General Motors has made another splashy deal in the hope of accelerating its path to producing and deploying fully self-driving cars. The automaker announced today that it bought 11-person sensor startup Strobe, which specializes in developing laser-based sensors called lidars.
Why GM is vertically integrating as it moves deeper into making self-driving cars
Its latest move: Buying 11-person sensor startup Strobe.
The acquisition gives General Motors control over the production and development of a sensor that many believe is critical to the development of autonomous vehicles. And GM is now one of the only carmakers or autonomous vehicle developers that owns a good portion of the major components of the self-driving supply chain: The car itself, the self-driving “brain” (via its 2016 acquisition of Cruise), a key part of the “eyes,” as well as the service layer, a proprietary ride-hail network.
The company wouldn’t divulge any of the financial details and there’s little that’s public about the startup. However, GM had already been cultivating a relationship via an investment from its venture arm, GM Ventures.
Have more information or any tips? Johana Bhuiyan is the senior transportation editor at Recode and can be reached at johana@recode.net or on Signal, Confide, WeChat or Telegram at 516-233-8877. You can also find her on Twitter at @JmBooyah.
Bringing lidar production in-house has become especially important as self-driving players scramble to meet impending deadlines to publicly deploy autonomous vehicles. Today, while there are many companies attempting to meet the demand of the nascent self-driving space, there are few that are mass producing lidars quickly enough and affordably.
As a result, the industry largely relies on one firm, Velodyne. However, paired with the high cost of lidar, Velodyne has also sometimes been slow to meet that high demand, some sources say, leaving companies waiting for the sensors for months.
“Existing commercially available solutions cost tens of thousands of dollars, are bulky and mechanically complex, and lack the performance needed to unlock self-driving operation at higher speeds and in more challenging weather,” Cruise CEO Kyle Vogt wrote in a blog post announcing and explaining the deal. “Strobe’s new chip-scale LIDAR technology will significantly enhance the capabilities of our self-driving cars. But perhaps more importantly, by collapsing the entire sensor down to a single chip, we’ll reduce the cost of each LIDAR on our self-driving cars by 99%.”
Bigger picture: GM is closer to becoming a vertically integrated autonomous vehicle manufacturer, which could help it move faster and build a better self-driving experience and business.
First there was its acquisition of self-driving startup Cruise in 2016. Then the company began experimenting with its own ride-hail network called Cruise Anywhere. And now the company owns the production and development of a critical sensor. (GM also has a partnership with no. 2 U.S. ride-hail player Lyft, but it’s clear the company is keeping its options open.)
The benefits of being vertically integrated are clear:
- Owning the major parts of the supply chain frees GM from relying too heavily on the production cycles of suppliers.
- It allows the company to develop the technology to its own custom specs.
- And it gives GM full control over the revenue that each of those parts bring in — say, for example, if GM decides to license its lidar technology to other companies instead of keeping it proprietary. (For now, GM’s acquisition takes this lidar tech off the market for its rivals.)
Additionally, there are benefits to having each of these components working side by side under one roof. A good example from another industry is how Apple has used its vertical integration of hardware, software and services to make the iPhone a much more cohesive product experience than smartphones running Android software from Google.
GM wouldn’t give any more details on how long it would take to start building the lidar sensors and outfitting its cars with them. But as it recently announced its factories were ready to manufacture cars equipped with the hardware to drive completely autonomously, it appears confident that it will be able to integrate the lidar into those production lines quickly.
There are technological advantages to moving quickly. While being the first to market with a fully self-driving car isn’t the end-all, be-all to competition within the space, it does give whoever is first the ability to test and validate its technology far longer and then update its software based on those learnings. That means, by the time the second player rolls out its fully self-driving cars onto public roads, the first one’s technology could already be far better.
GM is hardly the only company increasingly pushing to own more of its supply chain. Alphabet’s self-driving arm Waymo also announced it was bringing lidar development in house. Lyft, too, decided to begin developing its own autonomous software. In both cases, this could reduce how much the companies rely on external suppliers, while increasing their value to potential partners.
But the company that comes closest to mirroring GM’s integration is rival electric vehicle manufacturer Tesla — despite that Tesla CEO Elon Musk does not subscribe to the larger industry belief that lidar is a necessary technology for autonomous cars. Tesla’s scale is much smaller than GM’s, but it is manufacturing its own cars and building its autonomous software in-house. Musk has also talked about eventually creating an on-demand network of self-driving Teslas.
Lastly, this acquisition will likely have the same effect on the lidar and sensor industry that GM’s Cruise acquisition did. Backed by Spark Capital, Maven Ventures and others, Cruise’s exit was followed by a flurry of investments into and acquisitions of autonomous software companies. (See: Ford and Argo AI, Uber and Otto.) Since then, investors have been turning their focus toward the other components of autonomous cars, such as sensors.
This article originally appeared on Recode.net.
