Tesla's Next Product Is a Subscription, Not a Car transcript

So, which of these are you thinking of doing? Okay, buying your own Model Y or Cybertruck when it becomes available and then having that for your own use or would you just decide to not own it any car at all and just go 100% hailing robot taxis? If you own Tesla stock, that's the argument we've been hearing and reading about all week. And I think that actually both sides are picking the wrong answer. My [music] take is that Tesla's next big product is a monthly subscription just like Amazon Prime or Apple One.

Okay, you pay one price and you get whatever car you need for as long as you need it. Okay, so stay with me here because it gets very, very interesting. Joe Tech Mara got most of the way there with his slider piece, the one that we covered on a show I did on Wednesday. Today, I want to take it one stop of that slide that he talked about and go all the way down [music] it. Because if I'm right, it changes what owning a Tesla even means and it changes what kind of company you're actually holding.

So, I want to go through Joe's idea first and then what I think the actual product looks like, then the evidence that Tesla's already building it piece by piece, then the part I think is the strongest argument in the whole show, which is why every car subscription before this one died and why this one won't. Reality check after that and what it means for you as a shareholder. Now, let's get going. Let's start with the argument.

So, Thursday night I posted the full presentation from the Cybertruck event on X, the whole thing from the Tesla team and the replies underneath that split right down the middle. So, one camp says, "I'd never own a two-seater with no steering wheel. I'll just hail one." The other camp says, "Under 30 grand, it earns money while I sleep, sign me up." And both camps agree on one thing without saying it, either you own the car or you don't.

Joe Tech Mara was at the event. He lives out near this place called Canyon Lake in Texas in the country where a hail button doesn't really fit his life and he wrote up what he calls a slider. His post on September 6th says, "Robotaxi is not Uber with no driver versus a $50,000 a payment. It is a slider. Hail on Monday, you lock a session for groceries on Wednesday, own and earn on Friday. Hail again when you're in another city.

You don't need to ask people to pick a poll. His article lays out six stops on one bar. So, hail on one for one trip, a session where the same car stays with you through a chain of errands, an hourly or daily rental, a monthly subscription, and his words for that are a monthly bucket of hours with priority when the network is light. Then owning the car and lending it to the network when you're not using it. And at the far end, owning it outright with no network at all.

The line from Joe I kept coming back to is this one. I love that framing. Now, we went through Joe's piece on Wednesday, so I'm not going to redo that whole thing. What I want to do today is pick stop four, the subscription, and argue that it's where the whole bar ends up collapsing to. And Joe hints at at this himself at the end of his article. He says robo-taxi ends up becoming being something you subscribe to the same way you subscribe to your phone plan.

In his words, "The unit of service is not a gigabyte, it's an hour of a vehicle that can be yours, shared, or hailed." So, that's my starting point. So, I have a theory, and I get that nothing here is announced, it's just my theory. My guess is Tesla's going to stop selling you a car, and they will start selling you the use of a car. Okay? Probably one monthly price. Gets you everything you might ever need. Picture what's in it, okay?

A bucket of ride credit for the days you just want to hail a cyber cab to the office and think about nothing. The right to have a Tesla sitting in your garage, though. You can also do this for a week or a month, a whole season if you want. When life needs a car that's always there. Let's say you have a newborn baby. The car seat stays strapped in the back for 6 months. Now, you don't want to keep hailing cars, what you want is a car that stays in your garage.

That's the right to pick the body for the job. Cyber cab for the Tuesday commute, maybe. Maybe a day you want a Model Y with a car seat in it for the baby months. Maybe one day you want a Model Y L, the six-seater for the house move, or a camping trip. Maybe you want a Cybertruck for the hardware store weekend. A premium Tesla for your friend's wedding. And then folding into that one price, the stuff you pay for separately today.

Okay, insurance, connectivity, FSD, the entertainment on the screen, charging. Never buy again. You dial up and down. A quiet month, you drop to hailing. Two teenagers come home for the summer, you add a car to the driveway. You move to another day, another city, same account, same profile, same payment. Just to be clear, Tesla hasn't said any of this. Their own car Cybercab FAQ says the interest form is for buying, quote, a fleet or an individual Cybercab vehicle for commercial purposes.

There's no consumer configurator. So, this is me looking at where the pieces point, and the rest of the show is the pieces. Why does that matter for you as a shareholder? Because a company that sells a car once at a margin gets valued like an automaker, but a company that collects a payment from a household every month for years gets valued like software. That's the whole re-rating argument. And I'm going to I'm not going to put a number on it today, but that is what I truly believe is going to happen.

Okay, so here's the first piece of evidence. Have you noticed that Tesla already sells everything in that bundle, just separately? Start with FSD. So, in February, Tesla stopped selling FSD outright. Subscription only, 99 bucks a month. Elon said the price goes up as the capability goes up. By second quarter, they had 1.48 million active FSD subscriptions. That's up 56% from a year ago. Rough math, that's somewhere around $790 million a year coming in from software automatically.

That's a lot of money. Then, there's premium connectivity, right? 10 bucks a month for the streaming and the maps. Tesla insurance is live in 13 states now. Wrote about $238 million worth of premium in the first quarter, and for the first time their carriers actually made money on the underwriting. Robo-taxi rides are metered per mile in seven markets, and charging has been on your Tesla account since forever. So, think about Amazon Prime for a second.

Amazon already had shipping and video and music as separate things when Prime launched. Prime just put them on one invoice, and that's what turned a customer into a member. Tesla's sitting in exactly that spot right now. The pieces exist, the invoice doesn't. And the way I think about the FSD move in February, that was the first domino. Tesla looked at a thing customers used to buy once for 8 $12,000 and said, "No, you rent that from us now every month.

Every month." Once you've done that with the software, and the car is the last thing left to do with it. Now, second piece. Elon's own words on this, and I want to be careful here because he's describing something different from what I'm describing, and then I'll show you why I think it's the same thing. So, a clip from Elon went around on Saturday. Dosh Designer posted it, and Elon says, "For the fleet that is owned by our customers, it'll be like an Airbnb thing.

You can add or subtract your car to the fleet whenever you want. You can say I'm going away for a week. Just one tap on your Tesla app, your car gets added to the fleet, and it just makes money for you while you're gone. You can add it to the fleet for a few hours, a few days, or a few weeks. Whenever you want it back, you can say come back, and the car will come right back. I'm highly confident that the revenue made by the owner of the car will far exceed the actual monthly payment.

Thank you, Dosh Designer, for clipping that. And if that sounds new, it's 10 years old. Master Plan Part 2, July 2016, says you'll be able to add your car to the Tesla shared fleet just by tapping a button on the Tesla phone app and have it generate income for you while you're at work or on vacation. Same idea, same tap. Okay, so that's the Airbnb version. You own the car, the network borrows it. Here's where I land, though.

If the network can take your car for a week with one tap and bring it back when you say come back, then the network can hand you a car for a week with one tap and take it back when you're done. Same dispatch, same come back button, same billing on the same account. So, the Airbnb version and the subscription version are actually one piece of the software. And the title is just a different name. And the supply side for for it is forming already.

Tesla put up a form on launch day last week for people who want to buy fleets. So, our mayor posted the link. So, Chamath signed up the same day. Signed up, he said, "Let's build a huge fleet. Put me in, coach." Now, Ross Gerber's take was the opposite. He told Tesla owners to let Tesla take the risk and not be the landlord because Tesla, in his words, will take the 20-30% off the top like Airbnb and if no one uses the car cabs, it's on you, not Tesla.

Which honestly is is a bit of my point. Most families would rather pay a flat price and have the car show up than run a one-car rental business out of the driveway. That's the gap a subscription fills. And it sits right between what Chamath is saying and what Ross is saying. Okay, third piece, and I think this is the strongest argument in the show. Car subscriptions have been tried. They all died, though. And the reason they died is the exact thing FSD gets rid of.

Let's go through the list. So, there was a thing called Book by Cadillac, which was launched 2017, and it wound down by the end of 2018. There was Access by BMW, Audi Select, the Mercedes Collection, all gone. Care by Volvo hung on the longest and shut down in 2024 after a fight with its own dealers in California. And this week, September 9th, TechCrunch ran a story on a company called Autonomy. That's a startup that in 2022 promised to buy 23,000 EVs and built its whole pitch on Model 3 subscriptions.

Their fleet lost about a third of its value when Tesla cut prices. The founder had to bail it out with his own money, and now they're adding gas-powered Fords through a dealer in LA to stay alive. Hertz, same story. Remember that? They ordered 100,000 Teslas in 2021, sold most of them off by 2024. So, why did all of them die? Cadillac's own explanation was, quote, "Cost and problems with back-end technology involved in a fleet management." Plain words, every time a customer said, "I want the SUV this weekend." a human being had to drive the SUV over, pick up the sedan, clean both of them, and park one in a lot where it sat losing value until the next person wanted it.

The swap ate the margin, and the car sat there depreciating in between. Now, picture the same weekend on a Tesla network. You tap Model Y L for Saturday, Friday night. It drives itself to your driveway. The Suburban that's been sitting there drives itself back to the network, and goes to work earning fares while you're camping. Sunday night, reverse. Nobody drove anything. The swap cost went to basically zero. And the car that used to sit idle is out earning.

Sandy Munro had a line on Sunday that fits here. That Tesla's fleet management software could ultimately be almost as important economically as its autonomous driving software. This is exactly what he meant. That's the moat, by the way. Waymo can't do this. They don't sell cars. Uber can't. They don't own any. GM and Volvo couldn't because a human had to do the swap. Tesla's the only company with the cars, the driver, their insurance, and the billing on one account.

Now, the industry's already moving this direction in smaller steps. Waymo launched a membership in June, Waymo Premier. 30 bucks a month. You get priority pickups, 10% back in Waymo cash, a few free cancellations, early access to new cities. Uber One is 10 bucks a month, and has somewhere north of 50 million members. So, the question of whether people will pay monthly for robotaxi access is already answered. Yes. Waymo got people to pay $30 a month for basically certainty.

A car comes when you tap. Uber One's ride perk by the way is 6% back and priority matching during busy hours. That's it. Uber can't bundle insurance because Uber doesn't insure the car and Uber can't bundle the car because Uber doesn't own one. What neither of them can do is put is put a car in your garage. Tesla's version starts where they are stopped. Same monthly fee for priority and credit and then the thing only a car maker can add, the actual car for as long as you need it with the insurance attached.

J Mad had a post about that side of it last year. He said in a world of Waymo and Cybercab, insurance so you can drive on the roads versus a robot will become a very expensive luxury good. Think about that for a second. If insuring a human driver turns into the expensive thing, then a bundle where Tesla's software drives and Tesla's insurance covers it gets cheaper and cheaper by comparison. Okay, fourth piece. The pick your car part sounds like the science fiction bit and it's the part that already physically exists.

So Joe flew his drone over Giga Texas on Tuesday and here's what he saw in the outbound lot. Hundreds of Model YL's and many more Cybercabs and for the first time equipped with a Starlink module in the hatch in big numbers. Thank you, Joe. So take the lineup as it sits today, right? You got a Cybercab which is two seats, no wheel, cheapest ride on the network. You got Model Y, five seats, already doing robotaxi in seven markets.

Then there's Model YL, six seats. The launch series is about 62 grand and US deliveries are listed for October and November. Now, Cybertruck for the moving weekend. Joe's line on this, you do not hail a robotaxi, you request a capability and then the network assigns metal to the need. And the first version of pick your body is already live. You open the robotaxi app in Austin today and it asks you which one you want, Cybercab or Model Y and it quotes two different prices. 30 Tesla posted a comparison on launch day.

Cybercab is almost half the price of a Model Y robotaxi. So the choose the car screen exists already. It just has two options on it right now, and the ride is the only thing you can buy from it. Tesla can launch this with the lineup it has today. The only new part is the plan. And the Starlink on the hatch matters more than it looks. A subscribed car parked in in a driveway 40 minutes outside the geofence still has the phone home.

Elon said on the Q2 call that that's so we don't have robo taxis missing in action. Very good detail. Uh one thing on the wedding example, because I said a premium Tesla, and some of you are thinking Model S Model S and X production ended June 30th. So, the premium car in this bundle is whatever's left in the fleet or whatever comes next, which honestly might be the point. A subscription fleet is how Tesla keeps a small premium tier alive without ever building a consumer S or X again.

Uh that's one's I guess that one. And remember, Tesla already has a rental business. I don't remember how many cities is that now. Just like Hertz, you can go there and rent a Tesla for multiple days. I think it's up to 5 days that you can rent it paying a month a daily fee. So, that model is already existing today. Okay, last piece. What does it cost? I'll give you the ceiling instead of a guess. So, AAA does a study every year on what a new car costs to own all in.

Payment, insurance, fuel, depreciation, the works. Last year's number was 11,577. Call it $965 a month. Depreciation alone is about $4,300 of that. Separately, experience says the average new car payment is $767 a month before insurance or anything else. So, that's the number. Any Tesla bundle that lands under roughly $1,000 a month for a household that would otherwise own one car is competing with car ownership. That makes it a household product.

Anything above that and it's a luxury product. Joe puts the crossover better than I can. He says the two numbers that define the market are at what price a subscription beats a month of hails for regular rider, and at what utilization owning plus earning beats paying for a bucket. Those crossovers will differ by zip code, of course, and by whether there is a child in the house. That's a map, and it looks different in Austin than it does in something like Canyon Lake.

And the size of market changes with it. Per ride robo-taxi is fighting Uber for what? 40, 50 billion dollars of a year of US bookings? A subscription that replaces a car is fighting for the money Americans spend owning cars, and that's north of a trillion dollars a year. Probably even more. All right, here's the reality check. I'm going to give you the bear case here. Number one, nothing is announced. I said that up top, and it's worth repeating.

The FAQ says commercial purposes. On the Q2 call, Elon said, "We expect to be vertically integrated with robo-taxi, which points at Tesla running its own fleet." And his public framing is Airbnb. You own it, and you lend it. The subscription is my theory. Nobody at Tesla has said that word. Two, the car in your garage part has a problem if the car is a cybercab. No steering wheel means it can only go where unsupervised FSD is allowed, and right now, that's part of Texas and Florida.

So, for most of the country for a while, the garage tier is a Model Y product, and the cybercab is a hail. That works, but it's slower than the vision. Three, Austin right now. There's only 45 cybercabs registered at launch. On launch day, cybercab ride was about 35% cheaper than the same ride in a Model Y. $7.77 against $12.10. By Sunday, the demand had flipped it. Tesla Oracle priced a 1.3-mile ride on the 7th at $15 in a cybercab with an hour plus wait against $7.95 for Uber X.

So, obviously, they're still working things out. They don't have the mass yet. That's early adopter demand, but a subscription is a promise that a car shows up. And today, a a doesn't always shows up. So, on top of that, NHTSA opened an audit query into the self-certification the same day the service launched. Wells Fargo called the event underwhelming. Stock dropped 6% that Friday. For this company called Autonomy. The reason I told that story is also a reason to worry.

Their fleet lost a third of its value in a day when Tesla cut prices. If Tesla owns the subscription fleet, Tesla carries the depreciation on its own balance sheet, and a lot of it. Million cars on a subscription plan is rough math 25 30 billion dollars a model that Tesla has to finance and hold. That's a CapEx question the automaker version of Tesla never had to answer. Elon's answers to that is the fleet operators, the Chamaths of the world, and Ross is right that that moves the risk off Tesla.

It also moves the margin. And the one I don't have an answer for, insurance on a car that's yours on Tuesday and the network's on Wednesday. Who's the insured? I've looked, don't know quite know the answer to that quite yet. Going to leave that one open. So, how do I hold the thesis with all that? As a list of milestones. I'll give you that list in a minute. Nothing here has a date. What it means for you as a shareholder, three scenarios, and they're all they're pretty lopsided.

The bear version is Tesla sells arbitrage to fleet operators, takes a cut of the fares, never bundles anything. Still a very good business. It's an automaker with a software fee on top, and you value it like one. The middle version is the Waymo path. A $30 a month membership inside the robo-taxi app priority and some credit, no car in the garage where you can't keep one at home. Small and I'd expect it within a year because Waymo already forced the issue.

The version I actually think is coming is the full bundle. And if it comes, the number to watch is households on a monthly plan. And the line to watch is Tesla reporting recurring mobility revenue on its own. The day they break that out, the re-rating conversation starts the next morning. Personally, I think the middle version comes first, and then the full bundle grows out of it. The same way Prime started as a free shipping and turned into everything else.

So, here's the things to watch. The first one is the robotaxi app, any monthly membership or plan tab, the app code already shows multi-stop trips and scheduled rides, which are Joe sessions. So, the plumbing for stop two on his bar is being built right now. Second, Tesla's wording. The day the FAQ says fleet or individual without commercial purposes after it, that's the tell. Third, Tesla insurance showing up attached to robotaxi ride or to any kind of plan.

Fourth, Model Y L in the robotaxi fleet as a body you can pick in the app and the first fleet deliveries to outside operators. Joe's hoping for those by spring of 2027. And the one that settles it, Tesla reporting recurring mobility revenue as its own line, of course. Let's zoom out for a second. Tesla spent four years building a car nobody can drive. I think that's because they never planned to sell you one. They plan to sell you the use of it by the month and everything else Tesla makes gets folded into the same bill.

That's my theory. Could be wrong on the timing, could be wrong on the whole thing. And I'll update as the pieces show up. Hopefully, you found this one useful. And if you did, please subscribe so you can catch the follow-up when Tesla actually says something about this. Thanks for watching. I'll see you tomorrow.