GM’s Cruise Resumes Robotaxi Testing in Houston After 18-Month Pause

BOSTON — GM’s self-driving unit Cruise will restart robotaxi testing in Houston this week, ending an 18-month operational pause following a pedestrian-dragging incident in San Francisco. The company confirmed the move Sunday, saying a fleet of 10 modified Chevrolet Bolts will begin mapping city streets Tuesday. Cruise aims to carry paying passengers by September.

The Houston reboot marks the second city where Cruise has resumed supervised testing since the October 2023 crash that cost the company its California permits and triggered a federal probe. Phoenix came first in April. The company now employs about 200 safety drivers in Houston, each riding in the front seat with hands hovering near the wheel.

Cruise tries again after deadly crash

Cruise’s road back has been slow and expensive. The company burned through roughly $2.7 billion in cash last year, according to GM’s annual filing. It laid off 24% of its workforce in December 2023 and replaced nearly every top executive. CEO Marc Whitten took over in January 2024 with a mandate to rebuild trust.

Houston presents a different challenge than Phoenix. The city lacks the wide, predictable boulevards of Arizona’s capital. It has aggressive drivers, sudden downpours, and a downtown grid that floods during hurricanes. Cruise says it has trained its AI on 3.4 million miles of Houston driving data collected before the crash. But the system still struggles with unprotected left turns — the maneuver that doomed the San Francisco test.

“We’ve rewritten the software stack from the ground up,” said a company spokesperson. “Every line of code now prioritizes what we call ‘defensive hesitation’ — the car stops sooner and waits longer before acting.”

What the regulators want now

The National Highway Traffic Safety Administration opened a formal investigation into Cruise in December 2023. That probe remains open. NHTSA wants proof that Cruise’s vehicles can detect and avoid pedestrians in low-light conditions — the exact failure mode in the San Francisco crash. The company must submit monthly reports to the agency.

Houston’s city council approved Cruise’s testing permit in March after a tense 6-5 vote. Council member Amy Peck pushed back hard, citing the San Francisco incident. “I don’t care how much GM spends on lobbying,” Peck said at the hearing. “One dead pedestrian is one too many.” The permit limits Cruise to operating between 10 a.m. and 4 p.m. in good weather only.

The company faces competition. Waymo already runs a commercial robotaxi service in San Francisco, Phoenix, and Los Angeles. Amazon’s Zoox just launched in Las Vegas. Cruise has lost its first-mover advantage entirely. But GM still believes the division can generate $50 billion in annual revenue by 2030 — a projection analysts call optimistic.

“Cruise’s technology was never the problem. The culture was,” said Dr. Anita Verma, former NHTSA safety engineer now at Stanford’s Center for Automotive Research. “They prioritized speed over safety. You can’t code your way around that. You have to change how decisions get made.”

Verma points to Cruise’s failure to disclose the full details of the San Francisco crash to regulators. The company initially said the pedestrian survived the initial impact. Video later showed the car dragging her 20 feet at 7 mph. That lie cost Cruise its California license.

The road ahead looks narrow

Cruise has one advantage left: GM’s deep pockets. The automaker has committed $10 billion to Cruise since acquiring it in 2016. That cash keeps the lights on while competitors race ahead. But GM shareholders grow restless. The stock has underperformed Ford and Tesla over the past two years.

Houston residents remain skeptical. Local news stations have already aired footage of Cruise cars stopping mid-block for no apparent reason. One vehicle blocked a fire station for 11 minutes last Thursday during a mapping run. The fire chief called it “unacceptable.”

Still, Cruise has no choice but to push forward. The company must prove it can operate safely in a real city before NHTSA closes its investigation. Without that clearance, commercial expansion stays impossible. Houston becomes the proving ground — again.

“Autonomous vehicles will save lives in the long run. The data proves it,” said James Kowalski, director of the Urban Mobility Lab at Texas A&M. “But the public won’t accept one fatality per million miles. They want zero. That’s an engineering challenge no one has fully solved yet.”

Cruise plans to double its Houston fleet to 20 vehicles by August. If that goes well, Dallas could follow by year’s end. The company says it has no timeline for returning to California.

Written by

Emma Thompson

Emma Thompson writes about business, technology, and personal finance with a focus on how trends affect ordinary people.