Waymo: A Deep Dive into Its Robotaxi Operations and Cost Challenges

Waymo, which spun off from Google’s internal autonomous driving project, has attracted numerous renowned investors since 2020, raising over $11 billion. To date, Waymo has accumulated 2,000 miles of driving on public roads and 20 billion miles in simulation, and has conducted autonomous driving tests in more than 13 U.S. states.

1. Waymo Robotaxi Operations

In the five years since its official launch, Waymo has commenced operations in six U.S. cities, covering a total service area of approximately 600 square miles, with a current fleet size of 1,500 vehicles. Waymo initially launched its public service in Phoenix, Arizona, in October 2020. To date, it has rolled out services in 6 cities—Phoenix, San Francisco, Silicon Valley, Los Angeles, Austin, and Atlanta—operating within designated geofenced areas in each city, for a total service area of 588 square miles served by a fleet of 1,500 vehicles. Currently, users in Austin and Atlanta can only hail rides via the Uber App, while in other cities, rides are bookable through the dedicated Waymo One App. In 2026, Waymo plans to expand its operations to three new cities—Miami, Washington D.C., and Dallas. Furthermore, test operations in overseas markets such as Tokyo are underway.

In terms of ride volume, Waymo's weekly ride counts have steadily increased from approximately 10,000 in May 2023 to 250,000 in April 2025, reflecting the continuous expansion of its covered cities and regions. For reference, 250,000 weekly rides roughly equate to 3.25million quarterly rides. In comparison, Uber, with a global network and human drivers, records approximately 3 billion rides per quarter, making Waymo's quarterly ride volume approximately 0.1% of Uber's.

Looking at specific cities: California, a core region for Waymo's paid service, has seen an upward trend in weekly paid trips in both San Francisco and Los Angeles since its public launch in 2024, with a solid growth rate. This indicates that the public’s acceptance of autonomous taxis—an emerging service—is gradually increasing, and their willingness to pay for such services is also on the rise.

Figure: Waymo’s Total Weekly Ride-Hailing Volume (Unit: 1,000 Rides), Source: The Driverless Digest

2. High Costs Behind High Pricing: Safe, but Not Affordable and Scalable

Waymo's pricing strategy is cost-constrained, resulting in prices higher than those of manned ride-hailing services in the same area. Based on per-kilometer and time-based charges, Waymo's prices consistently exceed those of traditional ride-hailing providers like Lyft and Uber. For short trips (covering three distance ranges: 0.1-1.4km, 1.5-2.2km, and 2.3-2.9km), Waymo's charges—$26.5, $11.6, and $8.2 per kilometer respectively—are approximately 37%, 35%, and 34% higher than Uber’s. During peak hours (7-9 AM and 5-7 PM), Waymo’s prices are all higher than Uber’s, with price differences ranging from approximately 18% to as high as 58%.

Figure: Per-Kilometer Charges (USD) of Different Driving Distance Ranges, Source: Obi

We believe that Waymo's primary impediment to reducing its pricing below that of human-operated taxi services is the elevated costs associated with vehicle procurement and high-precision mapping.

We believe the primary reason Waymo cannot lower its prices below those of human-driven taxi services lies in its high vehicle costs and high-precision map costs.

Vehicle Costs: a Waymo Robotaxi currently costs approximately $160,000, while the cost of a conventional taxi is only one-tenth to one-fifteenth of that. Waymo's sixth-generation vehicle possesses a computing power of 2000 TOPS. Its autonomous driving system employs a multi-sensor fusion solution, utilizing 13 cameras, 4 LiDARs, and 6 radars to achieve 360-degree perception without blind spots and identify road conditions up to 500 meters. Compared to the fifth-generation system, the sixth-generation hardware cost is reduced by 40%, while machine learning optimization improves decision-making efficiency by 30% at night and in adverse weather. For example, in tests in Austin, Waymo’s Robotaxis successfully handled 127 cases of sudden lane-cutting, 38 detours around construction zones, with an average reaction time of only 0.2 seconds, significantly faster than the 0.8 seconds of human drivers.

High-Precision Map Costs: Waymo's autonomous driving relies heavily on high-precision maps. Every day, thousands of its vehicles transmit data collected by radar and cameras back to the headquarters servers in real time. The system then compares road condition differences with the previous map version with centimeter-level accuracy and updates the high-precision map for each vehicle in real time with minute-level latency. However, if a Waymo vehicle is removed from its designated operational area and driven elsewhere, its autonomous driving level immediately decreases from L4 to L2.

Therefore, Waymo's L4 autonomous driving capability has been achieved through a non-scalable, capital-intensive approach. In 2024, Waymo's revenue was $1.65 billion, while expenditures totaled $6.1 billion.

Conclusion:

Waymo’s high-cost technology route limits its pricing flexibility. As the public’s novelty toward Robotaxis fades, how many passengers will continue to choose Waymo Robotaxi for transportation? Furthermore, facing competition from rivals such as Tesla and Apollo Go, Waymo urgently needs to make changes.