If there was more screen real estate this section would be titled, “Why Uber is undervalued at ~$68 billion, why they might be the largest company in the world in twenty years, how they and their competitors are going to change the entire world, and why none of these statements are hyperbole.” Allow me to explain these dramatic claims.
When the first production autonomous vehicles come out, they will be very expensive and essentially novelties for the rich, or commercial endeavors undertaken by fleet operators. Uber and other ridesharing companies will be first in line to purchase a trial fleet of autonomous vehicles, regardless of the cost, to use in a pilot program to test their viability for replacing their human drivers. Several years ago Travis Kalanick, the CEO of Uber, stated that if Tesla met their goal of producing half a million vehicles by 2020 he’d buy every single one of them, but only if they’re autonomous.1
Ultimately Uber chose to develop their autonomous technology in-house and partnered with Volvo to provide the automotive platform for their Uber ATG, as seen below.
Now, even if these autonomous vehicles cost hundreds of thousands of dollars, which is quite likely in the beginning, that will still be much, much cheaper than Uber handing 75% of its revenue2 back to its drivers. Right now Uber is paying its drivers $7 billion/year and the company doubling in size every six months.3 That’s a lot of money to otherwise spend on buying autonomous vehicles and infrastructure to support them. And almost regardless of how expensive the autonomous tech is in the beginning, it still amortizes out to be much, much cheaper than paying human drivers for every minute and mile they’re transporting a fare.
Studies have shown that car sharing services like Zipcar are already reducing car ownership by 50%4 (and allowing consumers to enjoy commensurate fiscal savings) just through the power of the sharing economy, and that’s using a sharing model that’s fairly inefficient. With an efficient and mature autonomous car sharing fleet, the economics are much better for the consumer. This slide shows how long it’s going to take Uber to recoup the costs of an autonomous vehicle investment at different price points and usage patterns relative to the price of Uber’s human labor rate.
The first thing to be aware of is that this data is being modeled on cab usage statistics, because Uber at the moment is more than anything analogous to a taxicab. So we assume the vehicle is usable for 6 years, like a New York cab. In most markets the demand for cabs during peak times is enormous, whereas the demand at 4 am is very small, although not non-existent. Because the demand for shared transportation is highly variable, I’ve analyzed the economics of paying off an autonomous vehicle taking into account that highly variable demand.
This is the economics of paying off an autonomous vehicle relative to the cost of human drivers at various price points and usage patterns. The economics of the payoff period are very different depending on whether you use the vehicle for 4, 18 or 23 hours/day. You can see that even if the autonomous vehicle costs $150K/year, if you put it to use for 18 to 23 hours/day, and there is a small market for round the clock usage, it will still pay for itself in about a year. That means that after paying the running costs these vehicles are just generating profit. If you only used it 4 hours per day, which would correspond to only being used at peak times, it would take almost the whole 6 years to pay it off. If you could buy a $50k autonomous fleet vehicle and you used it 12 hours/day, which isn’t too far off actual cab usage, you could pay off the vehicle in a little over 7 months. But that price point isn’t going to be available for quite some time.
Autonomous vehicles are going to be very expensive at first. So expensive that in the beginning it will only make sense to invest in them for heavy usage. But as the vehicles get cheaper the payback period gets shorter and shorter. All this adds up to the reality that once Uber is using autonomous vehicles they will eventually be able to cut the price of their service substantially, especially after many generations of technology iteration and massive economies of scale, such that the autonomous vehicles cost $50K rather than $150K. Uber won’t want to reduce their prices that drastically but market forces will force them to do so.
Once Uber’s service is that cheap, their business will grow exponentially. That market growth will allow them to service suburban and residential markets with the same speed and convenience with which they’re currently serving urban environments. Fast forward that scenario a decade or more and now Uber’s entire international fleet is autonomous. This means they’ve probably purchased millions of autonomous vehicles by then. That volume of purchasing will have a substantial impact on the manufacturing economy of scale of autonomous vehicles (and electric vehicles), which will massively reduce the price of those technologies.
Now’s the point where you’re probably expecting me to say that Uber’s volume of purchasing will bring the price of autonomous vehicles within the reach of the middle class, thereby causing the tipping point where autonomous technology becomes widespread due to the middle class buying them in volume. Well, I’m actually not going to say that because I don’t think that’s the way this will unfold. I believe that bringing autonomous technology within the purchase grasp of the middle class will ultimately be irrelevant, because all this time that Uber was bringing down the price of autonomous technology through volume purchasing, they will also have been quietly causing the greatest mobility revolution since Henry Ford democratized transportation with the Model-T. This mobility revolution is going to be one of the greatest disruptions American society has ever experienced. Allow me to explain.
This is a graph that shows how much Uber could charge for its services if it could buy autonomous vehicles at various price points, given certain reasonable assumptions about a vehicle being used like a taxi, amortization periods and other things.
This graph is based upon estimates of the autonomous vehicle being used like a taxicab (vehicle replaced every 6 years, used a little more than 12 hours/day, ~70K miles/year, etc)5, combined with known payout schedules currently being done with human drivers, combined with leaked Uber financials from 2013.6
The first thing to be aware of is that Uber’s current aggregate charge is on the left at $1.81/mile. That’s about three times the cost of private car ownership and it’s that expensive because right now Uber users have to pay to rent the driver’s time and the car. Turns out that’s equivalent to buying autonomous fleet vehicles that cost $412K. Follow that? In case it’s not clear, Uber’s human labor rate is equivalent to purchasing autonomous vehicles that cost $412K. Right now as Americans we pay about 66 cents per mile to drive on average. That’s equivalent to having a fleet of autonomous Ubers that all cost $160K. Using autonomous vehicles that cost anywhere between $412 and $160K just makes Uber cheaper and consequently grows their market even more. This will allow them to service suburban and residential markets with the same speed and efficiency that they’re currently serving urban markets. So autonomous fleet vehicles priced between $160K and $412K makes Uber cheaper and further explodes their market share, but otherwise its business as usual.
Things start to get interesting when Uber can buy autonomous vehicles cheaper than $160K, though. Because being able to purchase vehicles below that price allows them to start undercutting the average price of private car ownership. If the day ever comes when they can buy autonomous fleet vehicles for $75K, at that point they can offer transportation for less than two-thirds the price of privately buying cars. If we ever get down to a $25K fleet vehicle, now Uber is offering their service for a third the price of private car ownership. Undercutting the price of private car ownership in this manner will have profound social and economic consequences, as detailed in the next section.