In my November 2024 article, “Robotaxis Accelerate, Ride-Hailing Leaders React”, I reviewed of progress made in the robotaxi space. Ten months ago, there were early signals that ride-hailing platforms were starting to react to the increasing volume of paid robotaxi rides in both the USA and China. This ecosystem has indeed regained significant momentum in the past year or two after a period of doubt and concentration. 

 

As a resident of Silicon Valley since 2017, I have become accustomed to seeing Waymo’s Jaguar iPaces all over in San Francisco where I live — there are about 800 of them now. They are also increasingly visible in Los Angeles where I often visit. Yet, I know how difficult it is for my European fellow citizens to imagine how common they have become and how comfortable, safe, and relaxing it feels riding in one versus taking a typical Uber or Lyft ride.

 

Waymo currently provides over 250k paid rides per week in five U.S. cities and has gained an estimated 20% of the overall ride-hailing market in San Francisco (see below current Jaguar iPace and new Zeekr in San Francisco). The Alphabet subsidiary is a distant leader in the USA, as Zoox and May Mobility are operating a few small-scale pilots. By comparison, China’s leading autonomous ride-hailing platform, Baidu’s Apollo Go, averaged 170k paid rides per week in Q2 2025, operating across a dozen cities where it is essentially challenged by Pony.ai and WeRide. 

 

As operational efficiency builds up, the scope of robotaxi fleets is expanding to more cities and wider operating design domains in existing geographies. Waymo and Apollo Go now operate fleets of approximately 2,000 and 1,000 vehicles respectively as of mid-2025 — Pony.ai has about 500. This poses a serious threat to ride-hailing operators — or rather presents a great opportunity.

 

Until recently, Waymo covered the complete robotaxi scope from the autonomous driving (AD) tech and remote assistance to fleet management, financing, and rider interface. Starting this year in Austin and Atlanta, the company disaggregated its business, narrowing its scope to the AD stack and remote assistance. It shifted (exclusive) distribution to Uber and asset management (financing, clean, charging, maintaining) to Spain-based Avomo. This approach will not only relieve Waymo from the margin dilutive, asset-heavy activities but more importantly enable its business to scale faster. Indeed, Uber delivers about 1,000 times as many rides as Waymo today and operates in 15,000 cities.

 

 

Robotaxi is a Strategic Play for Ride-hailing Platforms

Uber and Lyft both launched in-house AD initiatives — ATG in 2015 and Level 5 in 2017 — as they identified early on the potentially massive impact of the technology on their cost structure. It could unlock significant scaling. But they sold these entities to Aurora and Toyota respectively in 2020 and 2021 as they needed to reach profitability. Now that the robotaxi model is maturing, both ride-hailing operators are hastily partnering to bring these vehicles on their platforms.

 

Waymo rides are currently 30 to 40 percent more expensive than equivalent ones on Uber or Lyft but offer a superior service. Yet I expect significantly cheaper options in the coming years. In 2022, McKinsey anticipated that the use of a pooled robotaxi would cost 40% less than that of a personal vehicle and a fifth of that of human-driven ride-hailing — although still more than public transit.

 

Achieving profitability remains challenging as the costs of vehicle acquisition, fleet management, and remote operations are still significant. The path to sustainable robotaxi operations depends on scaling up fleets, increasing utilization and operational efficiency, and further reducing the capital cost for specific hardware and vehicle customization.

 

Uber and Lyft opened their platforms to robotaxi pilots in the past via partnerships with Motional, May Mobility, and even Waymo. Yet, collaborations like this have multiplied since mid-2023 and really accelerated in mid-2024. Uber has partnered with Waymo for the USA, Baidu for outside the U.S. and China, WeRide and Pony.ai for the Middle East, Momenta for Europe, Wayve for the UK, Volkswagen’s MOIA for Los Angeles, May Mobility for the USA, Nuro (and Lucid) for global deployment, and more. Lyft has implemented a similar strategy although at a smaller scale. 

 

Ride-hailing firms will increasingly integrate robotaxis in their operations. These vehicles will displace a portion of human-driven ones just like the latter displaced traditional taxis. Robotaxis will likely address the base network load as they are essentially a fixed cost business which requires high utilization. Traditional vehicles will cover load variations, adjusting supply to demand.

 

Early hybrid operations already provide interesting information. Waymo’s first 100 vehicles in Austin completed more trips per day than 99% of Uber’s human-driven ones — keeping in mind that many drivers operate part time on the platform. On a different note, Uber and Lyft have started to decrease bonuses paid to drivers as vehicle supply has increased, resulting in lower hourly pay (-6.9% y.o.y. in San Francisco vs. +1.0% nationwide).​