Carmakers have regained some of the upper hand in self-driving

In 2014, Uber and Lyft were the future. Automakers like Ford and GM were under pressure as these fledgling startups threatened to cause a fall in sales, as they aimed to curb the need for personal car ownership.

Initially, carmakers concluded they needed to partner with Uber or Lyft and help create autonomous vehicles or risk extinction. But in the short lifespan of the self-driving industry, a series of significant changes has shifted the power to the traditional automakers.

Ford, GM, Volvo and others have regained some of their leverage as Silicon Valley has come around to the idea that, well, building a car is hard, and no matter how readily they can create the technology, self-driving cars won’t hit the road unless they work closely with automakers.

The shift first became clear in 2016 when Google finally came to an agreement to outfit a limited number of Fiat Chrysler cars with autonomous technology — now about 500 cars — and Uber bought 100 Volvos to do the same.

Google and Uber started to realize how hard-pressed they were to find automakers wiling to become metal benders for them. The tech companies wanted automakers to manufacture a Google or an Uber car. Automakers resisted, refusing to strip their cars of their own branding.

Now they have the leverage to plug their own autonomous vehicles into ride-hail networks. Daimler is doing that with Uber; Jaguar Land Rover is doing that with Lyft.

The self-driving ecosystem

The autonomous ecosystem consists of three primary components: The automakers, the self-driving software providers (often referred to as the brains) and the path to market represented by the customer network developed by Uber and Lyft.

Early on, automakers struck relationships with ride-hailing companies to make sure they weren’t shut out of this budding market.

General Motors invested $500 million in Lyft, establishing a plan to plug their future self-driving cars into the network once they were ready to hit public roads.

Then, for a large part of 2016, it was the software companies building autonomous technology that held the power. The narrative was that building the brain that drives these cars was the hardest problem, Silicon Valley was the answer, and these laggard, century-old companies would be lost without tech’s self-driving engineers.

The tech industry was convinced that the software was the hardest problem to solve.

Automakers like Ford wanted to work with Alphabet’s self-driving arm, now called Waymo, or companies that could compete with Waymo. Even Uber was trying to work with Waymo before former CEO Travis Kalanick decided to venture out on his own.

In March, GM acquired Cruise, a self-driving software startup, for $1 billion; in a talent grab, Uber acquired self-driving trucking startup Otto for $680 million. That acquisition is now at the center of a trade secret misappropriation lawsuit between Alphabet and Uber.

Waymo

Late to the game, Ford also took a majority ownership stake for $1 billion in newcomer Argo.ai, another self-driving software startup. That’s just to name a few.

But in that time, dozens of self-driving software companies have sprung up. That created some competition among these new entrants, but as there are only so many major car manufacturers on which to deploy these technologies, Detroit got the upper hand.

That shift prompted some VCs to turn their attention away from investing in the brains of the ecosystem, and focusing more on what some are calling “ingredient” technologies. These ingredient suppliers are creating hardware that will need to be fitted into a car manufacturers’ build of a self-driving vehicle.

That includes companies working on developing smaller parts of the supply chain — like the laser-based radar, or lidar. For instance, newcomer Luminar recently raised a $36 million seed round from 1517 Fund, Canvas Ventures and GVA Capital.

That has also left the software developers to compete on relationships and talent rather than a different technology offering. In the case of startups like Aurora and nuTonomy — both of which create systems that help a car process what its sensors pick up — investors are betting on which company will deploy their cars first based on the caliber of the engineers developing the technology.

Some expect an impending period of consolidation where some of these smaller startups join forces and bring together their capital, infrastructure and talent, and others fall by the wayside.

“One of my core hypotheses about this space is the technology is complex, difficult and new enough that there will be few groups that will get to a working solution in the near term,” nuTonomy co-founder Karl Iagnemma told Recode. “If you believe that only a handful will get access to the technology, then my feeling is the people who hold that technology will have this rare and valuable commodity.”

In large part, these 30 or more self-driving startups aren’t trying to build a car from the ground up. They also know how important it is to make sure the hardware and software are closely developed and are integrated into the production line. This all requires a carmaker partner.

nuTonomy

“I’ve learned over the years in doing self-driving cars, you need support from the automakers,” Argo.ai CEO Brian Salesky previously told Recode. “The virtual driving system is an extremely complex set of hardware and software; you need the support of an [automaker] to integrate that safely and cost-effectively.”

Now carmakers have a myriad of companies to choose from — allowing them to be picky about which players to work with in addition to striking non-exclusive relationships.

Automakers like Ford can hedge their bets and work with multiple self-driving tech companies.

By that same token, there’s no reason an automaker couldn’t simply work with both Uber and Lyft — or even decide to create its own ride-hail apps.

Add to that, these century-old companies have a clear and likely unchanging value in the self-driving supply chain.

There are, of course, exceptions, and the dynamic may in fact shift again. As Iagnemma said, if only a few self-driving software companies are able to reach level 5, considered full autonomy, then these companies will become more valuable.

The prime exception that comes to mind is Lyft’s new autonomous strategy, wherein the company is serving as an agnostic platform for automakers and tech companies to click into, but also building its own self-driving software. But even that was done in a bid to gain leverage in the supply chain, making it a more appealing partner to automakers.

Each of these components — the car, the brains and the path to market — are necessary and crucial parts of the ecosystem. But it’s clear, at least for now, automakers hold a great deal of the leverage in this dynamic.

This article originally appeared on Recode.net.

We have a request

Vox's journalism is free because we believe that everyone deserves to understand the world that they live in. That kind of knowledge helps create better citizens, neighbors, friends, parents, consumers and stewards of this planet. In short, understanding benefits everyone. You can join in on this mission by making a financial gift to Vox today. Reader support helps keep our work free, for everyone. Will you join us?

One-Time Monthly Annual

$95/year

$120/year

$250/year

$350/year

Other

$

Yes, I'll give $120/year

Yes, I'll give $120/year

We accept credit card, Apple Pay, and Google Pay. You can also contribute via