NVIDIA Ties Hyperion to Bolt’s Madrid Robotaxi Data Plan
Bolt plans Madrid driving-data collection in 2027, subject to regulatory approval. Other European projects span passenger cars, freight and public transit.

NVIDIA is tying its Hyperion vehicle architecture to a widening set of European autonomous-driving projects, with Bolt planning Madrid data collection in 2027 for a Lucid-based robotaxi effort, subject to regulatory approval.
Bolt selected NVIDIA’s Alpamayo Stack for its European ride-hailing program, with software adjustments planned for local roads. Future Lucid vehicles built on Hyperion are the starting point, and the architecture can also run driving software created by partners.
The vehicle commitment remains a target: at least 25,000 future Lucid vehicles capable of Level 4 driving. Bolt has a separate objective of 100,000 autonomous vehicles by 2035 across Europe and the broader United Nations Economic Commission for Europe region. The new software choice and Madrid timeline add detail to the companies’ earlier partnership plan.
The partners expect Madrid to serve as the first real-world data-gathering phase in 2027, if regulators approve it. The initial work is meant to help develop and validate the system. Level 4 automation lets a vehicle handle driving within defined service areas without requiring a human driver to remain engaged. NVIDIA names Germany as a priority for future deployment, while Bolt also identifies France as a key market.
At GTC Berlin on Wednesday, Oct. 21, NVIDIA also described projects in other parts of the mobility sector. Mercedes-Benz is extending development of Level 3 and Level 4 driving capabilities. Volkswagen Commercial Vehicles presented an autonomous ID. Buzz AD platform using Hyperion, while Einride is developing a next-generation autonomous freight system with the architecture. TIER IV and the Technical University of Munich plan to use Hyperion for autonomous public-transportation work.
Hyperion provides a shared computing and sensor foundation intended to ease integration and validation across different vehicle programs. NVIDIA introduced Halos SAFE, a safety-evaluation framework that connects simulation with software and hardware testing, road trials and fleet monitoring. Its Alpamayo 2 Edge model has two billion parameters and is designed to interpret in-vehicle scenes and plan vehicle movement at 10 Hz.
“Robotaxis are the first mass-market application of robotics and physical AI,” NVIDIA founder and CEO Jensen Huang said.
The projects span different stages: Mercedes-Benz is extending driving-capability development, Volkswagen presented a platform, and the transit partners are planning development work. Bolt’s next stated milestone is the Madrid data-collection phase in 2027, subject to regulatory approval.