Autonomous Car and investing
From horseless carriage to driverless car — how to make money?
How to make money from the shift to selfdriving cars?
Warren Buffett once said that the best way to bet on the car in 1900 was to short horses. It was hard to predict Henry Ford, Gottlieb Daimler or Ferdinand Porsche. You don’t have the benefit of hindsight.
So where do we stand today?
The first commercially available autopilot is Tesla Model S and Model X. Both cars can do what no other car can. They can hold a lane, drive themselves on highways and do some self parking.
There is a lot of work being done behind the scenes. The value stack is large and opportunities look abundant. Let’s look at the value stack:
The semiconductor space is always complicated. Everybody claims to be involved. I focus on two companies, Nvidia and Intel. The former is cutting edge and the latter has the corporate DNA to figure out how to become cutting edge.
Selfdriving cars require sensor fusion and very low latency. Sensor fusion means that data from different types of sensors has to be processed at the same time. Nvidia has an edge in this type of problem because their GPUs are good at matrix calculations. Cameras, radars, liadars and other sensors accumulate data. This data needs to be processed fast.
Inference is when that data is used to make decisions. Inference requires very low latency so the car can make decisions fast enough.
Intel has an interesting approach to this problem. They are leaders in CPU. Now they augment the CPU with programmable chips solutions to drive down latency with inference. Another approach is to better aligning memory closer to the CPU to reduce latency.
Intel wants to develop semiconductor technology which can be used by OEMs to drive their autonomous cars. It’s not surprising that Intel is thinking about replicating the CPU-PC model in the car space.
Nvida shares are up 5X and hence reflect a lot of selfdrive potential. Intel shares trade at sub 13PE and have lots of upside if Intel can make autonomous driving a substantial part of their business.
2. Car companies
We don’t think any of the incumbents have the know how to develop self driving cars. They have been good at making and marketing cars that drive with an internal combustion engine. They have no expertise in developing software for robots.
The only company which might have an edge is Tesla because the company was founded on the premise of software. In autonomous car technology Tesla acts like a startup. It might or might not succeed. But at least there is no legacy baggage like with the other incumbents. So far, Tesla has proven to be very good at developing and implementing selfdrive technology. It looks promising. We are long Tesla shares and one of the main drivers, we believe, is the edge Tesla has in selfdrive technology.
3. Startups in the ride sharing space
Uber is putting a lot of resources into selfdrive technology. It makes a lot of sense for them to pursue this avenue. In order to build a sustainable ride sharing business you have to think autonomous driving. Sine Uber is private we don’t know much about their progress. But I would not count them out. They have a huge incentive to develop driving robots. They have software at their core and they have no legacy business.
4. Tech companies like Google and Apple
Google has a selfdriving technology. They will try to implement the technology along the Android model. Google wants to be the operating system of selfdrving cars. In this context Google wants to replicate their approach in the smartphone market. While this makes a lot of sense for Google it would be disastrous for OEMs. Just look at the smart phone OEMs using Android. It’s a zombie space with lots of profitless companies fighting for features nobody really cares about. We don’t believe this model will work because it’s not sustainable. But don’t count Google out. They have muscle and stamina.
Apple looks like the 800 pound gorilla always ready to penetrate the space. I personally would love to see a car designed by Apple. There are no indications that the company is close to releasing something. But they are working on it. Most likely it will be an all in solution like the Iphone where Apple takes care of software and hardware and buys crucial components like sensor technology.
Mobileye is another company with a stake in selfdriving cars. Mobileye is good at computer vision. The problem with Mobileye is that first, the stock is very richly valued. Second, the strategy of Mobileye is to take the car industry hostage and make everybody dependent of Mobileye technology. They want to be the platform that everybody uses. While laudable, the strategy is hard to implement. Nobody want so be hostage to an obscure tech company. Still, the computer vision piece of Mobileye is strong and might play a role longterm. But at >20 times sales the valuation is not interesting for investors.
We believe the selfdriving car technology is a huge shift and will generate a lot of wealth. Currently we see opportunity in shares of Intel and Tesla to take advantage of this shift. Nvidia has a role to play but we consider the shares as too pricey. The eternal risk with Nvidia is that competition catches up with their GPU technology and erodes their market share. Mobileye is good at computer vision but doesn’t follow a sustainable strategy. Apple and Google are not interesting from the perspective of selfdriving cars.