AUTONOMOUS VEHICLES

In case you haven’t already heard, Waymo made waves last week with a big announcement. They are partnering with Jaguar to buy 20,000 of their all-electric I-PACE SUVs. These vehicles will be added to their existing (and expanding) fleet of Chrysler Pacificas.

This announcement is momentous. It affirms that we are months away from the first commercial launch of an autonomous vehicle. But it also serve as a reminder of how early we are in this transportation revolution. Pressure has been building for years, but the dam has not yet broken. Despite the likes of Uber and Lyft and all the noise about autonomous vehicles, transportation has changed only at the barest margins in real terms over the past decade.

So let’s put the announcement in context. What does it mean and how does it jibe with the many claims that have been made about this technology? To do that I’m going to focus on the following points from the presentation:

• Waymo is acquiring 20,000 vehicles
  • Collectively these vehicles can complete 1,000,000 trips per day
• The vehicles are electric
• The vehicles will be manufactured by Jaguar
• The vehicles are being converted from Jaguar’s I-PACE
• NO pricing was announced

For this post, I’ll focus on the first item.

1,000,000 trips per day

Specifically, Waymo said it will be purchasing 20,000 vehicles over the next few years that will be outfitted with its driverless tech and that will collectively be capable of completing 1,000,000 trips in a day. Given the quality of the vehicles that are being purchased, this is well over a $1B deal. With Waymo’s hardware, the arrangement is probably closer to $2B in value. This a big deal and a big number, but we have a long ways to go before this technology is meaningfully impacting transportation. Below are some number to put this in context. These are crude measures, but they provide a sense of scale.

• 17.2 million new cars and light trucks were purchase in 2017. In the time that it takes Waymo to acquire these vehicles (~3 years) nearly 50 million new vehicles will have been purchased in America alone.
• Even more significantly, there are roughly 250 million registered vehicles in the country. Now, researchers have estimated that 1 AV could replace up to 12 manually driven vehicles, but that still means we would need over 20,000,000 AVs to replace all manually driven vehicles at a minimum.
  • Interestingly, if every vehicle sold in the US from today forward was a fleet-operated AV, we would have enough AVs to serve every trip in just 1 to 2 years. While 20,000 vehicles may not be much in the grand scheme, we have the manufacturing capacity today to make AVs a reality incredibly rapidly.
• Per USDOT, Americans take an estimated 1.1 billion trips every day. By this measure we would need roughly a 1000 fold increase in AVs to complete all the trips that occur in the country.
• In a 2013 case study, Columbia University’s Earth Institute found that a fleet of 18,000 AVs would be capable of serving 2/3 of the daily trips in Ann Arbor, MI. Ann Arbor has 285,000 residents, 200,000 personally owned vehicles and it experiences 740,000 trips on the average weekday. It is also the 228th largest city in America. In other words, the fleet that Waymo is acquiring from Jaguar would be capable of serving the bulk of travel that occurs in Ann Arbor.

Again, these are crude measures, but they put Waymo’s announcement in context. 20,000 AVs would, of course, make less of an impact in rural America than they would in New York City. But by every measure I list, we are several orders of magnitude away from significant impacts to transportation in America. That said, this transition has the potential to happen very quickly given the scale of existing automotive supply and manufacturing systems in this country.

As I’ve noted here, forecasts for the deployment of this technology range widely. Some have predicted that AVs will roll-out gradually over the next several decades. Others have predicted that AVs will largely replace manually driven vehicles by the end of the next decade. At this point, it is still too early to say which of these forecasts will be closest to reality. However, the first couple years of commercial service will give us some important clues to that future including the cost to consumers (which I’ll talk about in a subsequent post) and safety benefits.