⌕

THE CUSTOMERS ALREADY BOUGHT ONE

NOV 18, 2026 · SPLASH · THE MONEY · TESLA, CYBERCAB, CYBERTRUCK
The customers already bought one

THREE CARS KILLED IN ONE YEAR. The Cybertruck wind-down isn't a pricing failure — it's what happens when the product lasts longer than the replacement cycle, and the lineup has no flagship left to sell.

TESLA KILLED THREE CARS IN ONE YEAR AND ANNOUNCED A FOURTH THAT MOST OF ITS OWN CUSTOMERS CAN'T USE. The lineup is now a sedan, two SUVs, and a two-seat pod you can't drive outside a Texas zip code.

Everybody's asking whether the Cybercab is a good product. The better question is where the buyers went.

WHAT'S ACTUALLY GONE

Model S and Model X were discontinued in January 2026. On 6 November 2026, Elon Musk said Tesla will wind down Cybertruck production at Gigafactory Texas around mid-2027, freeing the line for Cybercab. Cybertruck sales were down 28% in 2026 on Cox Automotive's numbers.

That leaves the Model 3, the Model Y, the long-wheelbase Model Y L — and a $29,990 two-seater that requires a $299-a-month subscription to move, works only in supported Robotaxi zones in Texas and Florida, and can't be used to earn rideshare money without Tesla's written permission.

Tesla's consumer lineup just became a volume brand with no flagship. That's the part that should worry people more than the pod.

WHY THE SALES FELL — AND IT ISN'T THE PRODUCT

The Cybertruck take-rate is a pricing story. The Model S and X decline is something else, and it's the thing nobody at the company seems willing to say out loud:

The people who wanted one already bought one, and the one they bought still looks new.

A 2012 Model S with a bumper refresh passes for a 2022. There's no grille to date it, no body line that changed, no generation that announced itself in a parking lot. Tesla's design consistency was a flex for a decade and it turned into a sales problem the moment the early adopters were done.

That's the iPhone camera problem. If the new one looks like the old one and the only changes are a headlight, a screen and a chip, the upgrade has to be justified by want rather than need — and a car that still drives fine doesn't generate want on its own.

Tesla built cars that last, then discovered that cars that last don't get replaced. A Model S comfortably passes 150,000 miles without feeling mechanically old, and over-the-air updates keep a five-year-old car feeling current. That's excellent engineering and terrible replacement demand, and they're the same decision.

THE UBER PROBLEM

Then there's the thing that pushed the status buyer out of the middle of the lineup.

The Model 3 and Model Y became the rideshare uniform. Once a car is the default Uber, it stops signalling ownership and starts signalling a shift. That's what happened to the Prius and the Crown Victoria before it, and it happened to Tesla faster because the economics of driving for a living favour an inexpensive EV with cheap miles.

So the buyer who wanted a Tesla but not the Uber Tesla moved up to the Model S. And then Tesla killed the Model S.

That's the trap. The middle of the lineup carries the stigma, the top of the lineup carried the status, and the top is gone — replaced by a two-seater that looks like the fleet product it is. There is no longer a Tesla you buy to be seen in.

NOBODY IS SELLING TO NEW CUSTOMERS

Tesla doesn't advertise in any traditional sense, and its announcements are aimed at people who already paid attention: journalists, retail investors, and the fans who put deposits down years ago. That audience doesn't need convincing. It needs a product to buy, which is the problem.

Everybody else finds out from a YouTube reviewer — which means the first impression of a $30,000 Tesla is delivered by someone whose audience is watching for gadgets, not shopping for a car. That's not a marketing strategy. That's a distribution accident.

And it matters because the EV pitch was never self-evident. Early Tesla won by demonstrating: here is the car, here is the charger, here is the road trip. You cannot inherit a customer base you never explained yourself to, and the current lineup explanation — “buy a two-seat robotaxi pod” — doesn’t survive contact with a family of four.

YOU OWN IT. YOU DON'T CONTROL IT.

The subscription is the headline. The order agreement is the story, because of what it takes away from the person holding the title.

According to the agreement as reviewed by reviewers who ordered the car, you cannot tint the windows, wrap the exterior, or change its colour. The membership is what covers maintenance — tyres, wiper fluid, windshield replacement — rather than letting you handle any of it yourself. And it doesn’t cover your insurance or your charging.

Then the two clauses that actually matter:

You cannot sell the car without Tesla’s written approval — and Tesla decides what the price is. That is not a normal used-car market. That’s a resale market with a gatekeeper, on an asset you paid for in full.

Tesla can raise the monthly fee whenever it wants. There’s nothing in the agreement that caps it. A car that cannot move without the subscription, sold by the company that sets the subscription, is a lease wearing an ownership costume. You bought the hardware; you’re renting the right to use it, month to month, at a price somebody else controls.

Even Musk has said publicly that $299 is too high, floating something closer to $115. The terms didn’t change.

Run the arithmetic the way a buyer would: a couple of years of membership crosses $7,000 in fees alone, before insurance, before electricity, before the $29,990. At that point a Model 3 with the same hardware is the cheaper car — and it comes with a steering wheel, four seats, and the right to sell it to whoever you like.

WHO PAYS TO BE GEOFENCED

Which brings us to the actual question about the Cybercab, and it isn't about the price.

$29,990 plus $299 a month is roughly $48,000 over five years before insurance, registration or electricity — for a car with two seats, no steering wheel, a fixed service area, and a contract that forbids commercial use unless Tesla says otherwise. The original pitch for years was that owners would buy one and send it out to earn while they slept. The order agreement bans exactly that.

So who's the buyer? Not a fleet operator — they can't deploy it. Not a family — two seats. Not a status buyer — it's the least exclusive object Tesla has ever sold. The realistic customer is someone who wants a chauffeur and doesn't want a car, in a city where the geofence is generous enough to be useful.

And that customer has an alternative that already works: a black SUV with a human driver, which is what people actually order when they stop taking the regular ride.

Which is the argument we made about the limousine: privacy, space and control are the product, not the ride. A glass-canopy two-seater in a fixed zone is the opposite of all three. You can't spin the block, you can't detour to the spot you just spotted, and you can't change your mind without typing it into a screen. People pay to be dropped off; they pay more to be unseen and unbothered.

SPLASH’S READ

Tesla didn't build a bad car. It built a car for a customer who doesn't exist yet, and sold it to a customer base it had already exhausted.

The saturation story is the one the company can't tell, because it's an admission: the product was so good that the market finished. Nobody upgrades a car that still looks new and still works; that's not a failure of desire, it's a success that stops the treadmill. Apple solved the same problem by making the phone the upgrade — camera, screen, battery, every year a reason. Tesla made the car the upgrade once, in 2012, and then mostly changed the headlights.

So the pivot isn't strange, it's rational and desperate at the same time. Car sales are a low-multiple, cyclical business with a replacement cycle the company accidentally lengthened. Recurring software revenue trades at a much higher multiple. The Cybercab is the bridge between those two valuations — and the bridge is a subscription, which is the tell.

The honest summary: Tesla stopped selling cars to people and started selling miles to a market. Whether it works depends on whether regulators clear the geofences fast enough to matter — and on whether anyone wants to pay $48,000 for the privilege of being told where they can go.

WHAT WE COULDN’T CONFIRM

The three discontinuations — Model S and X in January 2026, and the Cybertruck wind-down announced on 6 November 2026 for around mid-2027 — are from Tesla's own announcements and CNBC, Reuters, USA Today and CNET's reporting, with the 28% Cybertruck sales decline from Cox Automotive's data.

Cybercab order terms — $29,990 before taxes, a $1,000 non-refundable deposit, the $299/month Autonomy Subscription, invite-only ordering, restrictions on wraps and tinting, and use limited to supported Robotaxi service areas in Texas and Florida — are from the order agreement as reported by Electrek, TechCrunch, The Verge, Teslarati and Not a Tesla App. We did not see the full contract ourselves, so terms here are as reported.

The rideshare ban is as reported: commercial use and third-party rental apps require Tesla's authorisation. Long-term promises about owners earning fares are the company's stated ambition, not a contractual right.

Figures circulating in commentary — an $18,000 build cost, a $3.3 billion German Cybercab plant, a $50,000 penalty clause, and a “46 days” teardown of the Fremont line — were not verified for this piece and are not used here.

Sales-saturation is an argument, not a measurement. Tesla doesn't publish model-level demand curves, so the claim that owners simply aren't replacing cars is reasoning from the numbers we do have, not a figure from the company.

MORE ON THE BOARD ▸
The Money · Cars & Machines · All Stories
Sources & claims: Tesla's discontinuation of the Model S and Model X in January 2026, and Musk's 6 November 2026 statement that Cybertruck production at Gigafactory Texas will wind down around mid-2027 to make room for Cybercab, are from Tesla's announcements and CNBC, Reuters, USA Today and CNET reporting; the 28% decline in Cybertruck sales in 2026 is from Cox Automotive data as reported by USA Today. The Cybercab personal-use order terms — $29,990 before taxes, a $1,000 non-refundable deposit, the $299 monthly Autonomy Subscription, invitation-only ordering, restrictions on wraps and tinting, and operation limited to supported Robotaxi service areas in Texas and Florida — are from the order agreement as reported by Electrek, TechCrunch, The Verge, Teslarati and Not a Tesla App. The rideshare prohibition is as reported by those outlets. Commentary figures that were not verified (build cost, German plant investment, penalty clause, Fremont teardown timing) are excluded and named in the closing section. The order-agreement details — the prohibition on tinting, wrapping or changing the car's colour, maintenance being covered by the membership while insurance and charging are not, the bar on reselling without Tesla's written approval with Tesla setting the price, the absence of any cap on the monthly fee, Musk's public statement that $299 is too high and his floating of about $115, and the roughly $7,000 of fees across two to three years — are from Tailosive EV's review of the order agreement and the same reporting.