The Autonomous Vehicles story has taken a major step forward today with GM’s Cruise announcing they have received a Driverless Deployment Permit from the California Public Utilities Commission. This permit allows them to not only operate their driverless vehicle service but for the first time, charge customers for autonomous trips in a Major Us City.
Having autonomous vehicles is a massive deal, as it dramatically changes the cost equation for both the service and the customer. For the service, not having to pay a driver, and increasing the passenger seat by 1, opens the door to a much more profitably service than the ride share services we have today.
For a passenger, we are likely to experience a far cheaper cost of travel, given the savings available from the lack of humans involved in the driving function. Compared to driving yourself, this enables people who want to go out and have a drink, to skip any temptation to drive their own car. For those business types, it means you can jump on your phone or laptop and let the car handle the driving, so if your time is valuable, this could be a great way to increase productivity on daily commutes.
With Cruise now able to offer commercial rides and the technology to operate the vehicles, you would think they would leverage this first-mover advantage in SF, to put as many vehicles on the road as fast as possible to effectively print money right? Here’s where things get interesting.
So here’s the question, why would you be rolling out fared rides gradually? This doesn’t make sense unless there’s actually an issue with the economics.
Founded all the way back in 2013, Cruise has had a long journey to get to this point. Over that time they have accumulated more than 2 million miles in California in their Autonomous development vehicles.
The company now has more than 300 all-electric AVs, but these custom Chevy Bolts used by the service need to be fitted with a large array of cameras, lidars, radars mounted on the roof of the vehicle.
This hardware array will not come cheap and is just half the picture, what you don’t see externally is the computer in the boot that’s doing the processing of inputs, making planning decisions about the best route ahead, while monitoring the environment around it.
This computer would also not be cheap and on top of the wholesale price of the vehicle, makes the total price of each vehicle quite expensive. We should also remember that AVs may remove the cost of the human driver, but there’s been almost a decade of R&D that’s gone into reaching this point.
Cruise’s technology approach is to leverage HD Maps (scans of the environment) and compare the current location of the vehicle to these maps. From videos shot by test users of the system, Cruise’s tech obviously works if they have achieved this permit to operate a commercial service, but one severe limitation is the restriction to specific geographic areas that have been scanned.
We’ve seen videos of Cruise vehicles overtaking double-parked cars etc, so the car is definitely not on rails, but what is in question is how fragile are these HD Maps. How much has to change in the environment to confuse the vehicle?
This reliance on a pre-mapped environment also places one of the biggest restrictions on the service’s ability to scale. When you hear they’ve received this approval for a paid service in San Francisco, you may imagine they’ll soon be in LA, then New York, then Washington DC, Vegas and more US cities. The reality is, it is likely to be a slow burn, particularly if Cruise the service (and each vehicle) isn’t profitable.
What is exciting is the Cruise Origin, while not due for a few years yet, this shuttle is a truly autonomous vehicle, with no steering wheel or pedals and seats that face each other, allowing passengers to converse during trips.
I’m really glad to see the autonomous vehicle space moving in the right direction, however, the important thing now to watch at Cruise is not the number of rides, but the economics, can they make this generation of vehicles turn a profit? If so, then they should make as many of them as humanly possible.