Opinion: is Tesla’s Robotaxi all bluster?

Elon Musk presented his shiny new Robotaxi to a captivated audience in Los Angeles with his usual broad brushstrokes. He painted a picture of a world where you get all the dead time spent driving back to do whatever is it you like to do. Where car parks become green spaces, crashes become few and far between, and where running a small fleet of Cybercabs becomes a lucrative hustle.

He then went on to say it’ll be in production by 2026 and cost from under $30,000. He also showed an autonomous Robovan that looks like an unhitched Airstream trailer without any commitment on dates, and a dancing robot that chatted with the partygoers and poured them drinks. It was a lot.

Photography: Harry Rudd

Let’s focus on the Robotaxi here, because the 2026 timeline and price targets have been met by snorts of laughter and cries of bull**** from large swathes of the internet. Tesla stock was actually down ten per cent on Friday, and you can see why – Musk’s made crazy, unfounded claims before.

He’s the king of unfulfilled promises (especially around autonomous driving) and lengthy product delays, and the presentation was packed with promises and shy on fine detail. Including info about the app you’d use to hail them and the legislation needed to run them. So why should we believe him now?

Lets start by giving him some credit, because as tech demos go, releasing a fleet of 20 Robotaxi prototypes (and the same number again of self-driving Model Ys) around the Warner Bros Studio lot, with punters jumping in and out and wandering all over the road, was a chaotic test bed. And nobody got injured… that we know of. You could certainly tell that the prototypes were hastily cobbled together, but the bet was big and with his trademark swashbuckling showmanship, he pulled it off.

The hard part comes next. Can he really convince California and Texas to allow Model 3s and Model Ys to roam around driverless as soon as next year? Especially when the purely camera-based tech that Tesla employs is considered to have fewer fail safes than the mixed lidar, radar and camera-based system used on Waymo’s driverless taxi fleet, already operating in San Francisco?

Tesla engineers argue that having multiple sensors actually causes problems and confusion for the computer when you have conflicting inputs and need to decide which sensor to believe. Either way, the self-driving technology that was pretty laughable just a few years ago has improved rapidly, and will continue to do so as more and more data is collected every day.

If Tesla can get its existing cars up and running for unsupervised self-driving in these small, selected areas by 2025, then there’s no reason the Robotaxis can’t follow in the same areas soon after, with a controlled expansion across other American states from there.

After all, for all its futuristic styling, it’s the same FSD tech as a Model Y underneath, albeit repackaged into a shape that’s optimised for aero efficiency, with an interior that allows two passengers to lounge about and consume content on the 21in screen. But there’s one crucial difference – there’s no steering wheel or pedals, so no obvious manual override. An engineer we spoke to said the manual override amounted to a button you could tap that forced the car to pull over into the next safest spot.

My instinct tells me Robotaxis could easily be rolling off a production line in small numbers by 2026, but having them running about in meaningful numbers, with the full app ecosystem open to all, will take another two years beyond that. Like any great technology transition it’s a leap – for legislators and for us the humans that have to learn to trust it, but like anything, with familiarity that trust will come.

I remember talking to a gentleman in the street when we first drove the Cybertruck who refused to accept it was a production car, despite me promising him it was. Now LA is crawling in them, one passes by and you barely bat an eyelid. The event last night wasn’t the conclusion or even apex of Tesla’s autonomous journey, but the beginning of it.

There’s a long way to go. Now we can catch up on our e-mails while we get there.