Self-driving cars will come to car sharing before showrooms.

Turns out, the first self-driving car you will ride in won't be one you own; it'll be one you order up on your smartphone from Lyft.

"The first mainstream deployment of autonomous vehicles won't be to customers but to a ride-share platform," General Motors President Dan Ammann told Mashable at the North American International Auto Show.

"We're going to have a car that operates only in downtown Austin that has a maximum speed of 30 mph and operates in controlled conditions"

Ammann later clarified he was speaking hypothetically; Although GM recently announced a partnership with Lyft, self-driving robo-taxis in Austin are not imminent.

This revelation comes just days after GM announced it was investing $500 million in a strategic partnership with the ride-sharing company.

The GM-powered Lyft cars could be more than just self-driving Chevy Volts or Malibu Hybrids, they will be digitally personalized to you — even before you open the door.
With your Lyft profile, the car will know who you are and your preferences and will arrive preset with all the things you like — think Spotify playlists and ideal seat settings. All you'll have to do is tell it where you're going within downtown Austin and it'll take you there autonomously.

Though Ammann was unwilling to give a specific timeline for the rollout of this autonomous test fleet, he said it would be sooner than a self-driving car being offered for sale to customers. There are two strong reasons why. First, the average car today sits unused 95% of the time, which is hugely inefficient. An autonomous Bolt in a Lyft fleet would be in use around 60-70% of the time.

Plainly, full autonomous technology is expensive — so is battery-electric technology as well as hydrogen fuel cells (Ammann tossed those three in together — not me). So right out the door, the economics are stacked against the likelihood of a customer choosing a self-driving car over a human-driven one. Lyft, however, could afford to pay such a price because it could run the car 16 hours a day while earning income for years on end. It makes much more sense.

The second reason is that it is easier for GM to create a car that works in a known city within certain limits at or below 30 mph. Unlike a car you might drive to the mountain or past a parade or through a desert, the autonomous system has much less programming to handle.

Importantly, expanding into car-sharing, General Motors isn't sacrificing its current business model but rather expanding it. Ammann explained that the majority of the carmaker's profits come from selling trucks and SUVs to people who live outside urban centers. Changing their business model inside cities doesn't affect that but rather opens up a big new profit opportunity.

That said, Ammann does see a business case for offering self-driving cars to retail customers some time down the road. In the short-term, however, GM is going to cut its autonomous teeth with Lyft.