Tesla CyberCab – hype over reality?

Last night Elon Musk revealed the much-anticipated CyberCab at an event at the Warner Brothers film studio in Hollywood – a location that has produced both massive blockbusters and monumental flops.  The key question for investors and anyone with an involvement or interest in the automotive industry was which category will the CyberCab fall into?

The event apparently started an hour late, and then only lasted 30 minutes.  They cynical amongst us might believe that there were software glitches that meant that the CyberCab ‘would not proceed’ as Rolls-Royce used to describe it in a more golden age of motoring, or that the self-driving software decided to take Elon on a less direct route from the backlot to the stage in order to rack up the fare.  Or perhaps Elon was just delayed on some Twitter/X or SpaceX business and didn’t think it mattered to keep the audience waiting whilst the Optimus robots did their best to entertain them?  I would certainly love to have been a fly on the wall backstage to know the answer to that one!

However, turning to what was said, the context for this launch was a combination of Musk’s claims that the future value of Tesla was not in cars but in AI-enabled autonomous driving technology, and the delay of the launch of the affordable ‘Model 2’ which had been expected this Spring.  There still seems to be a possibility that a cheaper car may be launched next year but how long it takes from reveal to hitting the roads is an open question that probably Elon himself can’t answer.  The CyberCab however we were told last night will be available for under US$30,000 with production starting “before 2027” – so in two years’ time.  The car that brought him on stage had no steering wheel or pedals so there is no potential for human intervention from within the car – meaning that the launch would depend not only on Tesla developing their self-driving technology to the point that neither killed the passengers nor abandoned them in the Badlands of South LA, but also the regulators agreeing that the technology had reached that level.  To be fair, Musk did say that the timing was subject to regulatory approval, so he has a ‘get out of jail free’ card insofar as the most recent focus of regulators in the US has been on potential criminal charges against Tesla, rather than rushing to expand approvals.

Setting aside the regulatory approval for a moment, the two critical points for the industry about the launch are that Tesla claim they can produce a BEV with a full self-driving capability for under $30,000 (and presumably see their way to making a profit at that price point), and that the self-driving capability itself can be delivered within that cost at a performance level that makes it both safe and reliable (i.e. not giving up in the middle of a busy road intersection).

We know from the Model 2 rumour mill that Tesla had some additional innovations in mind that would allow them to get more cost out of the product, such as the extension of their use of giga-casting to reduce assembly time.  Whilst not at Chinese levels of product cost, Tesla also has scale advantages that will help, and has long production runs that allow investment to be amortised over a much higher volume than would be typical.  With the reported ex-factory cost of a C-segment Chinese BEV being around RMB100,000 or US$14,000, then if Tesla got anywhere close to that, then they have some scope for the cost of the self-driving technology.  In the past that has not been cheap with a single LIDAR unit costing over $10,000 a decade ago, with four units fitted in the latest Waymo vehicles (recently approved for driverless taxi use in Los Angeles county).  The Waymo technology stack costs up to $100,000 per car depending on which reports you read, but they can justify that by amortising that cost over five years and fifty rides per day to get a cost per ride that is attractive relative to a human driver.  However, the cost of a LIDAR unit has now fallen to under $1,000 with an expectation that there will be a further fall to under $100 in the next decade.  Moreover, Tesla (controversially) does not use LIDAR, instead relying on cameras and advanced software, so challenging as it might seem for other manufacturers, a fully autonomous BEV at a selling price of under $30,000 might be possible.

The second product question is then whether the autonomous driving technology can be delivered by Tesla within two years.  This seems much more questionable.  Even accepting that use would be restricted to the most ADAS-friendly cities, Tesla seems to have a mountain to climb in terms of getting to a level of reliability that would make paying customers happy to jump into a CyberCab.  Their reliance on vision-based technology is at the root of many reported incidents involving Tesla cars using ‘Full Self-Driving’ mode colliding with other vehicles, sometimes with fatal consequences.  Existing users report the cameras being blinded by sun, darkness, dirt and rain – which is not helpful whichever part of the world you are in.  However, you should never under-estimate Elon Musk – he challenged orthodox thinking with Space X and ended up with a functioning reusable spacecraft that has shamed Boeing at a much lower cost.  Whether he can achieve an equivalent performance and reliability with his chosen technology in two years is another question however.

Which brings us to the big question and the ‘get out of jail free’ card – regulatory approval.  Even the most business and technology-friendly regulators take time to go through the approval process.  They want data – lots of it – and they want time to assess that.  They are answerable (at least in the West) to elected politicians who might love headlines, but they need to be of the right sort – not ‘unproven technology approved by regulators kills family’.  Even if the technology had been extensively tested now, without any incidents and very rare remote intervention, the regulatory approval process would surely take more than two years.

That in the end is the likely fate of the CyberCab.  It will probably not be production-ready before 2027, Musk will claim that the autonomous technology is proven, but it won’t be, and the regulators will not be close to approving the sale of CyberCabs to your local taxi company – even in Hollywood.  Not all movies have happy endings…