Almost two years ago, Motional stood at a crossroads of autonomous vehicles.
The company, formed from a $4 billion joint venture between Hyundai Motor Group and Aptiv, had already missed a deadline to launch a driverless robotaxi service with partner Lyft. The company had lost Aptiv as one of its backers, prompting Hyundai to invest another billion dollars to keep the company going. Several layoffs, including a 40% restructuring cut in May 2024, had reduced the company from a peak of about 1,400 employees to fewer than 600. Meanwhile, advances in AI were changing the way engineers developed the technology.
Motional should evolve or die. It paused everything and chose option #1.
Motional told TechCrunch that it has restarted its robotaxi plans with an AI-first approach to its self-driving system and a commitment to launch a commercial, unmanned service in Las Vegas by the end of 2026. The company has already opened up a robotaxi service – with a human safety operator behind the wheel – to its employees. It plans to offer this service to the public later this year with an undisclosed ride-hailing partner. (Motional has existing relationships with Lyft and Uber.) By the end of the year, the human safety operator will be removed from the robotaxis and a true commercial, driverless service will begin, the company said.
“We saw that there was tremendous potential with all the advancements happening in AI; and we also saw that while we had a secure driverless system, there was a gap in finding an affordable solution that could generalize and scale globally,” said Laura Major, president and CEO of Motional, during a presentation at the company’s facilities in Las Vegas. “And so we made the very difficult decision to stop our commercial activities and slow down in the short term so that we could accelerate.”
Image credits:Kirsten Korosec
This meant a shift from the classic robotics approach to an AI-based model-based approach. Motional was never devoid of AI. Motional’s self-driving system used individual machine learning models to process perception, tracking and semantic reasoning. But it also used more rule-based programs for other operations within the software stack. And the individual ML models turned it into a complex web of software, Major said.
Meanwhile, AI models originally built for language began to be applied to robots and other physical AI systems, including the development of autonomous driving. That transformer architecture made it possible to build large and complex AI models, ultimately leading to the rise and skyrocketing use of ChatGPT.
Motional looked for ways to combine and integrate these smaller models into a single backbone, enabling an end-to-end architecture. It has also serviced the smaller models for developers, which Major explained gives Motional the best of both worlds.
“This is really critical for two things; one is to generalize more easily to new cities, new environments and new scenarios,” she said. “And the other is to do this in a cost-optimized way. For example, the traffic lights may be different in the next city you go to, but you don’t have to redevelop or reanalyze it. You just collect some data, train the model, and it can function safely in that new city.”
TechCrunch was able to see Motional’s new approach firsthand during a 30-minute autonomous drive through Las Vegas. One demo cannot provide an accurate assessment of a self-driving system. However, it can reveal weaknesses and differences from previous iterations and measure progress.
Progress is what I saw as the Hyundai Ioniq 5 I was riding in autonomously navigated its way from Las Vegas Boulevard to the Aria Hotel pickup and drop-off area. These bustling areas are notorious in La Vegas and my experience was no different as the autonomous vehicle slowly weaved its way around a stopped taxi unloading passengers, changed lanes and then back again, passing dozens of people, giant flower pots and cars along the way.
Motional previously operated a ride-hailing service in Las Vegas with partner Lyft with vehicles that would handle parts of a ride autonomously. Parking lots and hotel valet and app ride pickup locations were never part of these activities. A human security guard, always behind the wheel, took over to navigate parking lots or the busy pick-up and drop-off points of hotel lobbies.
There is still more progress to be made. The graphics shown to riders in the vehicle are still in development. And while there was never any retreat during my demo ride—meaning the human safety officer takes over—the vehicle did take its time to maneuver around a double-parked Amazon delivery van.
Still, Major maintains that Motional is on the right track to deploy safely and cost-effectively. And majority shareholder Hyundai is in it for the long haul, she said.
“I think the real long-term vision, you know, for all of this is Level 4 for people’s personal cars,” Major said, referring to a term that means the system handles all the driving without the expectation of human intervention. “Robotaxis, that’s stop number one, and it has a huge impact. But ultimately I think every OEM would like to integrate that into their cars as well.”
