Tesla CEO Elon Musk has announced the company will unveil its robotaxi design, based on their next-generation vehicle platform, on August 8, 2024.
Earlier today we told you about a report from Reuters that said Tesla was focusing its efforts on developing a robotaxi, deciding to shelve plans, at least for now, for its affordable $25K model. While Musk refuted the story, saying through a post on X that Reuters was lying, a few hours later Musk took to X again to announce the company will unveil its robotaxi design later this summer.
Tesla Robotaxi unveil on 8/8
— Elon Musk (@elonmusk) April 5, 2024
Tesla has long talked about a robotaxi, initially claiming that all Tesla vehicle built with the appropriate hardware were capable of fully autonomous driving, once the software reached maturation.
In more recent years Musk has talked about developing a dedicated robotaxi. In 2022 at the Cyber Rodeo Musk revealed plans for the dedicated robotaxi, a vehicle that would be built specifically for autonomous driving without a steering wheel or pedals. As we learned last year, the robotaxi was going to be based on the same next-generation platform that would power the $25K model.
As for what the robotaxi will look like, we will have to wait until August 8, but there have been a few hints. Last year Musk gave us one hint, saying that it will be “quite futuristic looking.” Soon after, Walter Isaacson’s biography included an image of what was captioned as a “robotaxi concept” (seen above), showing a futuristic two-door design without a steering wheel.
A second image with Chief Designer Franz von Holzhausen standing in front of an “early robotaxi” model, matching the design in the drawing, was also included in the book.
Of course the most important part of the robotaxi will be the software powering it. Tesla did recently release v12 of their FSD software, which so far has proven to be very impressive. Based on statements made by Ashok Elluswamy, Director – Autopilot Software at Tesla, the switch to using end-to-end neural nets should accelerate the software’s rate of improvement over previous versions, as it is now capable of learning behaviours from millions of video clips, instead of having to rely on hundreds of thousands of lines of code.
