November 14, 2017
Driverless cars are now arriving
by Paul Carroll
here has been speculation for years now about how soon driverless cars would hit the road, and we finally have our answer: The time is now.
With Google’s Waymo arm announcing deployment in Phoenix, we have our first real fleet of cars on the road without someone sitting in the driver’s seat. Deployment will be phased in—Waymo, for instance, still has an employee sitting in the back of the autonomous minivans, able to hit a button to have the vehicle pull over and stop if anything goes wrong. But, now that we’ve crossed the starting line, the pace of the race should only pick up from here.
Our own Guy Fraker, ITL’s chief innovation officer, says Florida is the state to watch. Guy, who lives in the Florida Keys (and escaped catastrophic harm to his house despite being almost in the center of the eye of Hurricane Irma), has been working with Florida regulators on autonomous cars for some time now. He is, in fact, speaking this week at the Florida Autonomous Vehicle Summit in Miami.
Some 34 companies are already operating autonomous vehicles in Florida, almost all in stealth mode. A main reason is that Florida regulations mean that AVs don’t require any more insurance than an Uber or other vehicle operated by a TNC (transportation network company). Guy credits Jeff Brandes, a state senator, with leading the way as Florida peels back layers of regulation to spur the use of AVs. As a result, Florida is very likely to be a pioneer, with all sorts of lessons for the rest of us — including about liability and insurance.
I expect that Phoenix and Florida provide a pretty good road map, if you will, for how AVs will be deployed through roughly 2025.
The early deployments will be in limited geographic areas—”geofenced,” to use the term of art. The reason is that to get to full autonomy, the technology at this point seems to require extraordinarily detailed maps. Those are possible within limited areas, and cities are great candidates because the roads are generally better-maintained and -marked than those in suburban or rural areas.
Southern cities will draw companies that use electric vehicles (and the AV sensors require huge amounts of power) because batteries don’t like really cold weather. More expensive cities will also draw attention because they will provide a higher price umbrella that will let AV companies charge more. Some cities may be prized because their complexity means that anyone operating there must have top technology. Uber has been operating in Pittsburgh (with “safety drivers” up front) for some time, partly because of the difficult topography, uneven quality of the roads and challenging weather. GM’s Cruise has been running increasingly extensive tests of AVs at its home base of San Francisco, with its hyper-challenging hills and narrow streets. Cruise says it will run a test fleet in what, for AVs, is the Wild, Wild East (New York City) starting next year.
Many cities and states will likely imitate Florida and try to be pioneers once they become comfortable with that the technology really is safer. So, change could come fast when we reach the tipping point.
For now, the geofencing and expense will likely limit AVs mostly to shared uses as “robotaxis,” competing with Uber, Lyft, et al. But many expenses will follow a Moore’s Law curve of the sort that has been cutting the costs of electronics by 50% every year and a half to two years since the 1960s. For instance, GM announced recently that it had acquired a technology company that might be able to cut by “almost 100%” the cost of Lidar (the laser-based version of radar generated by those devices you see spinning on the tops of AVs). I initially thought the figure was a typo or some silly hyperbole. But the company, Strobe, really does say it can cut 99% out of the roughly $70,000 cost for today’s Lidar by replacing the spinning device with just two stationary chips.
Expenses should drop so much that, as technological limitations go away, personally owned AVs should become widely available by the middle of the 2020s. That’s the point at which the huge change should start to hit insurance, as accidents plummet and liability shifts to the makers of the cars and away from the drivers.
There’s still some question about how quickly the change will hit once we get to 2025 or so. It ordinarily takes more than 15 years for the inventory of cars on the road to turn over; regulators will, I assume, do whatever they can to accelerate the use of AVs once the safety benefits become clear—how could they not?—but the pace of change is still unclear.
What is clear is that AVs are now reality. When Chunka Mui and I wrote our book “Driverless Cars: Trillions Are Up for Grabs” more than 4 1/2 years ago, we laid out an aggressive timeline for how quickly AVs would develop—and events may be overtaking us.
Fasten your seatbelts. And keep an eye on Florida.
Editor in Chief