Skip to main content

Elon Musk: Tesla will stop selling cars [Update: at consumer pricing] once full self-driving is solved

Tesla CEO Elon Musk has been talking a lot about Tesla Network lately, part of Tesla’s “Master Plan, Part Deux” which will enable Tesla cars with full self-driving hardware to operate as autonomous robotaxis to generate revenue for owners and for Tesla itself.

This is all still a ways off, but that hasn’t stopped Musk and others from theorizing about what might happen when the technological problems behind self-driving are solved.  Recently, Musk stated that any Tesla bought today is an “appreciating asset” due to its potential to be used to generate revenue in the future.  But an asset wouldn’t really appreciate unless a new, similar asset couldn’t be bought at the same price.  So now, Musk has committed to making that happen, stating that once robotaxis become possible, Tesla will likely stop selling cars to consumers, at least at anywhere near the same price.

The exchange came, as it often does, as part of a nighttime tweetstorm from Musk.  Among various other questions about the timeline for upgrading HW2+ hardware to Tesla’s new FSD computer and a comment about Tesla’s potential to have a million-robotaxi-fleet by the end of next year, Musk was asked whether prospective buyers would be able to keep buying Teslas well into the future, or if their potential as a revenue generating asset would make that price unattainable for a typical consumer:

And Musk answered, short and sweet, in one word:

This plan is not unexpected for those of us who have been following Tesla’s robotaxi aspirations, but this is the first time Musk has stated explicitly that consumers have a limited time to buy a car at anything close to what most consumers would consider a reasonable price.

The concept here relies on a few assumptions.  1) That Tesla will be able to make a car that can drive itself fully, at all times, with no human inside it.  2) That a car driving itself will be safer and cheaper to operate than a car with a human driver in it, since the human driver won’t need to be paid.  3) That these robotaxis will be able to make enough money driving themselves around that the potential profits will significantly eclipse the purchase price + running costs of a $40,000-$50,000 car even when the time value of money is taken into account.

If all those things become true, then Tesla has a choice between selling a car for $40,000 once or keeping that car and operating it as a robotaxi and generating perhaps ten times that amount over the life of the vehicle.  If that’s the case, then the company would be foolish to sell the car and miss out on that future profit potential.

And Tesla shared some numbers showing that they think this will be the case.  At their recent Autonomy Investor Day, they showed a slide suggesting that an average robotaxi would be able to bring in $330,000 worth of profits over its useful lifetime.  In keeping with this projection, Tesla recently raised the price of the still-unreleased full self-driving option (just a couple months after temporarily lowering it in a pretty shady way).

So this is what Musk is getting at when he says that consumers have a limited time to buy a car.  All of a sudden, a $40,000 Model 3 would need to cost six figures for the equation to pencil out for Tesla, and consumers will likely balk at that.  This, then, would give Tesla little reason to retain a large retail presence and they would likely focus on offering robotaxi services or possibly fleet sales.

Theoretically, Tesla could still sell cars at high prices, but Musk’s laconic answer suggests that Tesla plans otherwise.  Tesla has already stated that they plan to buy back leased Model 3s and put them into service as robotaxis after the lease term is up.

This is Tesla’s “master plan,” anyway.  Tesla owners will be able to use cars they already own to participate in the robotaxi fleet, but if Tesla can make more profit by keeping those cars themselves, they will do so.  Musk recently indicated that Tesla is thinking about opening up Tesla Network early as a human-driven Uber/Lyft competitor so owners can make a little cash on the side with ridesharing before full self-driving is achieved.

An autonomous robotaxi fleet isn’t just Tesla’s idea, either.  Several companies are aiming for the same thing, including Uber, Waymo and others.  Tesla’s approach differs in that they do not use LIDAR and other companies do.  It remains to be seen which technology will win the race to self-driving, but whichever one does will likely result in massive revenues for the company or companies which solve the problem first – and likely massive resistance and lobbying playing up the “dangers” of self-driving vehicles from those companies which don’t.

Update: Musk now states that rather than consumers having a “limited time to buy a Tesla car” as he originally responded “yes” to, instead Teslas will still be available for purchase, but at a “significantly” higher price:

As pointed out above, it seems unlikely that a consumer would want to buy a formerly-$40k car for “several times” more than that price, so it seems that in this eventual scenario, Tesla would probably not focus on consumer sales anyway.  Again, Tesla’s (quite optimistic) projection for the lifetime profit value of a Model 3 is $330k, a price that consumers are likely to be reluctant to pay.  If Tesla’s assumptions are correct, then for all practical purposes consumer sales would fall precipitously, to be replaced by fleet sales/Tesla-operated vehicles, in keeping with the scenario imagined in Musk’s original tweet.

Electrek’s Take

For a long time Musk has stated that his intent with self-driving cars wasn’t to stop humans from being able to drive on their own, but simply to make your typical day-to-day driving more safe and convenient.  Time and time again he has stated that Tesla wants to still make a car that’s fun to drive, regardless of autopilot.

And the Model 3 is very fun to drive as is.  Giving it the capability to drive itself won’t stop that from being the case.

But if Tesla moves to focus on building cars that can only be used in robotaxi fleets and are not within the reach of regular car buyers, that sort of throws the whole “fun to drive” thing out the window.  Sure you’d still be able to rent one and have fun driving it, presumably, unless Tesla decides that human drivers are more likely to cause accidents and therefore remove a car from the fleet and reduce revenue generation potential.

Of course a lot of this seems inevitable from the standpoint of a futurist.  It’s clear that, in the long term, there will be fewer and fewer places to drive a car manually, and this won’t just happen because regulators want it to, but because the public will demand it.  When human drivers are seen as unsafe, this will end up being a public health issue and people will want to put a stop to it.

It’s a shame, because driving the Model 3 – and my manual-steering Roadster, for that matter – is so much fun.

Of course there are a lot of assumptions here, as stated above.  But this does seem like a reasonably likely future scenario, as none of the individual steps to get there seem particularly unrealistic.

What do you think about Tesla’s potential plan to stop selling vehicles and operate robotaxis instead?  Let us know in the comments below.

FTC: We use income earning auto affiliate links. More.

You’re reading Electrek— experts who break news about Tesla, electric vehicles, and green energy, day after day. Be sure to check out our homepage for all the latest news, and follow Electrek on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our YouTube channel for the latest reviews.



Avatar for Jameson Dow Jameson Dow

Jameson has been driving electric vehicles since 2009, and has been writing about them and about clean energy for since 2016.

You can contact him at