BEGAN THE FLEET WARS HAVE
The business of making an autonomous vehicle (AV) Vs. Making an AV-based business
Let me take you on a short trip to 2050, a San Francisco summer, and we are cruising down Market St in an air-conditioned robotaxi. I point to the cabin screen and say “Look at how this taxi tracks the other cars, it is using the latest …” wait, this doesn’t sound right… we will probably not be talking about the taxi now, would we? Fact is, that in a future AVed world, AVs make a very boring conversation, they are horizontal elevators.
Back in 2020, the business of developing AVs is thriving. Brilliant engineers are working in numerous startups on the different parts of the technology mix — radars, lidars, AI for the edge, smart data annotation, cybersecurity, and the list goes on; they are all in the business of problem-solving, developing solutions destined to become parts of an automated driver. The business model is clear — develop a product and then sell lots of it to many customers. They might also get acquired along the way, by a customer who fancies exclusivity, which is also a good outcome.
Now let’s examine the companies developing the entire automated driver, their business model is less obvious, so let’s dive in:
Technology’s maturity: if you want to make God laugh tell him about your plans, and if you want to make an AV engineer cry tell him that you found a new edge case scenario. The only way to launch a self-driving car (during our lifetime) is to make few restrictions over car’s operation (a.k.a Operational Design Domain) — only summer, only highway, only paved roads, no Dutch traffic signs. Master Yoda would say “a car as you know it, it is not”. This situation, of a complex robotic product, with an even more complicated user manual, is not reserved only for AVs; Some (working) products cannot be sold (or cannot be sold just yet) and the common practice in these cases is for the developer to operate his own product to sell its utility as a service; This model lets the developer monitor performance, find bugs, update software and restrict the usage as she sees fit, in a scrutiny-free environment.
Value creation and capture: how would you price an automated driver? One could argue that since it runs 24/7, it worth at least 3 times the annual salary of a professional human driver, every year. Someone else could argue that it worth the COGS (Cost of Goods Sold) of its sensors, chipset, wiring, and brackets + 10% for the manufacturer’s profit margin. In other words — if you were the developer of a self-driving car, would you sell 10,000 such cars to Uber for a fixed price, or would you rather register 10,000 cars as freelance drivers in Uber’s network to make 75% of the fare on every ride? Let’s switch places, now you are Uber, what would you do now? wait a minute…..is this why you are developing self-driving capabilities for yourself?…a ha! So it seems that the best answer for pricing is to try avoiding the question.
Developing a digital driver is unbelievably expensive, and the fruits of this hard work are sweetest, and ripe sooner, if the developer is also a fleet operator (a.k.a Transport Network Company). On the flip side, this could block the developer from doing business with other fleets, so what is more important? profit margins, time to market, or market share? The term market-share has this nice aroma of world-domination, so just bear in mind that in the robotaxi business, expanding to the next town is more complicated than franchising McDonald’s.
In reality, the two-sided tension I described is a 3-way tango between automated driver developers, automotive OEMs (carmakers), and Fleet operators (TNCs), each of which has all the reasons to believe in being the chosen one, the one best positioned to capture the lion share of the value. The carmakers (e.g. VW, Ford) are best positioned to design, build, and maintain the perfect city transporter; They are the automotive experts, and when it comes to owning big fleets — they only pay the COGS, for both cars and spare parts. The rideshare companies (e.g. Uber, Lyft) own the demand, they bring in the customers, and they swipe the credit cards, need I say more? The developers of the automated drivers, well… if it weren’t for them, we wouldn’t have this conversation, and we need them in the long run, optimizing the automated driver will be a never-ending story. So maybe everyone should just work together, each on his part, and we’ll split the pie three ways, 1/3 for each; If only things were that simple. With so much in stake, investors and C-level sponsors want to know what payday will look like, and it cannot depend on the goodwill of others. Why should anyone give up profit to a carmaker, when you can always find another carmaker (in China?) to buy from? The Ubers and the Lyfts also suffer a severe disadvantage with zero customer loyalty, by users who just want to get from A to B, for less; And I am sure you don’t expect the horizontal elevator companies to dominate this market, they will be white-labeled once the novelty expires, right? Ever wondered why automated-driver companies invest heavily in brand awareness? The answer is written on your laptop, a little sticker that says “intel-inside”. Capturing value is not a matter of greed, and I have never heard a CEO downplay a partner, but getting a good chunk of the pie is an existential need for covering costs and sustaining a business, and so… the plot thickens.
“So certain were you. Go back and closer you must look” / Master Yoda.
At this point, I have to apologize for oversimplifying the situation, I might have given you the wrong impression that the fleet operators have the easiest job, which couldn’t be further from the truth. The whole idea of a robotaxi is to get you from A to B without a human driver, so if the car drives itself, we can take out the human driver! Right?…Wrong! As it turns out, the person behind the wheel does more than just controlling the car and complaining about the government; She is also the first responder to any medical situation if such occurs inside the car, he is a valuable source of information, a local guide, for the commuting tourist, she is the sheriff, the keeper of peace, in a pay-per-seat shared ride to the airport, he will wait to see the light turning on in the staircase after dropping off a young lady at night, she is attentive to any squires and whistles coming from under the hood, he will be efficiently creative if the drop-off spot was taken by a FedEx truck, she will make sure the passenger took his belongings, he will clean the mess left by the previous group of riders, and she will cut through the parking lot for her passenger to catch the last train. This is a (very partial) list of tasks, to be handled by the robotaxi service provider. Fulfilling these requires teleoperation capabilities, 24/7 service center, on-call maintenance teams, automated in-cabin monitoring. I will stop here, but we only scratched the surface, it gets much worse (or better, if you are the one who figured it all out)
Lets recap, developing an AV is hard and expensive, same goes for manufacturing a reliable car, and so is developing the capabilities to operate a robotaxi service. Beyond development, the capital investment for owning and running such service is imaginary, and will only make sense in large scale deployments. Handling all of this is too much for any single company, both financially, and in terms of management focus, and DNA (Although the name of a specific company does come to mind). All of this would not be possible without the development of AVs which are merely horizontal elevators. So if the market will be fragmented, what will be the split of captured value across the vertical? And don’t forget the other competition, between developers of alternative solutions at each stage of the chain. BTW, no discussion can be complete without mentioning the Tier 1 suppliers which play a pivotal role in this transformation, but I’ll leave it to another post.
“Difficult to see. Always in motion is the future” / Master Yoda.
Reality tells us that no one feels safe, and no one puts all the eggs in a single basket, Ford-VW-Argo, GM-Cruise, FCA-Waymo, Uber ATG, Lyft level 5, companies work together while keeping their options wide open, in case their partners will not become the winning horses; they also explore into adjacent turfs, either “for real” or just to send a signal to their future partners. In 2018 OEMs shopped for technology, and in the current markets, I will not be surprised to see a tech giant buying a carmaker. If it wasn’t clear till now, I personally admire everyone who works and invests in the future of mobility, I am using the term horizontal elevators just to remind myself that all the magic under the hood, which gets me excited, is not a customer benefit. Society as a whole will profit from this technology, and I wish we could say the same about all the companies that contribute to the effort. It’s complicated, there is no silver bullet, you can win in tech but lose in business, you can base a business on the tech of others, and trying to be the first might kill you, leaving the riches to the procrastinators. Oh boy…I miss the days when being a great engineer was enough.
If fear is the path to the dark side, we have every reason to stay in the light. The business opportunities in the future of mobility are huge, with so many verticals to serve, this is not a winner-takes-it-all game. Not everyone will make it to the finals, but everyone sure does enjoy playing; and how can one not, if working towards a noble cause, surrounded by brilliant people.
The magic spark, born during the DARPA Grand Challenges lives on and catches on new souls every day. Remember, the force will be with you, always.
