Waymo Timeline Also Ranks All Robotaxi Players, With Tesla In Last
Waymo was founded at Google in 2009, and is now widely acknowledged as the clear leader among Robotaxi players. It's been a long path over 15 years. This new infographic plots the path using objective milestones. In addition, it places other active Robotaxi companies who have gotten past the initial milestones, based on where they are today if placed on Waymo's timeline. This suggests how far they have to go reach Waymo's current level, namely "beginning to scale."
The milestones used are based on the Yoda principle: "There is no try, only do." In particular there is also no "talk." While companies release demos and show off, they generally do not release the detailed internal statistics on their safety performance that you need to compare their progress with others. In some cases, they even release false statistics. You can't judge any vehicle just by riding in it—the ride had better be perfect or the company is not a contender. In fact it has to be near perfect over an entire human lifetime, something no one human can experience. Only statistics of extensive use can tell the story. Companies do reveal their data in what they do. A demo ride or video means nothing, but being willing to remove the safety driver says a lot—it says their internal statistics showed that was a reasonable move, even though any serious incident, with no human there to fix it, can destroy or seriously delay their entire project, as we saw with Cruise and Uber.
Unfortunately, California's required "disengagement reports" reveal very little, as companies use very different methods to count the disengagements, and what we want to see is a number nobody releases, namely the number of significant safety incidents that were the fault of the digital driver. All teams seek to make their record for those better than that of human drivers, for then they are making the roads safer.
As such steps that matter are ones like letting the public take rides, not requiring the riders to sign confidentiality agreements, and most of all, taking out the safety driver. Once you do that any mistake can quickly lead to a major problem and doing this means the company has data internally that says it is ready—or reckless. Once you take out that driver, letting employees, then beta testers and finally the general public are major milestones, as well as expanding territory and handling more difficult road situations.
While I, and other analysts can express opinions on team strength, ride quality, sensor selection, algorithms and hardware design, these are more subjective views. There are some other objective data points, such as team size and funding level, which can tell us how serious a project is—but even Apple, the richest company in the world, failed at their project and they never got to milestone one. Some companies (such as Aurora and Waabi) are not doing public demonstration and feel they can get to a high level entirely in sim. They won't show up. This chart only shows major robotaxi ventures with deployment, so you won't see Wayve, Waabi, Aurora or Nuro (which just recently switched from being a pure delivery company) ranked.
Notes On The Companies- is the leader and was the foundation for this chart as they got to these milestones ahead of all others.
- Baidu Apollo, with 75,000 rides/week, is chasing Waymo's heels, though it is much harder to get dependable data on China due to the different nature of news reporting and social media. These vehicles are operating in very complex environments and many cities.
- got the furthers before they were temporarily shut down. They vow to return soon.
- AutoX, Pony and WeRide: These are grouped together due to the lack of getting good data on Chinese operations, but all of them are operating no-safety-driver robotaxi services in Chinese cities, which is more than anybody else below them is doing.
- has done very limited no-safety-driver operations with employees. They said they would carry the public in Las Vegas in 2024, but have pushed it to 2025.
- deployed a test robotaxi service, but shut down all robotaxi efforts. They intend to let their customers operate the robotaxis, including the Rimac "Verne" service announced in 2024. They also have customers in China with plans.
- has been operating a test robotaxi service with safety drivers on board in partnership with ride-hail companies. They have yet to take out the safety driver. Absent recent turmoil, they might rank above MobilEye.
- Other players who have not gotten far in robotaxi include companies like , which has run delivery robots with nobody aboard, and plan to deploy in robotaxi, and , which does plan cars but currently does only trucks, but has had successful trucking runs. There are also companies that have done no public demos like Wayve, Waabi and many companies running shuttles like May, Beep/Holon, Navya and others.
- is more difficult to classify and is discussed below. They have never operated robotaxi service but did a big launch with minimal details.
- Some are off the list, with just basic announcements, like Mercedes, Toyota Woven Planet and other automakers have not announced any robotaxi ambitions.
Tesla excels in several ways. They are a huge company and have pre-sold over half a million "full self driving (supervised)" packages, though it is a driver-assist system, not a self-driving system, in spite of the name. Watched by drivers, it is doing more miles than anybody, but it is part of no robotaxi service. Tesla has shown some custom prototypes of their robotaxi.
While Tesla wins for transparency, in letting people see the real quality of their system, that real quality is quite poor at present. Tesla owners report the system as being able to do perhaps 10 drives in a row without a critical safety disengagement. This leaves many Tesla owners incorrectly impressed, as they don't realize that doing a complete drive is a meager accomplishment when the goal is to get where Waymo is, doing tens of thousands of drives in a row without problems. (Waymo is now giving 100,000 rides/week with nobody in the vehicle to intervene in the event of a mistake, and reports of Waymo at-fault incidents are extremely rare.) Tesla's ten-in-a-row puts it where Waymo was around 2015-2016 — eight years behind.
