Tesla is unveiling the Robotax on August 8, 2024 there is an ongoing ramp of the FSD (full self driving) software. What do we no know and not know about this rapidly changing situation?
This is all happening as Tesla’s sales growth is “between two growth waves”. Tesla has cut over 10% of its staff today.
Tesla cut the monthly subscription price of FSD to $99 per month.
If 10% of the 2 million possible US and Canada Tesla owners were to buy the subscription this would be $240 million per year.
If 20% of the 2 million possible US and Canada Tesla owners were to buy the subscription this would be $480 million per year.
If 10% of the 1.5 million possible China Tesla owners were to buy the subscription this would be $180 million per year.
If 20% of the 1.5 million possible China Tesla owners were to buy the subscription this would be $360 million per year.
Any near term actual production ramp for a new line for Robotaxi’s were require factory staff. This would be in Austin and/or Shanghai. This would require a few thousand to ten thousand factory staff.
Cutting global staff today would suggest that Tesla is at last 12 months away from pushing hard on a new Robotaxi or model 2 line.
Tesla Shanghai was able to seamlessly ramp the new Model 3 line a few months ago. This produced 30,000 units per month. This would seem be the maximum near term level of additional robotaxi production. Ramping with a few thousand units for a dozen Waymo scale (250 robotaxis) with safety drivers before removing the driver in six-12 months seems like a necessary step from August 2024 to around August 2025.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Let’s assume for the moment that Tesla meets or exceed time between failures – including crashes and simple stoppages (equivalent to a human having a heart attack or seizure behind the wheel & getting stuck in the middle of the road, though most people could at least pull onto the shoulder in that circumstance anyway). This is as measured by miles driven representative of how people actually drive: highway, suburban, urban, etc.
OK, that still leaves SOME failures and worse, they are less predictable than human failures. People crash more often in bad weather, at night, and compensate accordingly so they don’t (typically). Self-driving EVs seem to have problems almost randomly by comparison, which adds unpredictability to both passengers and other driver and pedestrians.
That’s not even the major problem.
Assignment of liability is still completely unaddressed. For Waymo and Cruise, which discontinued or were forced by the state to curtail their operations due to accidents and stoppages in San Francisco, the liability is a mixture of corporate and passenger who signed a waiver or other drivers were at fault: “The vast majority of crash incidents Cruise reported to regulators involved a situation in which the Cruise car was not at fault because it was legally stopped, rear-ended or had the right of way, according to a company spokesperson.”
But this is not a long term scalable solution.
Until there is a no-fault option, FSD robotaxis will never scale. Imagine an ordinary Tesla owner getting a call that their robotaxi just crashed and the passenger died and now they are on the hook for millions in liability because their insurance won’t pay past half a million, or maybe not at all through some loophole.
A no-fault option, regulated to a minimum degree of safety, but risk-assessed in the open market which can put pressure on manufacturers to make their cars safer by charging higher premiums – not the customers, who have no control over the manufacture of the cars – is necessary. Investor-bought bonds would raise the money for the no-fault pools, probably divided by geography at first, where only pockets of the country allow self-driving cars at all. Eventually, bonded FSD insurance could be rolled out nationwide with corporate participation too.
Until the liability question is resolved, it doesn’t matter how good FSD is. The first accident that has a multi-million dollar settlement will scuttle the most ambitious plans, including Tesla’s.
Are you aware Tesla has already settled its first lawsuit over wrongful death?
where is the sensationalism? All I hear is crickets.
Remember auto companies survived the pinto and continue to survive a never ending number of trivial as well as deadly recalls.
People have already been suing faulty drivers & their insurance companies as well as auto makers for millions, you just factor it into the cost of business.
As for the losses, Are you aware Tesla has opened its own insurance company?
Why would they do that?
because they will occasionally be at fault driving, even with they’re 1000 times better than a human.
With millions of cars on the road it’s bound to happen.
So what can you do to alleviate that issue?
Run your own insurance so you can distribute the damages across the customer base.
Which is what Tesla has done.
And as the cars get safer at driving, the cost of Tesla insurance will drop, hand in hand with the companies reviews, with no extra and expensive data gathering division required.
In other words Tesla insurance will have less overhead for Tesla vehicles and be cheaper.
Tesla can phase in Robotaxi level FSD by putting out the Tesla Network App and the entire Stack allowing it to use any participating Tesla as a Robotaxi, and operate it in beta as a ride share app in competition with Uber and Lyft – with driver supervision in place.
Tesla’s main purpose is data collection both to refine the whole system and further train FSD – so they can pass on 80%-85% of the revenue vs 75% for Uber and charge more.
That would let them collect data on real world performance / interventions on a very large scale to give regulators.
FSD-TN would do everything, the driver would be along for the ride and to make things legal.
Tesla betting on Robotaxi is delusionally optimistic given recent trajectory of real-world FSD releases. There has been no improvement in distance to critical disengagement over 3 updates in last few months. https://www.teslafsdtracker.com/
Distance to disengagement would need to be a steep exponential for this strategy to have any chance of panning out, it has to improve by about 1000x (ie eliminate 999 out of 1000 disengagements) to reach distances that humans average between crashes.
Tesla has two areas, which could give them massive gain or they can fail a lot. Both are high risk – high reward.
1. Self driving. Lots of promises and some nice results but still far from true self driving.
2. Humanoid robots on a massive scale.
The main limiting factor for new car models and persisting issue are batteries. I don’t think Tesla solved dry coating tech or applied it as they wanted it. Battery tech looks promising in the lab, but can fail miserably at massive scale. This is the reason we are not seeing Cybertruck or semi on the massive scale after all those years.
I agree.
If I was Elon, I would’ve bet the farm on Teslabot, not robotaxi.
Keep pushing FSD to get better, but I wouldn’t even be talking (publicly) about a robotaxi.
If a bot has an issue, it gets hung up, or breaks something like a part on a assembly line, or a vase in your home.
A much better headline then “Tesla’s robotaxi smashes into school playground”.
Far less risk, and the reward is equally huge, I think bigger.
Good points