Pricing for AV TaaS — Analogues from the Energy Sector
It was announced recently that Waymo has started experimenting with prices in Phoenix for rides in its Autonomous Vehicles (AVs). It also became the first company to get a DMV permit to test a full AV solution without a human safety driver in CA. Cruise, a GM subsidiary, has also mentioned to launch it’s AV service to early riders in San Francisco bay area in 2019.
We’re in the early innings of autonomous Transportation as a Service (TaaS), a model that Uber and Lyft have pioneered in the US for some time. For TaaS, and more so for AV TaaS, pricing and monetization of these services will become more important. How might this be structured? I think there are several analogues here to the energy grid — a service that’s taken for granted to be always present. One can argue that it’s a commodity, which TaaS at scale might also be, that’s why customer experience will be important (a topic that I’ve written about earlier here).
The three main elements of the energy supply are generation, transmission, and distribution. In a regulated monopoly, the rates for residential, commercial, and industrial customers is determined by a regulatory process, overseen by a Public Utility Commission (PUC). While in deregulated market, like Texas in the US, customers are free to choose electricity from variety of Retail Energy Providers (REPs), also called energy retailers. The energy rates for residential and commercial customers are typically higher than industrial customers, big factor of which is infrastructure investment in stepping down the voltage at the distribution level.
For the purpose of article, we’ll use residential / commercial customers in deregulated markets as an analogue.
The starting point of the energy price is the wholesale price. The wholesale price varies by different generation sources (e.g. combined cycle gas, coal, nuclear, solar) and usually, Levelized Cost of Energy (LCOE) — NPV of the unit cost of energy over the lifetime of the generation asset, is good for comparing costs of various options. Various generation assets bid in a day ahead market in blocks, and are fulfilled starting with the cheapest, until the expected demand is aligned with supply. There is no analogue to centralized generation to TaaS, but a fleet of vehicles — with associated optimized size, uptime, comes as a close proxy. This sets the wholesale price for that fleet service. It’s important to emphasize that both in energy and TaaS, these will be in unit of service to the customer — $ / kWh and $ / mile, respectively. There are several analyst estimates, including Deutsche Bank, that forecast the cost of operating a AV TaaS fleet at ~$0.5 / mile (I’ve written about it earlier here).
The energy costs to end customers is significantly higher than the wholesale price. Several factors affect the end rate, including:
- Tiered rates: electricity rates are normally tiered, with a low rate for the lowest kWh consumption bracket, and then progressively increasing as the consumption increases. More infrastructure investment need to be made for supporting higher energy demand. In fact, PG&E rates for residential customers can be more than 2x for higher tiers. Similarly, it’s expected to have tiered rates for TaaS, which will be inversely proportional to miles. Longer rides will cost lower $ / mile because the asset (vehicle) can be amortized over longer distance and have higher uptime.
- Demand Charges: Energy demand charges (in $ / kW) is incurred when there is an uptick in demand across customers at once (e.g. hot day in July and majority of the ACs are turned on simultaneously). These charges cover utilities fixed costs — e.g. natural gas peaker plant, to satisfy customer needs. However, these assets might only be called up for <10% of the time is a year. This is analogous to surge pricing in TaaS, where in more drivers choose to participate and provide that ‘peaker plant’ service, and might choose to do it only a few hours per day. How will this be managed in an AV TaaS fleet is yet to be seen, but don’t be surprised to see a ‘peaker plant’ analogue, new business models, as these services scale.
- Time of Use (TOU): TOU rates depend on not just how many kWh customers use, but when they use it. The intent is to drive the customer behavior so that overall energy demand stays aligned with supply. That’s why doing laundry in off-peak time can help lower electricity bill. For a given AV TaaS fleet, we can expect similar rate structures to smoothen demand over time. In addition to managing variability during a day, expect these pricing mechanisms to evolve to account for weekday / weekend and also different seasons.
There have also been incentives for adoption of new technologies in the energy sector — e.g. Net Energy Metering (NEM) for solar. Solar, in addition to better carbon footprint is also expected to alleviate the strain on the grid. In fact, these incentives have been so popular and driven the adoption of solar that they have resulted in a new problem — CA’s duck curve. It lowers the demand on the grid during the day when solar production is high, and shifts it to the evening when generation assets to have to ramp up in 1–2 hours. Interesting, we’ve seeing high adoption of newer modes of transportation — e-bikes, e-scooters, for shorter commutes in urban areas. As these solution cut into the demand of TaaS, we’ll start seeing mobility sector’s version of duck curve, with the belly at around 2–3 mile commute distance.
Beyond rate structures that determine the end customer price, there are couple of additional energy industry analogues that are relevant for AV TaaS.
Because there is not a large scale energy storage solution that’s commercial yet, the energy that’s generated doesn’t have a shelf life — kWh that are generated have to be matched with demand in ‘real time’. Because there is no buffer, that’s why Independent System Operator (ISO), operate in a mission critical manner to balance energy traffic and ensure reliability of the grid. For AV TaaS, expect creation of buffers that can absorb imbalance between riders and available vehicles. Expect these assets to be utilized for secondary uses — e.g. goods delivery, in addition to moving people.
Given the competitive landscape of energy retailers, customer stickiness is one of the key pain points. Customers are typically bound by short term contracts (typically year or less) and often switch to another retailer at the end of the term, often for lower price. That’s why leading retailers are highly acquisitive and provide bundle of services — e.g. Direct Energy, a business of Centria, offers energy, natural gas, water heater rentals, HVAC services, home warranty products, among several others. Customer stickiness with a key for AV TaaS as well, and it will become even more important to have several engagement points with the end customers and provide basket of services. Uber’s acquisition of Jump, to have an e-bike option on it’s platform, and aspirations to be ‘Amazon of Transportation’ provide clues towards that goal.
The industry parallels between energy and transportation are hard to overstate. As the AV TaaS scales, the learning from the energy sector provide relevant data points for the investor to place right bets, companies to develop accretive M&A strategy, and service providers for better monetization strategy.