Do Robo-Taxis Really Make Sense?
Kjeld Lindsted
Head of Products & Sales | Connected Vehicles & IoT | 10X TAM growth via product strategy innovation
Starting in about 2016 I got hugely excited about the idea that autonomous vehicles were just around the corner and, when built on electric platforms, would enable a new era of robot services. I believe we'd quickly find ourselves in a Star Trek like future where all drivers were computers, all vehicles electric, and the economics of the whole industry shifted so dramatically that million mile cars could only ever be owned by taxi companies able to drive them 24 hours a day for profit. Private ownership wouldn't make financial sense, and driving yourself would likely be unfavorable (or even illegal) shortly thereafter. Afterall, how many people could afford to buy a $100,000 vehicle just to park it 95% of the time when you could rent rides for pennies on the dollar instead?
After spending the last 8 years deep in the connected vehicle industry supporting governments as they rolled out infrastructure to manage new mobility, I've learned a few key things that completely upended this picture for me. I've learned that sexy tech doesn't necessarily equate to valuable business in the real world.
AVs are hard
It turns out that computers struggle with driving. They are amazing at well-known scenarios, but real life is varied and complex and computers have a hard time making choices in unknown settings (see the amusement of the traffic cones that shut them down in San Fran recently). This might not just be a matter of more development time. I am starting to suspect that our current computing paradigms might actually not be ideal for AI anything and that driving based on a glorified logic tree approach may not ever work all that well. We might need something more like a human brain, and we've not quite built that yet.
Public Policy is even harder
But let's say AV companies do manage to survive the VC dessert we're in and keep spending billions until they get it right. Will they ever get enough public buy-in to really scale? Sure, there are lots of states who allow autonomous vehicle testing, and even commercial operations (some even without much permitting in the way). Tons of questions remain completely unanswered.
I believe we won't have much difficulty sorting out legal liability and other ownership issues often raised as red herrings, but with driving under a barrage of regulatory pressure ranging from safety to fuel economy, to charging standards, to how roadways are taxed and funded, it's not clear to me where AVs fit in profitably.
Don't forget infrastructure funding
One of the biggest challenges in transportation right now isn't actually vehicles themselves. It's the infrastructure. Building roads is expensive. Fixing them is worse. Adding technology can help in some cases, but that's also expensive. Agencies struggle with funding options for all levels of transportation investments and the public generally isn't friendly to the taxes required to fix some of these issues. The gas tax, a major source of revenue for roadways, hasn't been updated in decades and efforts to find other ways to have drivers contribute to the costs have so many hurdles it's worth a whole other article to digest half of them. Explore "road usage charging" if you're interested in this area of policy.
It's become clear to me in my work over the last decade that the challenges with transportation aren't limited to vehicle side at all and automating the car won't even touch many of the issues.
But what about Vehicles?
Setting aside that AVs might be way harder to actually do than we think, and that infrastructure constraints are a bigger hurdle to modern mobility than most realize, it's not clear to me that AVs make great taxis either way.
领英推荐
Let's talk about safety
No, not crashes, individual safety. In college, I interned in DC for part of a school year. One of my classmates in the program took a taxi home from the bar late one night. In the course of the 5 mile trip, she was nearly raped by the driver and lost her purse in her getaway.
The app-ride solutions help a bit, someone is tracking all of it at least. But talk to Uber or Lyft drivers whenever you can. They have stories too. A few I've heard over the years from actual drivers.
Passengers:
And I have no doubt this list is super incomplete, this is just what I have personally heard directly from drivers I've ridden with. With Uber and Lyft, you have a driver who's being tracked so rape and other driver crimes are a little harder. That helps over traditional taxis. You also have a driver who's paying attention to the vehicle (they own it after all), and looking after the occupants in the above types of situations. Remove that driver and see what happens at your own risk.
And then we come to economics
Bird, Lime and all their ilk were once the coolest new idea in mobility. Scooters were fun, they where electric, and to most people, they were an obvious way to get around that didn't require sweaty Lycra suits or vehicle emissions. It was once obvious that mobility as a service was here to stay. Except it hasn't quite worked out. Some of these contenders have gone bankrupt, others got regulated away. Those that remain struggle with profitability. It turns out, it's waaaay harder to make money from mobility as a service than you'd think (at least if "you" don't actually work in transportation and so view it with rose colored glasses).
One of the worst costs for scooter companies is the vehicle repairs. Turns out, when people don't own something, they don't take much care of it. Don't forget all the vandals and thieves. It also happens to be really expensive to move scooters around all the time. If only people's destinations weren't so one-way and imbalanced.
The problem with AV taxis
I often fly out of Ontario Airport (a little inland from LAX). It's the closest one to me, but still about a 45 minute drive without traffic. These days, I mostly drive and park, it's annoying to wait for a Lyft and the cost is wild. But there are a few times I take Lyft and every time, I feel a little bad for the driver. They get a nice fee to drive me home, but then they're stuck in the country, at night, with little choice but to drive themselves home on their own dollar.
That's the dynamic that I think will make Robo Taxis super hard to do financially. It sounds great to have the autonomous electric car driving for a fee 24 hours a day, but it's not realistic. Just like with scooters, there will be a significant cost to rebalancing the network. Now, the cars presumably could just drive themselves to the balance point, no problem. But there's a cost, both in time and energy, to this balancing. Add in all the issues with vandalism and theft, and the problems of personal safety (which will require monitoring and intervention support activities) and I think the end solution is likely to cost as much as ever. This is especially true when you add in the margins needed to satisfy the VC funding currently trying to build this system.
If I'm even half right, it might not be much cheaper to ride in a robo taxi than to buy your own Tesla with partial self-driving capabilities, the safety and convenance of private ownership, and the luxury of private vehicle care you invest in your belongings.
Time will tell. I'm a tech fan. I work in technology and believe in the power of computing to do much good in the world. But not every problem is actually about computing at all. Sometimes the non-digital stuff is the issue all along.