Since Tesla Has No Robotaxi, Just What Will They Show On Robotaxi Day?

Tesla has been promoting their upcoming "Robotaxi Day" presentation, set for October 10 at Warner Brothers Studios in Los Angeles. Delayed from August, it will ostensibly reveal something about Tesla's robotaxi plans and designs. Musk tweeted, with the title, "We, Robot" that it would be "one for the history books." When it was delayed, Tesla stock took a temporary but significant hit, due to the high expectations.

There's one big problem. Tesla doesn't have a robotaxi system. It's not even close. While some reviewers have effused their admiration for its ability to complete entire trips without needing intervention, it can do this only a few times in a row. That's where Waymo was around 2014, and now in 2024, Waymo is just starting to scale up, and is not yet in production with a commercial service. Some advocates for Tesla mistakenly think that they are further along because their driver-assist "FSD supervised" (which I purchased in my car) runs supervised on most roads, but a taxi service inherently has a service area. Waymo's cars could almost certainly drive much better than Tesla's on all roads if Waymo wanted to make them do that, but they have no reason to do so. Tesla has done impressive things but they are still very far from the target. A working robotaxi has to travel with no supervision and do Tesla's "complete drive" accomplishment 50,000 times in a row—as Waymo does—not just a few times in a row.

So What Might They Show?

If they don't have the system to make a working robotaxi, just what will they show? I've collected my own speculations, and speculations of other industry watchers to imagine just what they might show, given the hype. If they simply show a new car model using Tesla FSD, the industry press and savvy investors will not feel it was one for the history books.

The Basics

There will surely be some sort of vehicle with no steering wheel or pedals and a nice interior for taxi riding. It might just be like an existing Tesla with the controls removed—they are already sleek inside. It might be a smaller two-seater or a cybertruck sized luxury workspace. Problem is, it will be hard to surprise people as there have been lots of concept and real vehicles for this purpose already seen. Some of the "leaked" models look quite a bit like the recently revealed Rimac "Verne" robotaxi. We've seen small vehicles and concepts from Google, Volvo, Toyota, Mercedes and many others. We've also seen larger vehicles, sometimes called "toaster" shaped from Zoox and Cruise, using face-to-face seating and offering spacious interiors and easy entry. We can hope Tesla will surprise us, but it's likely to contain many elements we've already seen.

The vehicle will probably be able to operate with no supervision on the closed movie set lot, to give people a picture of the future. Musk has, for over 8 years, predicted that future was just a year away, but he's lost most credibility on that prediction as a result. Many companies have done private property demos of this sort (though generally including a LIDAR.) Waymo/Google produced a golf-cart based robotaxi service for use on the Google campus in 2009 as their first project, though it was not shown to the public. Back then, Google was 15 years away from that project scaling up.

So You've Built A Robotaxi, Now Where's Your Infrastructure?

There should be an app to do ride-hailing and perhaps some nice new user experience. And there might also be some interesting new in-car experience demo—many companies have tried to create concepts for things people might do while riding in a robotaxi, but the reality is very likely to be "stare at their phones" as people do in the back of Ubers and on transit.

Another "Next Year" Prediction

Musk's predictions that Tesla cars would reach self-driving within a year have become a source of ridicule within the industry. A self-driving car is one you could fall asleep in, betting your life, and it's a very tall order. As such, while Tesla isn't close to having that, it hasn't stopped them from saying it. So it could be that this event will simply have them show off a planned course demo (perhaps even in a model 2 or concept car) and declaring this again.

Optimus Robot

The day's title is officially "We Robot," which suggests we might also see announcements about Tesla's humanoid robot project. While past demos have been underwhelming (the first literally had humans in robot suits) the pace of progress in humanoid robotics is generally interesting.

While it would be a gimmick, Tesla could install a modified version of its FSD stack in its robot and have it drive any car, though even more poorly than FSD itself due to the reduced view of its cameras. The hard part would be making a robot that can get in and out of a car. On a planned course on a movie lot, this could be safe enough, and while not a realistic plan, would certainly get people talking and feeling it lived up to the hype.

More useful would be robotic charging. They may show other new capabilities for their robot—people love humanoid robot demos.

The Robotaxi Pitch

Tesla is sure to pitch the virtue of robotaxis, because that's real, even if they can't yet deliver on it. This includes their safety, time-saving and economics , plus the ability to restructure transport and the shape of cities. I can hardly argue with them, I've been a high priest of these concepts for 17 years, One idea that I no longer judge as valuable is what Tesla has called the "Tesla Network," namely an ability for private Tesla owners to hire out their cars for robotaxi service via a Tesla ride app, making money from their car when it's not in use. This makes much less sense with a custom steering-wheel-less robotaxi, but even with regular cars, the economics are more dubious.

Musk has frequently said recently that Tesla's real value will come from its robotics. There is a great deal to unlock, so they are sure to talk about it. No surprises to industry insiders but unknown to much of the public.

Fake-Out and Model 2

Tesla could go in an entirely different direction and unveil their new low cost EV, often called the "model 2" by fans, rumored to cost $25,000-$29,000 an be cheaper than typical low-end sedans. While Reuters reported that this project had been scrapped, Musk has said they are wrong. This would make a big splash, and make people forget they were promised a robotaxi.

Robotic Charging & Wireless Charging

One valuable use of self-driving technology is to make cars that charge themselves. That starts with cars that can drive to chargers (possibly just slowly and at night, which is something Tesla could probably pull off long before robotaxi) but also it requires cars that can somehow connect to the charger without a human present.

One way to do that would be with the wireless charging that Tesla has said it's working on. Wireless charging is expensive and has some losses, but it's very convenient, and it doesn't require a human—and can be lower loss if the car can position itself perfectly. However, wireless charging is not installed anywhere and it's only going to be slower level 2 charging where the car must stay for hours.

Another option is self plug-in. The car can position itself just right, so you don't need much of a robotic arm to insert the plug. Tesla demonstrated a "snake" plug years ago, but a robotic car doesn't need anything like that. Or to get very cute, they could try to get their humanoid robot to be able to plug in cars. That's a plus because it could handle all the stalls with one robot, and even plug itself in for robotic recharge, and this could happen at existing stations without much modification.

Highway self-driving

Mercedes sells a car which will do self-driving on freeways in traffic jams. They just upgraded it to handle doing 60mph in the right lane with other traffic. Once that gets to 70mph, and able to do open road, it's actually a very valuable feature, one that Tesla FSD owners would very much like, if they can pull it off. They could how work in this direction.

It's unlikely that they could show this on the movie lot, but many have asked why the Teslas that drive in Boring company tunnels are still manually driven. It just hasn't been a priority for Tesla, but it would be nice to see.

Map-based car or LIDAR

Tesla regularly declares it is a bad idea to use detailed maps and LIDAR laser sensors. Every other leading team uses these, and many of them have working robotaxis. Tesla would have to eat a lot of crow to change their mind on this, but what better way to eat crow than to do it in a splash announcement. "We were wrong. We can learn. Look at the great things we will now do!" Deflect from the mistake by owning it in a big public spectacle. But probably not, it's a lot of crow.

No matter what happens, there is sure to be plenty of buzz.