About the buzz of Uber autonomous e-scooters
A couple of weeks ago the mobility buzz was capitalized by Uber’s announcement to start developing autonomous micromobility assets (AKA bikes and scooters).
Throughout this post, we are going to try to give some light to this announcement. We are going to do this showing some relevant data to have in mind. It is clear that with this technology micromobility providers can improve a lot in two of their most crucial fields: operations and a good integration with the city curb. Imagine you can summon an e-scooter that comes from a hub near you without disturbing other sidewalk users and without cost in human operations. A really huge promise.
The initial problem is a matter of balancing. Right now around the globe, there are not a single two wheels vehicles that are in hand of users that has self-balancing capabilities. Honda and other OEMs (BMW, Yamaha) have shown self-balancing-AV motorbikes. But is far from being commercialized.
Source: Honda Riding Assist
You might remember this April’s fools joke by Google; the self-balancing bike in the Netherlands. It was a joke, a funny and clever one but no more than that. Having in mind the advantage this functionality and other ADAS functionalities can give honda against its competitor the answer about why is not in mass production is clear: it is not ready yet.
At this point, it is important to mention a Singaporean startup called Scootbee. They have been promising AV scooter since a year ago, and have shared a video days ago of one of its scooters crossing a road by itself.
Let’s go by parts. They used from the beginning a scooter that has three wheels. The self-balancing problem is solved. Well is solved for them. For other players that have started with two-wheeled assets, this supposes an enormous investment in the purchase of new assets.
Now comes one of the biggest problems; the cost of hardware needed to transform a vehicle into an autonomous vehicle. Basically, there are three types of hardware that companies are using: LiDAR, Sonar, and cameras (computer vision). Most of the players are using the three, but Tesla, for instance, is not using LiDAR. The cost projected by City Group for a Lv. 3 and Lv. 4 in highways (we know these assets are not going to be in a highway, but enough level of autonomy to have the minimal performance required in the urban streets are the same).
Each column represents the total in different years: from 2019 to 2032. They are estimates made by the descent of cost production thanks to mass production and the development of new materials and manufacture techniques.
Now let’s see the economics behind an e-scooter sharing service. Below, are two economics: with and without AV HW. If you want to watch another economics visit one of our last blog post or this recent Mckinsey post.
Economics without AV HW
Economics with AV HW
Important things to note in our calculations:
- We are not including operations cost in both examples (People, combustibles…).
- In the example which includes AV hardware, we have increased costs by 30%, this is attributed to the fact that this hardware has increased battery consumption and that due to the higher asset cost, the life cycle of the scooter will be extended resulting in higher overall maintenance costs (the 1100 rides should be amortized).
There is one incognito that must be solved, are they going to use all the hardware or they can really just in cameras and computer vision? This could change a lot the economics.
Make your own conclusions with the above data. We think that this is, right now, an overpromise. Yes, Scootbee has shown the video, but let’s take it rationally. It is a 36-second video of an e-scooter moving just 15 meters. It is not a real proof of capabilities.
