Baidu and the Global Competition for Robotaxis
Baidu's Apollo Go operates in 22 cities already, but is largely unknown in the United States
Self driving cars have begun taking over the streets in American cities like San Francisco and Austin, but while the modern self-driving car industry traces its origins to the United States — and especially to the DARPA Urban Challenge in 2007— it has not stayed contained to Silicon Valley. In fact, if you go to Wuhan today, you can get a ride in a Baidu Apollo Go robotaxi.
Since 2017, Baidu has been working on its self-driving car program — Apollo — which has now spread out to 22 cities across multiple countries. Baidu also has, in contrast to a previous blog post I wrote, begun mass-production of the first purpose-built autonomous robotaxi, the RT6, and they delivered over 2.2 million fully driverless rides in Q2 2025 — with a total of over 17 million rides given so far in the program’s existence.
This is a massive program: Baidu says that Apollo Go covers 3,000 square kilometers in Wuhan, including highways with 70-80 kph speed limits. Waymo, by contrast, just expanded to cover the whole of Silicon Valley from San Francisco down to San Jose — an area of over 260 square miles or roughly 670 square kilometers. They have offered more than 17 million total rides so far. Baidu reported achieving 250,000 rides per week, which is approximately the same number as Waymo.
And they’re rolling out more, with 1000 vehicles planned for Dubai, more in Abu Dhabi, and a partnership with Lyft to expand across Europe.
Origins
The Baidu Apollo program started as it’s open-source “moonshot” self-driving car project back in 2017. It quickly built a network of partners and began to iterate, developing simulations and reference hardware as well as starting on-the-road testing. Similar to Google, Baidu could build off of its ownership of a popular maps service (maps.baidu.com). Their first commercial rides were in Beijing in 2021. Baidu released the 8th version of the open-source Apollo by December 2022 and had begun commercial operations in Wuhan by 2023.
The Baidu RT6 Robotaxi
The RT6 is the result of years of iteration on the base robotaxi design since the launch of Baidu’s Apollo program in 2017. The most recent version is sleek, effective, and cheap — less than $30,000 per vehicle. According to Baidu's CEO in the Q2 2025 earnings call, they’ve achieved unit breakeven in their flagship city of Wuhan, a huge milestone for any self-driving car program. Specifically, he stated: “We first achieved UE breakeven in Wuhan, where taxi fares are over 30% cheaper than in China’s Tier 1 cities and far below many overseas markets.” This means that they are rapidly approaching true economic viability before even factoring in any global expansion plans and moves into more lucrative markets.
Compared to the new Waymo Zeekr robotaxi, the Baidu vehicle also has a lot more sensors — lidar, radar, ultrasonic, and cameras, for a total of 38 different sensors, though they've said they’re testing camera-centric autonomy more in line with what Tesla has been rolling out in Austin. Baidu can work more closely with its manufacturing partners when building their vehicles, which they say is also an advantage over other Chinese robotaxi companies like Pony.ai.
Uniquely, their vehicles are also capable of autonomous battery swapping, something that lets them stay in operation much longer and decrease operational costs.
How Does it Compare to Waymo?
It’s only natural to want to compare Baidu with the leading American robotaxi company, Waymo. The two companies have a very similar philosophy, focusing on safety and using a wide variety of sensors.
In fact, in a lot of ways, the two companies’ trajectories look similar. Baidu’s rollout began in Wuhan for a similar reason to Waymo beginning in Phoenix: a friendly local government that helped make sure the deployment could start smoothly.
More generally, there’s a really interesting breakdown by Bryant Walker Smith comparing the two. In general, he says Baidu’s cars dealt with more complex scenarios, but had more obvious human interventions as well, and the pickup process was less smooth (i.e. there were fewer, predefined points where you could be picked up by a Baidu; whereas a Waymo works mostly the same way as Uber). At the time of his writing, Baidu supported freeways and Waymo did not; he also notes that Baidu rides seemed faster — nearly as fast as a human cab — as opposed to Waymo which is still notably slower.
On safety, Baidu claims to have a significantly better record, although it’s tough to compare exactly. Waymo claims 0.35 airbag deployments per million miles driven, versus one airbag deployment per 10.1 million kilometers for Baidu, or about 0.16 incidents per million miles driven. But this could be a feature of American roads and deployment environments, where cars are larger and speeds tend to be higher on equivalent roads. Neither company has seen any fatal incidents.
International Expansion
Baidu has aggressive international expansion plans, much like Waymo. With initial deployments in China, Hong Kong, and the United Arab Emirates, they plan to move into Europe in 2026, partnering with PostBus in Switzerland and Lyft in Germany and the United Kingdom.
Unfortunately, if you’re in the United States and want to get in a Baidu robotaxi, you may be waiting a long time — when I talked to a Baidu representative, they said it would likely be a long time before their service would come to North America. If you want to see Waymo and Baidu going head to head, the first location with both services will most likely be London, where Waymo plans to begin rides in 2026.
Final Thoughts
Waymo, Tesla, Wayve, Baidu and a handful of other companies are in a race to deploy robotaxis, and it seems inevitable we will see increasingly large deployments over the next couple years. I hope all of these succeed; a world with plentiful, safe, and affordable transportation from the Waymos, Teslas, and Baidus of the world will be a better one to live in. One reason for writing this article in particular is because the Chinese robotaxi companies have been relatively unknown to me (and I assume to many people reading this); it’s important to know that it isn’t just Waymo and Tesla scaling driverless rides right now across many different cities, and that this is a global phenomenon.
Previous article: