General Motors is taking a major step forward in autonomous driving development, launching supervised public-road testing of its next-generation automated technology across California and Michigan. The move signals a transition from large-scale data collection to active, real-world validation of advanced driver systems.
The new technology—trained on millions of real-world miles and extensively validated through simulated scenarios—will now be tested on limited-access highways. More than 200 development vehicles are expected to operate in live traffic environments in the near future, each equipped with a trained safety driver ready to take control at any moment.
From Data Collection to Real-World Validation
GM’s automated driving program has already accumulated over one million miles of data collection across 34 U.S. states. That extensive dataset now underpins the next phase of development, as the automaker refines its systems through supervised testing in real traffic conditions.
This milestone reflects GM’s structured, incremental approach to scaling autonomous technology for personal vehicles—prioritizing safety, validation, and continuous improvement.
Eyes-Off Driving Set for 2028 Debut
GM previously confirmed plans to introduce its “eyes-off” driving system starting in 2028, debuting in the Cadillac Escalade IQ. The system is designed to operate without requiring constant driver supervision in specific conditions, marking a significant leap beyond today’s hands-free technologies.
The rollout will begin on highways before expanding to more complex, door-to-door driving scenarios. GM aims to deploy this capability across a broad range of vehicles, from premium Cadillac models to mainstream Chevrolet offerings.
Powered by a New Centralized Architecture
A key enabler of GM’s autonomous ambitions is its next-generation centralized computing platform. This architecture consolidates vehicle intelligence—previously distributed across dozens of modules—into a unified system, allowing for faster updates and scalability across multiple vehicle segments.
By leveraging this approach, GM can deploy advanced automated features across its lineup without redesigning systems for each individual model.
Built on Billions of Miles of Experience
GM’s autonomous ecosystem benefits from extensive real-world and simulated driving data. The company’s Super Cruise has logged more than 800 million customer-driven miles across 23 vehicle models. Meanwhile, its subsidiary Cruise has accumulated over 5 million fully autonomous miles in complex urban environments.
This combination of supervised and fully autonomous experience provides a robust foundation for the next-generation system.
Simulation at Unmatched Scale
In addition to on-road testing, GM leverages advanced simulation tools capable of replicating approximately 100 years of human driving every single day. These virtual environments allow engineers to test edge cases, refine AI models, and accelerate development cycles far beyond what real-world testing alone could achieve.
Advancing Toward Scalable Autonomy
The latest testing phase underscores GM’s commitment to delivering safe, scalable autonomous driving solutions. By combining real-world data, simulation, and a new centralized architecture, the company is positioning itself to bring “eyes-off” driving to market within the next few years.
