Bringing autonomous driving into a new city is not only a perception challenge. It is a question of how quickly a system can understand a new environment, adapt to unfamiliar road structures, and move from first exposure to reliable autonomous performance. Using Munich as the central case study, this session will examine what it takes to accelerate L4 readiness in a new urban environment. It will introduce Air2Road, which combines real-world perception with aerial context to support a new kind of world model for autonomy. The presentation will focus on what was learned on real roads: how autonomy can operate without HD maps, minimize manual adaptation, and scale across new cities and ODDs.

• A New Perspective on Autonomy - Air2Road transforms vehicle perception into a real-time drone-like world model, providing a richer understanding of the driving environment than traditional ego-centric approaches.
• Faster Expansion to New ODDs - By leveraging a unified world representation instead of extensive HD mapping and manual tuning, Air2Road significantly reduces the effort required to deploy autonomy in new cities
• Vision-First, Cost-Effective Scaling - Air2Road enables Level 4 autonomy using a lean sensor configuration, making large-scale deployment economically viable for robotaxis and consumer vehicles.
• From Perception to World Models - The future of autonomous driving is not simply detecting objects but understanding relationships, intent, and future evolution of the scene through a dynamic world model.
• Proven in Munich, Built for Global Scale - Air2Road’s deployment in Munich demonstrates a practical path from pilot programs to scalable autonomous mobility, creating the foundation for rapid expansion across Europe and beyond.