Motional places artificial intelligence at the forefront of its robotaxi relaunch, aiming to introduce a fully autonomous service by 2026

Motional’s Pivotal Shift Toward AI-Driven Autonomy

Almost two years ago, Motional found itself at a critical juncture in the autonomous vehicle industry.

Formed through a $4 billion collaboration between Hyundai Motor Group and Aptiv, Motional had already missed its target to launch a driverless robotaxi service with Lyft. The company faced a major setback when Aptiv withdrew its financial support, leading Hyundai to inject nearly $1 billion to keep operations afloat. A series of workforce reductions—including a 40% staff cut in May 2024—reduced Motional’s team from a peak of 1,400 to fewer than 600 employees. Meanwhile, rapid advancements in artificial intelligence were reshaping the landscape for self-driving technology development.

Faced with these challenges, Motional chose to reinvent itself rather than fade away, pausing its previous plans to pursue a new direction.

The company shared with TechCrunch that it has relaunched its robotaxi initiative with a renewed focus on AI at the core of its autonomous driving system. Motional aims to introduce a fully driverless commercial service in Las Vegas by the end of 2026. Currently, employees can access a robotaxi service with a human safety operator present, and the company intends to open this service to the public—partnering with a yet-to-be-named ride-hailing company—later this year. By year’s end, Motional expects to remove the safety operator entirely, marking the start of a true commercial driverless operation. (Motional has previously worked with both Lyft and Uber.)

“We recognized the immense promise in the latest AI breakthroughs. Although we had developed a safe driverless platform, there remained a gap in achieving a scalable and cost-effective solution for global deployment,” explained Laura Major, Motional’s president and CEO, during a presentation at the company’s Las Vegas site. “That’s why we made the tough call to temporarily halt our commercial rollout—slowing down now to accelerate progress in the future.”

Transitioning from Traditional Robotics to AI Foundations

This strategic pivot involved moving away from a conventional robotics framework to one built on foundational AI models. While Motional’s earlier systems already incorporated AI—using specialized machine learning models for perception, tracking, and semantic understanding—many other functions still relied on rule-based programming. The result was a complex, fragmented software architecture, according to Major.

As AI models originally designed for language processing began to find applications in robotics and autonomous vehicles, the transformer architecture enabled the creation of large, sophisticated AI systems. This evolution paved the way for tools like ChatGPT and accelerated the integration of advanced AI into self-driving technology.

Building a Unified AI Backbone

Motional explored ways to merge its smaller, specialized models into a unified, end-to-end AI backbone. At the same time, it retained these modular models for developers, allowing the company to benefit from both flexibility and scalability.

“This dual approach is crucial for two reasons: it makes it easier to adapt to new cities and scenarios, and it keeps costs in check,” Major noted. “For instance, if traffic signals differ in a new city, there’s no need to redesign the system—just gather data, retrain the model, and it can safely operate in the new environment.”

Firsthand Experience with Motional’s Evolving Technology

TechCrunch had the opportunity to experience Motional’s new system during a 30-minute autonomous ride through Las Vegas. While a single demonstration can’t fully evaluate the technology, it does highlight areas of progress and remaining challenges.

During the ride in a Hyundai Ioniq 5, the vehicle autonomously navigated the busy pickup and drop-off zone at the Aria Hotel, skillfully maneuvering around obstacles such as stopped taxis, unloading passengers, and crowded sidewalks. This level of navigation marks a significant step forward, as previous Motional services with Lyft did not include complex environments like hotel valet areas or parking lots—those were always handled by a human safety operator.

There are still areas for improvement. The in-car graphics for passengers are still being developed, and although the vehicle never required the safety operator to intervene during the demo, it did cautiously edge around a double-parked Amazon delivery van.

Despite these ongoing refinements, Major is confident that Motional is on the right trajectory for safe and cost-effective deployment. She also emphasized Hyundai’s long-term commitment to the project.

“Ultimately, the big vision is to bring Level 4 autonomy—where the system handles all driving without human input—to personal vehicles,” Major said. “Robotaxis are just the first step, but the goal is for automakers to eventually integrate this technology into everyday cars.”