We at Alphaville have for a long time argued the economics of self-driving taxi fleets don’t work.
If you’re still not convinced, we recommend Uber’s IPO docs, which were out Thursday.
First off, a cursory reading of the documents implies the main reason Uber is being propelled into the self-driving race is not necessarily because it thinks it’s going to work for them, but because it fears if it doesn’t the likes of Google and Tesla will eat their lunch.
From the S-1 (our emphasis):
We have invested, and we expect to continue to invest, substantial amounts in autonomous vehicle technologies. As discussed elsewhere in this prospectus, we believe that autonomous vehicle technologies may have the ability to meaningfully impact the industries in which we compete. While we believe that autonomous vehicles present substantial opportunities, the development of such technology is expensive and time-consuming and may not be successful. Several other companies, including Waymo, Cruise Automation, Tesla, Apple, Zoox, Aptiv, May Mobility, Pronto.ai, Aurora, and Nuro, are also developing autonomous vehicle technologies, either alone or through collaborations with car manufacturers, and we expect that they will use such technology to further compete with us in the personal mobility, meal delivery, or logistics industries. We expect certain competitors to commercialise autonomous vehicle technologies at scale before we do. Waymo has already introduced a commercialised ride-hailing fleet of autonomous vehicles, and it is possible that other of our competitors could introduce autonomous vehicle offerings earlier than we will. In the event that our competitors bring autonomous vehicles to market before we do, or their technology is or is perceived to be superior to ours, they may be able to leverage such technology to compete more effectively with us, which would adversely impact our financial performance and our prospects. For example, use of autonomous vehicles could substantially reduce the cost of providing ride-sharing, meal delivery, or logistics services, which could allow competitors to offer such services at a substantially lower price as compared to the price available to consumers on our platform. If a significant number of consumers choose to use our competitors’ offerings over ours, our financial performance and prospects would be adversely impacted.
And yet, the documents also reveal even Uber can’t envisage a world where self driving fleets will ever be able to cater to all demand scenarios, and hence it expects to be reliant on human drivers (and their capital) for a long time yet.
For example (our emphasis):
Moreover, high-demand events, such as concerts or sporting events, will probably exceed the capacity of a highly utilised, fully autonomous vehicle fleet and require the dynamic addition of Drivers to the network in real time. Our regional on-the-ground operations teams will be critical to maintaining reliable supply for such high-demand events. Deciding which trip receives a vehicle driven by a Driver and which receives an autonomous vehicle, and deploying both in real time while maintaining liquidity in all situations, is a dynamic that we believe is imperative for the success of an autonomous vehicle future. Accordingly, we believe that we will be uniquely suited for this dynamic during the expected long hybrid period of coexistence of Drivers and autonomous vehicles. Drivers are therefore a critical and differentiating advantage for us and will continue to be our valued partners for the long-term. We will continue to partner with original equipment manufacturers (“OEMs”) and other technology companies to determine how to most effectively leverage our network during the transition to autonomous vehicle technologies.
This is important stuff. It pertains to our observation that autonomous fleets would become highly commoditised very quickly, and as a result be extremely sensitive to price swings and capital costs, sabotaging most operators’ profitability.
For example, to cater to high demand scenarios an operator would have to bear the huge capital cost of acquiring and maintaining excessive car fleets even though they will stand idle for some of the week. This is a likely loss generating exercise, as it’s highly improbable the operator’s frequently used inventory would be able to make enough money to cover the idle stock’s costs.
Even if the purchased stock of autonomous vehicles was utilised at full capacity, this means an operator absorbing far higher depreciation costs per annum. The bull case for self-driving cars often starts with the fact that “cars are parked 95 per cent of the time”, and thus a crucial stock of private capital is perennially underutilised. But underutilisation also elongates the life cycle of a vehicle for the owner. The reverse will therefore shorten it — if a car is in use 100 per cent of the time, how long will it last before it has to be written off or substantially repaired by an operator? In this scenario, there’s one sector self-driving is certainly bullish for: the auto manufacturers.
Uber’s solution to the problem seems focused on purposefully undersupplying the market so its self-driving fleets are reserved only for high-use periods, drawing on human drivers to plug the supply gap. This, at first sight, seems a clever way to transfer the capital costs of car acquisition, depreciation and maintenance to other entities. But the reality is that it also alienates the professional driver who depends on a reliable income to make ends meet. To compensate for the lack of dependable cash flows, surge prices would have to get exceptionally high to make it worthwhile for drivers to hang around on standby until high demand scenarios materialise.
Even higher and more erratic surge prices, however, are likely to spook customers who have become used to Uber’s affordability. They’re also likely to turn Uber into a mostly off-peak service, potentially undermining its vision of becoming a genuine substitute for personally owned vehicles or public transport.
Some might argue that the market will adjust to these issues. Professional drivers will disappear and be replaced by amateur multitasking jobbers who come in to fill gaps only when needed. Aside from the fact that this sort of jack-of-all-trades vision challenges the very notion of how efficiency is achieved under division of labour, it relies on the assumption that such jobbers already have cars to access. While it’s true that in medieval times, a freelance knight’s prospects were intimately linked to their horse-owning status, it’s also true those were medieval times, thus not necessarily times to be emulated in the spirit of efficiency and modernisation. More so, medieval knights didn’t have to surrender some share of their earnings to autonomous horses while compensating by couriering peace treaties or flipping burgers at medieval festivals.
Though, as least we can be thankful Uber hasn’t gone down the Elon Musk route of suggesting the capital gap should be plugged by private autonomous vehicles, rented out when not in use. For more on why that wouldn’t work, read here. (But in short, just like refineries — which only swing into action when the price of gasoline is high enough to generate enough of a margin to make it worth their while — the same would apply to the private owners of autonomous cars. The likelihood that this price is any more competitive than the price that attracts physical drivers is fairly low. There’s also the problem of sharing your private car with an unsupervised stranger, of course.)
Driver dissatisfaction: don’t bite the hand that feeds you
Uber’s IPO documents acknowledge how tough the transition to an autonomous network is going to be. Most glaringly, there’s the driver dissastifisaction they are likely to face as they begin to transfer models. From the docs:
Further, we are investing in our autonomous vehicle strategy, which may add to Driver dissatisfaction over time, as it may reduce the need for Drivers. Driver dissatisfaction has in the past resulted in protests by Drivers, most recently in India, the United Kingdom, and the United States. Such protests have resulted, and any future protests may result, in interruptions to our business. Continued Driver dissatisfaction may also result in a decline in our number of platform users, which would reduce our network liquidity, and which in turn may cause a further decline in platform usage. Any decline in the number of Drivers, consumers, restaurants, shippers, or carriers using our platform would reduce the value of our network and would harm our future operating results.
The above is a huge clue as to how dependent Uber is on its drivers, not just from a service provision perspective but also from the perspective of transferring car fleet capital costs to them. Given the above, it’s hard to envisage how the “humans on the margins” solution is ever going to be accepted.
But if that wasn’t enough, there’s the sheer capital cost of developing the technology in the first place, as well as the unexpected liabilities. According to the IPO docs, Uber spent $475m on autonomous fleet development in 2018 alone. This experimentation also resulted in one well documented public death in March 2018, emphasising the “significant risks and liabilities” for the ride-hailing platform.
In terms of the future capital costs …
But here’s the really important factor for would-be buyers of the stock on IPO day. Uber says autonomous driving is essential for it to continue to effectively compete, but it also says these development efforts are capital and operations intensive (the opposite of its supposed asset-light business model today).
And yet, as they also note (our emphasis):
… we may require additional capital to expand these products or continue these development efforts. If we raise additional funds through further issuances of equity or convertible debt securities, our existing stockholders may suffer significant dilution, and any new equity securities we issue may have rights, preferences, and privileges superior to those of existing stockholders. Certain of our existing debt instruments contain, and any debt financing we secure in the future could contain, restrictive covenants relating to our ability to incur additional indebtedness and other financial and operational matters that make it more difficult for us to obtain additional capital with which to pursue business opportunities. For example, our existing debt instruments contain significant restrictions on our ability to incur additional secured indebtedness. We may not be able to obtain additional financing on favourable terms, if at all. If we are unable to obtain adequate financing or financing on terms satisfactory to us when required, our ability to continue to support our business growth and to respond to business challenges and competition may be significantly limited.
That alone is the sell case in one neat paragraph.