Transportation-as-a-Service (TaaS)— Cost Structure

The Information recently published an article on Uber’s 2017 financials (table below). A lot has been discussed in the press on how Autonomous Vehicles (AVs) will disrupt the traditional automotive OEM business or personal car ownership, but Uber’s recent financials give a glimpse on how AVs might also disrupt the nascent Transportation as a Service (TaaS) industry, which still is ~1% of of total Vehicle Miles Traveled (VMT) in the US.

It got me thinking on what’s the potential of AVs even for TaaS industry. Focusing on Q3'17 column, its good to think about line items on a per mile basis. With ~4B / year TaaS miles in the US, and ~77% market share of Uber, that’s about 770 million miles per quarter on Uber’s platform. That’s about 40 million rides monthly with average ride being 6.4 miles.

The Company’s revenue is not broken by regions, and though Uber is present in 83 countries, there are regional market share leaders in several local markets (e.g. Didi, Ola, Careem, Grab). Conservatively estimating 50% of the Uber’s revenue from the US and assuming same cost structure across regions, the P&L from above table for Q3'17 on per mile basis are:

Note that $ / mile revenue is higher than what’s reported elsewhere (here and here) likely due to average across all the markets in the US (vs. say top 20 metros) or revenue from other businesses, like UberEATS, which is estimated to be $3B / yr business. Even based on these conservative estimates, Uber is losing ~75 cents / mile at EBIT level.

Contrast the above table for the scenario of autonomous cars based on Deutsche Bank estimates. Again conservatively, keeping all the line items but for three (in gray) same, the financials change drastically. The big cost with driver-related earnings in eliminated, but the operating expenses due to depreciation, maintenance, fuel etc increase due to higher upfront cost of the vehicle and higher utilization. Obviously, there are technology and regulatory constraints for across the board deployment, but the table below shows the potential of AV technology to Uber’s (or TaaS service provider) bottom-line.

Though not discussed much, the emphasis of cost structure for an AV TaaS will shift to Operations and potentially customer acquisition. 3 potential areas are:

  1. Maintenance of the fleet — the world of AV TaaS will look very similar to the current airline industry. The service providers (brands the customers interact with) will strategically outsource maintenance of less-critical (e.g. vehicle cleaning), while keeping strategic components (e.g. battery maintenance and repair, proprietary sensors) in-house. Maintenance, Repair, and Overhaul (MRO) is 12–15% of airline operating costs — expect a bigger emphasis here as AV fleets get deployed.
  2. Useful life — very closely related to (1) above is, vehicle design and maintenance schedule that leads to long useful life. A key component of that will be Electric Vehicles (EV), which due to their simpler design has less than 150 moving parts vs. 10,000+ for a typical internal combustion engine (ICE). It’s not a surprise that Tesloop, which rents Teslas on it’s fleet, has vehicles that have clocked 300,000 miles in 2 years at 1/8th the cost of a ICE vehicle.
  3. Vehicle Utilization— personal cars today at utilized for ~4% of the time, with perhaps 3–5x for vehicles on a TaaS platform. Vehicles on AV TaaS fleet will need to be utilized at 60–80% of the time. Expect new business models when these vehicles are utilized for goods delivery, particularly for last-mile, in non-peak hours when they are not moving people.

The TaaS of the future will be significantly different from what we’ve today, not just for the bottom line of the service providers but also for customers. For having a scalable AND profitable fleet, the Operations and associated costs will become more important. Expect more discussion as Waymo, Cruise, and other services get launched and scaled.