Uber self driving crash

I have blogged a number of times about self-driving cars being the key to working out the path for oil prices and so it is important to work out if the recent pedestrian death by an Uber self-driving car is a sign that self-driving cars are further away than we expected.

Electric Car Economics

An electric car is more expensive than an oil powered car (due to the cost of the battery), but cheaper to run (oil is expensive per unit of energy, lower running costs due to fewer moving parts, less energy loss with batteries and more efficient power generation).

This means the economic calculation is about amortising the cost of the battery over the distance travelled. 

By my numbers, the current cross-over is about 100,000km per year and falling with decreasing battery cost. This means that a private electric car is not economically better (because most drivers don’t drive 100,000km per year), but an electric taxi is.

Levelised cost of driving

Driverless Taxi Economics

Self-driving cars will be very cheap as the most expensive component (the driver) is removed. The simplest way is to look at car hire companies:

  • Car hire companies are happy to rent a car to users for $60 a day with about a 75% utilisation rate = $45 a day to the car company. Let’s call that the capital cost of a car.
  • Now let’s add a third to the cost to cover the battery and the driverless technology – puts us back at $60 per day.
  • Given its driverless, it can operate for 24 hours a day – but call it 15 hours to take into account time to charge the battery, cleaning, slow periods, servicing and to make the math easier.
  • So, the car owner needs to make $4 per hour to make a return.
  • Assuming 1.5 trips per hour (roughly the average), add in electricity costs, 10% to the booking company and about $3.00 per trip is the end cost to the user.

It is pretty compelling. Basically, a driverless taxi would be cost competitive with (human driven) public transport.

The timeline

There are a range of tests currently being run. The ones I am looking at most closely include:

  • Waymo/Google have done the most testing and have progressed to testing without human backup drivers in Arizona. They are planning on rolling out 20,000 taxis in partnership with Jaguar Land Rover starting in late 2018.
  • GM/Cruise are possibly more technologically advanced than Waymo with an integrated (i.e. manufacturing + technology) solution in the US. They are targeting a 2019 rollout and have a petition to change laws for driverless cars in the US which will be ruled on in coming months.
  • nuTonomy is running pilot operations in Singapore, is looking to expand to 100 cars in 2018. Singapore conditions (a small heavily populated land mass, no snow and relatively autocratic government ) are one of the most likely to support driverless cars in initial stages.  
  • Easymile has been running driverless bus trials (on mainly fixed routes) in a range of cities around the world.
  • Baidu (AKA China’s Google) is investing heavily and has released a free software platform (Apollo) which allows other Chinese car companies to use Baidu’s technology. Given government support and the Baidu platform, China will likely be a key player in driverless technology.

The Crash

A pedestrian walking a bike, not on a pedestrian crossing, was killed by a Volvo using Uber’s driverless technology in Arizona. The vehicle appears not to have seen the pedestrian and the safety driver appears not to have been watching the road. Uber has ceased operations while the investigation continues.

The question is whether this will set the driverless community back several years or not. 


There are lots of companies trying to develop driverless technology and standardised information is hard to get.  One rough guide is the Californian statistics which show the number of “miles per disengagement” – the number of miles before a driver has to take control of the vehicle:

Source: Arstechnica, Calfornia DMV

Now there are lots of caveats to this chart. It only contains Californian statistics, companies self-report and so could be under-reporting and it says nothing about types of roads or conditions. But Waymo/Google and GM/Cruise are clearly a long way ahead of the competition. 

Uber is not on this chart, Uber reportedly have internal goals to reach 13 miles per intervention. Waymo/Google are at 5,600 miles per intervention and 1,200 miles for GM/Cruise.

A recent report by Navigant also has Uber well behind other players in the space: 

Finally, there are suggestions from internal sources that Uber has been cutting corners and may not be as safety conscious as the traditional car makers.

Similar claims arose about Tesla following a high profile crash last year.


While there may be some extra caution from regulators, given Uber appears to be considerably behind the market leaders it would seem that this would be an issue specific to Uber.

The outcry over this death doesn’t seem to support banning trials, and I’m guessing that even if the investigation is damning that Uber will be singled out for restrictions rather than the entire industry. 

It seems commercial trials will start this year, and my expectation is that the takeup will be quick. Given the current trajectory, by 2025 the majority of new cars are likely to be electric, driverless and taxis. 

Its hard to prejudge the winners given the economics are still unknown. But the losers are more obvious. Oil. Car mechanics. Anyone who drives for a living. Possibly the entire car sales retail chain. Prestige cars will be more resilient.



Damien Klassen is Head of Investments at Nucleus Wealth.

The information on this blog contains general information and does not take into account your personal objectives, financial situation or needs. Past performance is not an indication of future performance. Damien Klassen is an authorised representative of Nucleus Wealth Management, a Corporate Authorised Representative of Integrity Private Wealth Pty Ltd, AFSL 436298.