Momenta Robotaxi Obtains Shenzhen Road Test Permit, Accelerates Self-Driving...

[CNMO Tech News] July 21, autonomous driving company Momenta announced that its Robotaxi has officially obtained the Shenzhen Intelligent Connected Vehicle Road Test Permit, and will commence on-site testing in Shenzhen shortly. This is another key testing qualification obtained by Momenta in China's first-tier cities after Shanghai, Suzhou, and Wuxi.

Momenta

Currently, Momenta Robotaxi has successively implemented testing in cities such as Shanghai, Suzhou, Wuxi, Shenzhen in China, as well as Munich, Germany, and Abu Dhabi, UAE. The company plans to implement high-level autonomous driving in multiple new cities and regions globally within 2026, steadily promoting technical scenario verification and people's livelihood travel service exploration. Its mass production solution has been implemented in over 10 countries and regions in Asia, Europe, Oceania, Latin America, and North Africa.

In terms of the cooperation ecosystem, Momenta Robotaxi has established cooperation with global ride-hailing platforms such as Uber, Grab, Lumo, Xiangdao Travel, as well as automakers like Mercedes-Benz. Previously, representatives from relevant Singapore government departments, Lyft, and other government/enterprise sectors visited Momenta for exchanges, demonstrating the international industry's attention on the progress of China's autonomous driving technology.

Momenta Partners

Momenta released the R7 Large Model based on World Models and Reinforcement Learning in April 2026. This Physical AI Foundation Model will support Momenta in the scale implementation of fields such as passenger cars, Robotaxi, and Robovan (autonomous logistics vehicles).

The core logic of the R7 World Model lies in enabling vehicles to build cognition of real-world physical rules. Distinct from traditional solutions relying on manual rules and image recognition, this system relies on massive driving data to learn fundamental physical laws such as inertia, friction, and object motion trajectories. It can independently deduce various rare long-tail road conditions and handle scenarios such as sudden falling objects, mixed traffic at intersections, and non-motorized vehicle weaving, which are difficult for traditional algorithms to cover. Its technical architecture is divided into three layers: World Model Pre-training Layer compresses physical laws, World Model Simulation Layer deduces long-tail scenarios, Reinforcement Learning Layer achieves autonomous decision-making optimization.

Momenta Collaborative Models

Momenta Vice President Gu Gongyao defined the R7 World Model as "The ChatGPT Moment for Physical AI", believing this is the key lever for AI moving from the digital world to the physical world. In terms of mass production implementation, Cadillac has announced it will be the industry's first to mass produce models equipped with the Momenta R7 World Model. SAIC Volkswagen ID.ERA 9X will also be among the first to be equipped with the R7 World Model. Roewe Jiayue 07 will also be equipped with the latest generation Momenta R7 World Model advanced intelligent assisted driving solution.