Tesla's biggest problem with self-driving cars is that it actually has to sell them to customers

• Unlike Cruise and Waymo, Tesla has created a business around building cars and selling them to consumers.
• This business model has been wildly successful for Tesla, now valued as highly as three of General Motors.
• But Tesla is hamstrung by this model, as far as the potentially $8 trillion autonomous ride-hailing and ride-sharing business is concerned.
• Tesla is trapped, having created great products that consumers want and love, but also watching as competitors design all-new businesses optimized for self-driving economics.
• Visit Business Insider's homepage for more stories.

Tesla is the dominant electric carmaker in the industry - an impressive achievement in just over 15 years of existence. But it wants to be more - a purveyor of fully autonomous vehicles and the operator of a robotaxi service.

But unlike a lot of other companies in the spotlight for self-driving technology, such Cruise and Waymo, Tesla has a potentially big constraint: that its business is built around selling cars to consumers, just like traditional automakers, not simply picking up and dropping off in them.

That could lead it to miss out on the next big thing: the unowned car.

As the industry moves toward electricity and autonomy, one of the most compelling business models for a wannabe self-driving car company is as follows: operate a fleet of vehicles in a city where there are lots of riders who can access the service using smartphones, run those vehicles day and night, 24/7, and power them with electricity.

There are two companies that currently have a reasonable shot at attacking what could be, according to one of them, an $8 trillion market. Other firms could enter the space by offering different products - Ford and Volkswagen have partnered with Argo AI to provide a range of services, and Mobileye has positioned itself to be a key technology supplier, data broker, and regulatory leader.

But only Alphabet's Waymo and General Motors-supported Cruise are positioned to run what amounts to an urban-taxi service. Both are testing just that - Waymo primarily in the Phoenix area, and Cruise in San Francisco. That might lead you to wonder about Tesla.

Well, CEO Elon Musk is far from oblivious to the big narrative shift away from electric cars for consumers and toward self-driving vehicles. He's promised that Teslas would drive themselves from coast to coast (hasn't happened), and that Tesla would launch a robotaxi service (also hasn't happened).

That doesn't mean Tesla doesn't have a compelling technology proposition; it does, if it can radically expand its Autopilot system.

The critical challenge isn't really technological. Tesla is hamstrung in this area because its entire business is premised on selling cars to individual buyers. The company sold a record number of vehicles in 2019, about 370,000, but the company's so-called "fleet" sales aren't much of a factor; for major automaker such as General Motors and Ford, they are.

In fact, big carmakers are often criticized for pushing slow-selling vehicles into fleet channels, and although that's a fair complaint at times, rental companies and governments have to buy vehicles, too.

Tesla has to be innovative about doing more with less. Its staggering market capitalization - at close to $150 billion after an epic stock rally that commenced last year, Tesla is on paper valued as three GMs - masks everything from its limited manufacturing capacity (one car factory, with a second just coming online in China) to its skimpy cash-on-hand (roughly $7 billion, enough to run its business for a year or so) to its meager profits (its most recent quarter in the black closed out 2019 on a high note, but emissions credits have a lot to do with that).

For years, Tesla has been stretching its vehicle mix and pricing to cover multiple segments, and it has convinced some mass-market buyers to move up to either expensive BMW and Mercedes alternatives such as the Model S or Model X, or to drift down to the cheaper Model 3. If you think about it, with the Model S essentially unchanged since its introduction in 2012 and the Model X likewise not updated since 2015, Tesla is selling one new car: the Model 3.

Consequently, if Tesla wants to offer a robotaxi service, it has to figure out a way to do it with the platforms it builds - platforms that are awkwardly and wholly devoted for the moment to the retail market.

Cruise and Waymo are directly attacking the autonomous-mobility market

Cruise, meanwhile, has GM building it dedicated, all-electric cars at a factory in Michigan; they're Chevy Bolts that have been manufactured with the self-driving tech built in. The next step is a fully driverless shuttle, the Cruise Origin, that was jointly developed by GM and Cruise investment partner Honda. Cruise would own these vehicles; retail customers wouldn't.

Waymo, on the other hand, is adding its technology to everything from passenger vehicles to semi trucks.

Neither Cruise nor Waymo has any meaningful consumer friction to contend with as far as its services go. Critically, it doesn't have to worry about financing customers because no individual owners are involved. (Uber, for example, had to work out a leasing program for drivers who lacked wheels, later selling that program to another company called Fair.)

Tesla appears to have two countering ideas when it comes to self-driving vehicles.

The first would give owners the opportunity to commit their vehicles to a ride-sharing services that would put the cars to work when they'd otherwise be parked. This sounds like a decent play, and there are Tesla owners who lend their vehicles to rental services. But Teslas remain depreciating assets, and nothing undermines residual value quite like miles on the odometer and Burger King wrappers in the back seat.

The second would involve creating a Cruise-type fleet from scratch - a base-level Model 3, perhaps, or a more Cruise Origin-type shuttle. But Tesla would require a significant number of vehicles to do this, and its entire existing manufacturing capacity is dedicated to Model S, X, and 3 demands on the consumer side.

GM, meanwhile, is just going to transform one of its plants, Detroit-Hamtramck, into an EV factory and build all the Origins required by Cruise there.

Big in EVs, but weak on AVs

Tesla really has a uniquely Tesla problem here: It's been so successful at creating its own EV market that it can't now easily pivot to the potentially much more lucrative 21st-century business of self-driving robotaxi services. This is Tesla's version of what the legacy automakers confront with internal-combustion engines versus electric drivetrains.

Musk saw this coming years ago and has been trying, in impressive fashion, to wrangle an autonomous-mobility strategy into existence for Tesla. This is probably what keeps him up at night. He knows that Tesla has been overvalued for the time being because investors think the next few decades could witness a massive shift to electric propulsion.

But he also understands that replacing the drivetrain doesn't change the business model: Every car built requires an owner to buy it. And the building part is very, very expensive, making the prospective profit margin, best case, something like 20%.

The ride-hailing and sharing model could be more like 30%, at which point the economics look more Apple-esque than Detroit-y.

Tesla can't evade this problem. Its curse is that it has created fantastic products that people are willing to borrow tens of thousands of dollars to acquire. That small beast must be fed, as it has been by automakers worldwide for over a century.

And if EVs displace gas-burners, the planet should be grateful. So for Musk and Tesla, mission accomplished.

But from a business perspective, Tesla might have ironically picked the wrong business to be in, if capitalizing on the 21st-century greatest opportunity means never owning a car again.