Cruise’s relaunch comes at a pivotal moment for the U.S. robotaxi industry. Elon Musk will showcase

Tesla
’s

“purpose-built robotaxi” in August.

Alphabet

-owned Waymo is expanding its Northern California territory beyond San Francisco into the Bay Area suburbs. And newcomer Zoox, owned by

Amazon
,

just announced plans to test in Austin and Miami.

The safety experts Barron’s spoke with expressed skepticism that driverless technology has improved much since the gruesome accident last fall that led to Cruise halting driverless operations nationwide after California suspended its permits. “There’s been no quantum leap in technology,” said Phil Koopman, a Carnegie Mellon University autonomous vehicle safety expert who testified at a congressional hearing last summer. 

There’s also been pushback from residents, labor groups, and politicians. A California state Senate bill introduced by Democrat David Cortese currently working its way through the state legislature would enable cities to set vehicle caps and limit robotaxis’ hours of operation, effectively slowing the rollout.

While driverless taxi companies say their vehicles are as safe, if not safer, than those driven by humans, investigations announced by federal regulators in recent weeks underscore broad concerns that the autonomous driving systems at the heart of robotaxis still aren’t ready for prime time.

Waiting on the robotaxi revolution

It might be hard to imagine a future where most cars are whizzing around roads without a driver, but that’s exactly what companies like Tesla and many industry watchers envision. 

Tesla alone could generate $100 billion in annual revenue on robotaxis in the U.S. by 2040, RBC Capital Markets analyst Tom Narayan estimates, with the entire U.S. market bringing in $500 billion. That’s a tall order, considering that Tesla doesn’t even have a robotaxi on the roads today.

General Motors

—the only U.S. company that publicly discloses its driverless car revenue—reported a nearly $3.5 billion loss from Cruise in 2023. Alphabet doesn’t break out Waymo’s finances, but Waymo said in May that it now completes 50,000 rides a week across San Francisco, Los Angeles, and Phoenix.

Getting regulatory approval is the biggest hurdle to broad robotaxi rollouts. What’s needed, Musk has said, is “conclusive data that the autonomous car is safer than a human driven car.”

Never again?

Cruise says the horrific dragging incident that led to its permits getting pulled in California could never happen again. “We made changes to our software to ensure that if the same incident occurred we would stay stopped,” Steve Kenner, Cruise’s new chief safety officer, told Barron’s in an interview. That’s a result of both improved artificial intelligence and new algorithms written by engineers to prevent it.

“We’re always improving,” Kenner added. “The bar we’re shooting for is to be a role model driver not just an average ride hail driver.”

Regulators are watching closely. In May, the National Highway Traffic Safety Administration launched an investigation into more than 30 incidents involving vehicles from rival driverless taxi firm Waymo, ranging from driving into opposing traffic on the wrong side of the road to crashing into parked vehicles. Amazon’s start-up Zoox is facing a similar inquiry.

Missing safety data

Cruise and Waymo both say their autonomous vehicles are safer than human drivers. But several embarrassing glitches and mishaps caught on video lay bare the technology’s shortcomings, namely “their inability to reason under uncertainty,” says Missy Cummings, director of Mason Autonomy and Robotics Center at George Mason University. 

“Several of the incidents involved collisions with clearly visible objects that a competent driver would be expected to avoid,” the NHTSA wrote in a May 23 letter to Waymo explaining why it is expanding its investigation.

Research conducted by robotaxi companies points out the limitations of existing data, namely that there simply isn’t enough of it to reach statistical significance for severe injuries and fatalities.

The problem, says Carnegie Mellon’s Koopman, is that autonomous vehicles haven’t driven enough miles to gather enough data to show how many severe injuries and fatalities there will be. “In the U.S. human drivers might see one fatality per 100 million vehicle miles. Companies who have a small fraction of that number of miles—7.1 million in the latest Waymo study—are nowhere near knowing how fatalities will turn out.”

That’s not the only problem. A 2023 study claiming Waymo robotaxis had “zero bodily-injury claims over 3.8 million miles” only included at-fault crashes where both an insurance claim and a payment were made. 

Waymo has made strides in the past year to address “a very long tail of scenarios that we still need to work on,” Chris Ludwick, the company’s director of product management, told Barron’s. For example, its cars now play audio clips at incident scenes where first responders are present to let them know the vehicle is planning to back up and exit the scene.

The (bumpy) road ahead

Going forward, Waymo says the next frontier is freeway driving. While it has permits to test on them, speeds are currently capped at 65 mph. It’s also ramping up in Austin, Texas, where it plans to offer commercial rides later this year. 

Cruise, meanwhile, resumed testing in Dallas in early June with human drivers. It’s looking to restart commercial rides “at the end of this year,” Kenner said, but is in no rush. The reboot “will be gated by safety and ensuring that we complete all the safety deliverables,” he noted.

The biggest wild card is Tesla. While Musk has given scant details on the robotaxi unveiling later this summer, he’s betting the company on autonomy. “If somebody doesn’t believe Tesla’s going to solve autonomy, I think they should not be an investor in the company,” he said in an April call with investors.

Meanwhile, the pushback against robotaxis shows no signs of stopping. “We just can’t let tech run rampant,” David Canepa, a San Mateo, Calif., county supervisor who supports increased regulation of driverless taxis, told Barron’s. “I’m afraid of the wild wild west of autonomous vehicles being deployed.”

That tension between those who think driverless cars are moving too fast and those who don’t is likely to play out for years to come.

Write to Anita Hamilton at anita.hamilton@barrons.com