand, while human drivers would add capacity during rush hours, bad weather, concerts and other peaks. The division is meant to match different kinds of supply to changing demand—not to show that robotaxis have already made the business more profitable.

Robotaxis for steady demand, drivers for surges

Uber’s proposal assigns robotaxis to baseline demand: the more predictable flow of everyday rides. Human drivers would help meet spikes when demand rises, including during rush hour, bad weather and concerts.

Sarfraz Maredia, Uber’s president of autonomous mobility and delivery, also described a late-night charging scenario: robotaxis could recharge when electricity is cheapest while human drivers cover the rush after bars close. It is an example of the proposed arrangement, not a reported measure of savings.

What Uber’s trip comparison says

Uber says Waymo vehicles on its Austin and Atlanta network complete about 30% more trips per vehicle per day than autonomous vehicles in other major robotaxi markets. The claim concerns those vehicles on Uber’s network; it is not a general performance figure for Waymo.

Pay trends and the future for drivers

Reported comparisons of Uber and Lyft driver earnings looked at Q2 against the same period two years earlier. They found lower base pay per hour in San Francisco and Los Angeles, and slower growth in both base and gross hourly pay in Austin than among rideshare drivers elsewhere. Uber separately said driver earnings in Austin and Atlanta had stayed consistent.

Uber has also acknowledged that some drivers will eventually be displaced. It said part-time drivers who rely on Uber as a safety net—including women, caregivers and lower-income workers—could be affected disproportionately.

Robotaxi economics remain unproven

Maredia said the economics of operating robotaxis had a long way to go. Whether autonomous fleets and their supporting infrastructure can cost less than today’s ride-hailing model remains unproven, so the proposed division of labor is not yet a demonstrated cost advantage.