This post invites you to explore the relationships of relevant players in the world of self-driving cars.
What Is Happening in the World of Self-driving Cars?
Every day, there is a flood of blog posts and news about startups, technologies and investments in the world of self-driving vehicles (SDVs) - or autonomous driving (AD) as others prefer to call this technology.
These are particularly exciting times for tech people in the fields of robotics, logistics, artificial intelligence and machine learning, especially considering all the funding and attention that the domain currently experiences. However, knowing what is going on in this field is quite relevant for other people as well, as current changes to transportation technologies may significantly impact our methods of transport in the not-so-distant future.
That's why we would like to provide an overview of self-driving cars, its major players, their relationships and their products.
The Players and Their Relationships
Self-driving cars have the chance to revolutionize our world of mobility. Accordingly, many companies and institutions from various backgrounds are trying their best to become key players in this new field.
We suggest grouping the companies that contribute to the development of self-driving systems into four clusters.
Original equipment manufacturers (OEMs) have the greatest expertise concerning how to design and build reliable cars cost-effectively on a massive scale. Regarding the dimensions, complexity and history of the car business, typical OEMs are huge organizations with a settled culture and focus on a broad range of products for different target consumers. All this tends to make company-wide transitions slow. Tesla though, as a very young player in this big game and with its all-electric product portfolio, is considered to be more agile and less bound to old conventions.
While various driver assistance systems and modern user interfaces are key modules for new cars, OEMs are still fairly new to cutting-edge software innovations and tend to stay stuck in their customary 5-year innovation cycles. In contrast, the progress in the self-driving car space rushes much faster, forcing OEMs to accelerate their R&D efforts.
These are established companies that possess special skills in their respective automotive sub-domains. Suppliers have the potential to offer targeted niche products, for example for engine control units, connectivity solutions or infotainment. With autonomous driving being a domain that affects the setup of the entire car and has implications for mobility services, the extended scope makes it harder for suppliers to integrate their parts to the system or to offer their products as platform solutions.
It is interesting to observe how leading suppliers, such as Bosch or Delphi, are building up entire autonomous cars by themselves, aiming to sell even the complete hardware and software bundle to OEMs.
We call all these companies startups which are comparably young and have the abilities to move faster than their bigger competitors. An interesting question around this ‘definition’ is whether or how long you consider a company still a startup once it has been acquired.
It seems like a lot of the startups in the field of self-driving cars were founded by people in the artificial intelligence and robotics area that saw new software-driven areas of applications and thought they have a competitive advantage over ‘traditional’ car manufacturers. Some startups see it as a huge advantage ‘to start from scratch and build a robot that transports people’ rather than applying incremental changes to existing cars. Also, one has to see the numerous challenges around design, manufacturing and testing a car, swamping massive financial and engineering resources.
Similar to a lot of startups, big software companies (most importantly Google and Apple but eventually also Amazon for delivery and Facebook for drones) see a huge market potential in the field of transportation services. They aim to conquer this field using their advanced software and artificial intelligence capabilities. On the other side, they traditionally have little expertise in hardware products beyond the size of iPhones and virtual reality goggles and no expertise in actually shipping vehicles.
Besides these ‘player definitions’ we structure the inter-company relationships into three categories.
We call two companies partners when there is only a limited overlap of ‘common interests and cooperations’. One example is the partnership between Intel, Mobileye and BMW. Each of these three companies has its own defined market and the area of collaboration is quite bounded in terms of vision and computing solutions for self-driving cars.
Here the relationship is mainly a financial one which may also involve common interests but does not necessarily result in the exchange of other resources such as IP.
This is the strongest form of relation and embodies that one company owns another one. we can still consider the company that was bought a startup as long as it operates in a closed environment from the parent company.
We would like to emphasize that our categorization is probably not unambiguous and we intentionally decided not to cover some lose partnerships or acquisitions with minor players where we did not have enough certainty about the type and scope of relationship. This becomes even more difficult as early-stage deals and partnerships might not disclose their term sheets. We are not speculating for relationships and companies where we do not have data for.
Based on these different kinds of players and possible relationships, one could come up with hypotheses how the ‘most promising’ partnerships and synergies could look like. For this, let’s highlight two facts:
First, autonomous driving is extremely challenging and resource intensive with respect to both technology and funding.
Second, the capabilities to address these challenges are unevenly distributed: OEMs are experts for car development, software giants are at the forefront of artificial intelligence and large-scale data processing, suppliers have precise domain knowledge and startups push new ideas at lightning speed.
Based on this nice piece of theory, we expect:
The most beneficial collaborations are supposed to be across company types and core competences.
But let’s look at our interactive landscape now to proof our hypothesis.
The Self-Driving Car Landscape
(click and pull company logos around, double-click logo to highlight companies, or click on the blue edges to open more information about the relationships)
We would like to emphasize that the graph mainly aims to stimulate exploration!
Additionally, let’s look for some key insights:
Connectivity ≠ success. Some companies are actually much more ‘connected’ than others. On the one side this means in general high activity in the self-driving car space, but probably also higher dependencies. Google’s spin-off Waymo on the other side is almost isolated (besides its partnership with Fiat-Chrysler), while belonging without doubt to the few very advanced companies in the field.
The importance of money. We could extend this graph with dozens of more startups that also sound promising and exciting. On the other hand signs indicate that only a handful of these startups will have the resources to seriously compete with the big ones. The reason for this is very simple: It is very expensive to build, develop, test and deliver self-driving cars. And yes, there will be more acquisitions or ‘acqui-hires’ of companies like Cruise Automation for example but let's be honest: are the small fish really going to change the market? Probably not.
There is no clear winner (yet). At the moment it is hard to evaluate who is really leading the tech race towards a fully autonomous car. Some companies communicate openly their advancements, like NVIDIA and Mercedes-Benz at the CES 2017, while others, like Zoox, stay in ‘stealth mode’ and focus on their product development.
Visualizations & Contributing
We think that we, the authors, and you, dear readers, can grasp the complexity of the self-driving car landscape in the best way with a clear visualization.
That’s why we decided to encode relationships between companies active in the autonomous driving space with an undirected graph. In case you are curious about the underlying libraries, please check our Github project. The key visualization is build based on D3's Force-Directed Graph.
If you think the graph lacks some relevant companies or key relationships, please leave us a comment (just scroll down here) and we will update the visualization as soon as we can.
What Happens Next?
This is just the beginning of a series of posts about self-driving cars, the industry and what impact self-driving cars will have on our future of mobility.
Stay tuned for our next posts where we will look closer into the fascinating tech details. As always, if you have special suggestions or interests please leave a comment here.