GM faces hurdles with launch of robotaxi fleet

Editor’s note: An earlier version of this story included an employment goal for GM regarding autonomous vehicles that was misidentified as an objective for Cruise only. The reference has been deleted.

Halfway through 2019, General Motors is running short on time to meet its goal of launching a large-scale fleet of robotaxis by the end of the year.

Dan Ammann, the company's former president who now leads its autonomous- vehicle unit in San Francisco, has GM Cruise cranking away to meet the ambitious plans, but it may be too little too late to overcome unexpected government hurdles and reported technical issues to launch the service on time, at least how the automaker initially planned. GM has yet to submit an application to the State of California to get approval for the service, a process that has never been done and could take weeks, if not months.

"We're working aggressively," GM CEO Mary Barra said last month when asked about Cruise. "Our rate of iteration continues to improve. So, that is the position we're in and that's the approach that we have, and we're very much looking forward to rolling out this technology because we do believe it will save lives."

The vision GM laid out in November 2017 called for launching a public ride-hailing service using self-driving vehicles in "dense urban environments" and with no backup driver aboard.

But given the obstacles that still remain, it's more likely any service GM starts this year would be much more limited in scope, perhaps only in San Francisco with human safety drivers behind the wheel.

It would still be a major milestone but not the aggressively timed breakthrough GM promised.

"Everybody in the industry underestimated how hard a problem this was going to be, and I think GM probably was a little overambitious," said Sam Abuelsamid, a principal research analyst at Navigant Research. "It's a mix of … a harder problem to solve and they're just not making the progress they had hoped for."

Whether GM accomplishes the goal is arguably irrelevant in the automaker's long-term plans to capitalize on what it believes to be a trillion-dollar industry, but it's the latest evidence of the complexity and difficulty in revolutionizing an industry and taking the driver out of the driving.

GM executives have been steadfast in saying safety would be its deciding factor regarding the launch of the fleet, but none has given a meaningful update on the plans in some time.

Cruise spokesman Ray Wert last week declined to comment directly on the company's progress, saying only that the company would be guided by safety.

If GM does not launch the ride-hailing service at scale this year, it will add to a growing list of missed goals for Cruise that also includes testing in New York City and increasing vehicle testing to a pace of 1 million miles a month by early 2018.

Abuelsamid says that, although missing goals can reflect negatively on Cruise, waiting to ensure the vehicles are safe remains the best option.

"From a business perspective, it hurts because that puts off the time when you start to generate revenues," he said. "But from a public standpoint, from a safety standpoint, it's better to wait and get it right than rush it out there and have technology that isn't ready yet — especially in an urban environment."

Waymo, Google's self- driving vehicle unit, is the most prominent company currently offering a limited autonomous ride-hailing service to the public. Waymo One, which started in the Phoenix area in December, uses human safety drivers and offered 400 people who participated in a pilot program first crack at using the service.

Abuelsamid said he thinks GM might follow a cadence similar to Waymo's in launching its fleet.

Since GM detailed its plans to deploy the autonomous ride-hailing fleet, the company has petitioned the U.S. government to let it use vehicles without manual controls such as a steering wheel and pedals.

Meanwhile, several reports have surfaced that GM's vehicles have had unexpected problems and software glitches, including shutdowns of the autonomous control system. In October, Reuters reported that Cruise vehicles were struggling to identify pedestrians and determine whether objects on the road were moving or stationary, among other issues.

GM's petition to launch a ride-hailing fleet without manual controls sat stagnant for more than a year until March, when NHTSA said it would solicit public comments and conduct a review. That process concluded in May.

NHTSA, in an emailed statement last week, said it is assessing the merits of the petition and considering the comments submitted.

If granted, the waiver would allow GM to launch as many as 2,500 vehicles a year as driverless taxis that could be summoned via a smartphone app and commanded using touch screens inside the vehicles.

GM is testing self-driving cars with manual controls and safety drivers in Arizona, California and Michigan, including as part of a pilot ride-hailing program for employees in San Francisco.

Cruise founder Kyle Vogt, who is now the unit's president and chief technology officer, has previously said the third-generation vehicles being produced now are the first that meet the redundancy and safety requirements to operate without a driver.

The company, if it wanted, could launch these vehicles without an exemption from NHTSA.

Cruise also faces hurdles from state and local regulations. San Francisco, where Cruise is expected to launch its robotaxis, has been placing stricter limitations on ride-hailing services.