Hail a ride from DiDi Chuxing (China’s version of Uber) and you’re likely to be picked up by an ordinary person driving any one of China’s dozens of mass production passenger cars. Expect the driver to be yammering away on the phone, while a good luck charm–likely a Daoist eternal knot–swings from the rearview mirror.

But order a ride from DiDi’s Premier service, and a dark-suited chauffer will show up in a polished luxury sedan, most likely a black Volkswagen Lamando. Complimentary bottles of own-label DiDi Premier mineral water await. The driver will gladly turn down the light classical music, while you take a business call in the back seat.

Facing long-term driver shortages and bad publicity over alleged crimes committed by their drivers, Didi is turning to these kinds of in-house alternatives over the independent driver model. Here it has an advantage over the likes of Uber and Lyft because, unlike its American rivals, DiDi doesn’t have to pretend to be a rider-driver matching service. It can operate openly as a self-described “mobility” company.

Uber’s existential crisis

For American ride-hailing companies like Uber and Lyft, the shift toward in-house services creates an existential crisis: their business models depend on maintaining the legal status of their drivers as independent contractors. That legal fine print allows them to avoid being regulated as taxi operators.

It’s a legal status that isn’t without controversy. In California, the state Supreme Court has recently established more stringent criteria for independent contractors. But the real threat to ride-hailing doesn’t come from the courts; it comes from self-driving cars.

Both Uber and Lyft are experimenting with self-driving cars or autonomous vehicles (AVs). Company-owned AVs certainly solve the driver shortage problem. But as these companies eventually transition to AVs, they will cease to be sharing apps and become straightforward mobility service providers.

For Uber and Lyft, that means a new (and much more capital-intensive) business model. It also means new regulation. For it’s hard to see how they can claim to be nothing more than sharing apps, when the cars they’re sharing are being driven by their own computers.

What is a car?

Like Uber and Lyft, Beijing-based DiDi is also investing heavily in AV technology. But unlike its American rivals, DiDi doesn’t have to operate in a legal grey zone. And the Chinese government actively supports its transition from the “ride sharing” business model to offering AV robo-taxi services.

If DiDi succeeds in bringing AV ride-hailing to China, it could change the very meaning of what a “car” is . Today, most people own the cars they use. Even Uber and Lyft drivers still own their cars. But company-owned robo-taxis have the potential to revolutionize the public’s relationship with personal transport.

In China, that revolution has already arrived–for bicycles. Sales of branded bicycles have collapsed as the country’s cyclists have turned away from owning bicycles in favor of renting them from bike-sharing apps. Bright orange Mobikes and yellow Ofos are ubiquitous in China, where it’s much more convenient to find a bike when you need one than to buy your own.

Mobike and Ofo still buy lots of bikes–millions of them–but their suppliers have been reduced from high-margin branded manufacturers to generic original equipment manufacturers (OEMs). Today, few people in China seem to know the names of the companies that manufacture the bikes they ride. And there are fewer of those OEMs than you might think.

Mobike and Ofo rely on a small number of OEMs to supply all their bicycles. Ofo has turned to China’s Tianjin Fuji-Ta Bicycle Company, a long-established branded manufacturer. And Mobike’s supplier is somewhat surprisingly not even a bicycle company at all. It’s Foxconn, the same OEM that assembles the Apple iPhone.

Market consolidation at China speed

China has been the world’s largest auto market by unit sales since 2009, but it is also one of the world’s most fragmented. More than a dozen local manufacturers compete for market share alongside nearly every global firm and multiple joint ventures. Such fragmentation makes it hard for any except luxury brands to establish a distinctive reputation.

If DiDi succeeds in doing to cars what Mobike and Ofo have done to bicycles, local auto brands may never develop at all.

In late April and early May, rumors swirled about a comprehensive deal between DiDi and Volkswagen under which VW would design and manufacture cars–including AVs–specifically for DiDi’s in-house fleet. Originally reported as due to be signed in early May, that was later punted to early June. Neither company has yet made it official.

Though a longterm partnership may yield substantial unit sales, Volkswagen may be cautious about becoming little more than a local OEM for the DiDi brand. General Motors, VW’s biggest foreign competitor in China, has apparently decided to go it alone in developing its own-branded robo-taxis. That may be the only way to avoid relegation into the branding second tier, as consumers increasingly choose a car based on the company that provides it, not the company that makes it.

Companies that lose their connection to the consumer end up competing on price instead of competing on brand. Just look at Foxconn, which makes most of the world’s iPhones, but earns the slimmest of margins on each one. In today’s auto industry, the parts suppliers are the OEMs that operate on razor-thin margins under constant price pressure from the big automakers. Tomorrow, it may be the automakers themselves who are OEMs to the mobility apps.

In the rapidly evolving world of AVs, everyone is talking to everyone. But whoever ends up owning the customer will end up on top. The rest of the pack will be demoted down to the ranks of OEM and generic technology providers to the surviving big brands. In China, it seems almost certain that DiDi will be one of the brands that makes it. Volkswagen beware.