Last October, Waymo did something remarkable: the company launched a fully driverless commercial taxi service called Waymo One. Customers in a 50-square-mile corner of suburban Phoenix can now use their smartphones to hail a Chrysler Pacifica minivan with no one in the driver’s seat.

And then… nothing. Seven months later, Waymo has neither expanded the footprint of the Phoenix service nor has it announced a timeline for launching in a second city.

It’s as if Steve Jobs had unveiled the iPhone, shipped a few thousand phones to an Apple Store in Phoenix, and then didn’t ship any more for months—and wouldn’t explain why.

Last Friday, two Waymo employees participated in an “ask me anything” thread on the SelfDrivingCars subreddit, a watering hole for self-driving industry insiders. Questions about expansion plans dominated the conversation.

“How are you going to scale?” one redditor asked. “What are the impediments to service expansion at this time?”

The Waymonauts responded with maddening generalities.

“We feel the same urgency to scale quickly that others do, but a ton of work goes into doing it safely,” wrote Waymo’s Sam Kansara.

Obviously, any expansion must be done safely. But the pair’s answers didn’t really add up. If Waymo is confident its technology is safe to operate in Chandler, Arizona, why wouldn’t Waymo self-driving be equally safe in neighboring suburbs like Gilbert or Mesa? Those cities have very similar weather, road conditions, and traffic patterns.

Maybe Waymo One just isn’t profitable?

Until recently, the best explanation I could come up with was about the cost of running Waymo’s service: maybe the cost of the service is just too high for Waymo to turn a profit.

One reason for this would be if Waymo’s self-driving technology isn’t actually that automated. If a driverless vehicle gets confused, the safe thing to do is often to slow down and contact Waymo headquarters for assistance. Waymo has a staff of fleet operators that oversees the cars and provides guidance—such as confirming that the car has chosen a safe route through a construction zone (Waymo says that its remote operators never drive the cars directly).

If this happens often, then Waymo might need a large operational staff, which could drive up costs. However, I doubt that this is the main issue. Waymo’s Lety Cavalcante wrote last Friday that “each Ops team member supports multiple vehicles.”

It’s also possible that more mundane factors could be driving up costs. For example, with no drivers to clean up messes, Waymo needs a cleaning staff in its garages. Waymo also has a call center to respond when a customer pushes a button inside the car to call for help.

A less obvious but potentially significant expense was pointed out to me by Joel Johnson, a Phoenix-area college student who has taken dozens of Waymo rides and documented them on YouTube. Johnson has observed that Waymo’s driverless Pacificas are frequently tailed by a Waymo roadside-assistance van. If a Pacifica gets stuck, a van can quickly arrive at the scene to fix the problem.

Johnson believes that in the early months of driverless operation, Waymo had a roadside-assistance van following every Waymo One vehicle, which would be a significant expense.

Hardware costs could also be a factor. The gear for early Waymo prototypes reportedly cost hundreds of thousands of dollars. With equipment that expensive, depreciation alone could eat up a significant fraction of the fares paid by Waymo customers.

Add all this up and it’s easy to imagine that Waymo’s cost to complete each ride might exceed the revenue it earns. In that case, scaling up the service might just lead to bigger losses rather than putting Waymo on the path to profitability.

Cost alone can’t explain it

Still, I don’t think these cost considerations can fully explain Waymo’s lack of expansion. Alphabet has already spent hundreds of millions—probably billions—of dollars developing and testing Waymo’s self-driving technology. The company presumably wants to turn that investment into a market-leading position in the emerging self-driving car industry. Right now, Waymo seems to only have a few dozen vehicles in its driverless commercial fleet in the Phoenix area. Whatever it costs to operate those vehicles, it must be a tiny fraction of Waymo’s spending to date.

More importantly, the only way for Waymo to drive down costs is by gaining experience and scale. This is most obviously true in hardware. Waymo’s first prototypes cost stacks of money precisely because Waymo was building them one at a time. For example, it had to pay around $75,000 for exotic lidar sensors. Over time, Waymo has developed more cost-effective lidar designs. It will see further cost declines when it can start putting in large-volume orders.

Similar points apply to other costs of the Waymo One service. For example, if the company only has a few dozen cars, it might be necessary to have a roadside-assistance van following each one so that one can arrive promptly if there’s a problem. If Waymo had hundreds of cars in the same area, on the other hand, it could assign assistance vans to zones rather than individual vehicles, enabling each assistance van to provide backup to several Waymo One vehicles.

And, of course, the long-term solution is for Waymo to improve its software so that assistance vans are needed less and less frequently. The best way to do that is by racking up more miles on the road. It seems better to generate some revenue by carrying a passenger rather than generate no revenue with a test vehicle.

Another theory I’m pretty sure is wrong is that it’s too difficult or expensive for Waymo to generate high-density maps of new areas. Waymo’s cars require extremely detailed 3-D maps, with objects like curbs, signs, and streetlights marked.

I’m sure building these maps takes time and money, but it can’t possibly take that much money. Waymo has racked up millions of miles test-driving in the Phoenix area—enough to traverse every street in the area many times over. The cost of paying someone to drive a mapping car on every street in a new service area must be comparatively tiny.

Suburban Phoenix is a terrible place to run a taxi service

Things start to make more sense if you think about Waymo One from the demand side. In other words: what are people supposed to use Waymo’s taxi service for?

Waymo One is centered in Chandler, Arizona, which is extremely suburban. Most households there own a car. Most stores, offices, and homes have ample parking. As a result, most people in Waymo’s service area don’t need a taxi very often.

An infrequently used ride-hail service is likely to be unprofitable. A ride-hail network needs a minimum density of vehicles on the road to give customers a reasonable pickup time.

Waymo says that, right now, wait times are its biggest source of customer complaints. To address that, Waymo would have to expand its fleet so that there’s always a car fairly close to any given customer. But it can’t do that profitably unless there are enough customers to keep the vehicles busy.

How might Waymo increase usage?

One of the most common reasons people in the suburbs take taxis is to go to the airport. Phoenix’s main airport is about 10 miles northwest of Waymo’s current service territory.

https://arstechnica.com/?p=1762459