GM and Lyft are teaming up to build self-driving cars — and take down Uber

The self-driving car wars are heating up. A few weeks after we learned that Ford and Google are creating a joint venture to develop self-driving cars, General Motors and Lyft have announced a self-driving joint venture of their own.

Lyft is raising $500 million in investment capital from GM along with another $500 million from other investors. Those other investors include two Chinese companies that are huge in Asia, albeit obscure in the United States — Alibaba in e-commerce and Didi Kuaidi in ride-hailing. Together, it amounts to a global alliance to battle Uber, a direct Lyft competitor that is working on its own self-driving technology.

The odds of this new alliance triumphing seem slim, as it doesn't have either the resources or the artificial intelligence expertise of Google, Tesla, Uber, and some of the other companies working in this space. But the deal does tell us a lot about how the global ride-hailing market is evolving — and why competition in the ride-hailing market is really just the prelude to a much bigger fight over dominance of the self-driving car market over the next couple of decades.

Here are four big lessons from GM's deal with Lyft.

1) The future of self-driving cars is rental, not ownership

(Spencer Platt/Getty Images)

We're used to thinking about cars as personal property that people own. Taxis and rental cars exist, of course, but they're niche products used in unusual situations or in a handful of central cities. Most households do most of their trips in cars the household owns.

I've argued before that self-driving cars will flip this around. Without the need to pay a driver, self-driving taxis will be so cheap that ordinary middle-class consumers, even in the suburbs and smaller towns, will find them cost-effective for most trips. Indeed, because rental allows several people to effectively share a single vehicle, getting around via self-driving taxi will likely become more affordable — not to mention more convenient and versatile — than owning a dedicated vehicle.

If on-demand rentals are the future of self-driving car technology, then teaming up with an on-demand rental service is a smart strategic move for a carmaker like GM. By the time self-driving cars start showing up in the marketplace — likely sometime in the 2020s — many customers will have grown accustomed to hailing cars using a smartphone-based service like Uber or Lyft. That will give them a lot of influence over consumers' use of self-driving technology, and so it makes sense that GM is acting now to make sure it has an ally in the ride-hailing market.

2) Car companies know self-driving vehicles are an existential threat

(Justin Sullivan/Getty Images)

New car technologies come along all the time, and it would be easy for car companies to assume that self-driving capabilities will be just another feature to add to their existing vehicles. But there's reason to believe self-driving technology is a much bigger deal, and that car companies that don't adapt quickly won't survive the transition.

Putting software in control of cars is likely to be a lot more than a cosmetic change. Just as the internet is fundamentally transforming industries from music to retail — and threatening incumbents in those industries — so self-driving cars are likely to prompt a fundamental rethink of how cars work. If cars are primarily rented, rather than owned, then they can be optimized for shorter trips and more heavily specialized for different use cases. Short-range, high-efficiency electric vehicles will become more practical. Companies may make cars in a wider variety of sizes and shapes, from hyper-efficient one-seaters to luxury minivans to support family vacations.

And car companies will also face new reliability and security challenges they've never faced before. When cars are controlled by software, they become vulnerable to hacking, and car companies' current manufacturing techniques — which involve delegating most of the work to hundreds of subcontractors — make car software almost impossible to audit. If car companies don't change their software development techniques, companies like Uber, Tesla, Google, and Apple are going to run circles around them.

Forming a joint venture with a prominent software company like Lyft gives GM a chance to start with a fresh slate. It could even be a sign that GM management recognizes that truly radical innovations will only take place if they're insulated from GM's bureaucratic culture.

3) The race to dominate the self-driving car business will be global

(ChinaFotoPress/Getty Images)

GM is Lyft's biggest new investor, but the latest round of funding also includes money from Alibaba (China's answer to Amazon.com) and Didi Kuaidi (Uber's biggest rival in China).

These investments represent a growing recognition that self-driving cars are likely to be a global market. Developing self-driving technology won't be cheap, and so companies that can spread those development costs across multiple big markets will have a distinct advantage. The United States and China are the world's two biggest markets, so it makes sense that companies in these markets would team up. Once GM and Lyft have developed self-driving car technology, they'll be able to turn to Alibaba and Didi Kuaidi to help sell it to Chinese consumers.

4) The new alliance has a lot of catching up to do

(Joe Raedle/Getty Images)

Lyft is seeking new allies because it's a huge underdog, as reflected by the terms of this new deal. Lyft sold shares to its new investors on terms that valued the company as a whole at $5.5 billion. A few years ago, that would have been considered a high valuation for a ride-hailing company, but it's tiny compared with the more than $60 billion valuation of Lyft's biggest rival, Uber. Lyft raised $1 billion in its latest fundraising round and $680 million total in 2015. By contrast, Uber has raised nearly $5 billion in the last year.

Uber has been spending all of that cash not only consolidating its lead in US markets but also extending its reach internationally. Its global reach means that if and when Uber's own self-driving car project — which it has been working on for a year — comes to fruition, Uber will be able to deploy the cars in dozens of countries around the world.

The Lyft-GM alliance will also have to play catch-up with other companies that have big head starts in self-driving technology:

• Google has had working self-driving car prototypes on the road for several years now, and has logged more than a million miles of real-world driving experience.
• Tesla's cars already have some rudimentary self-driving capabilities, and Tesla is working hard to make its vehicles fully self-driving. CEO Elon Musk said in Octoberthat fully self-driving cars are just a few years away.
• Apple has also been working on a car project in the past year. Not much is known about what the company is working on, but with deep engineering talents and tens of billions of dollars in the bank, it shouldn't be underestimated.
• /By Vox/