Wayve and Waabi are shifting the self-driving paradigm from map-centric, component-based engineering to end-to-end AI powered by world models—dynamic, physics-aware simulators that enable scalable, safe testing and learning. Wayve’s GAIA models and Waabi’s closed-loop simulator prioritize generalization, sensor flexibility, and real-world deployment across consumer vehicles and robotaxis, backed by partnerships with Uber, Nissan, and Volvo. Both companies emphasize capital efficiency and strategic licensing over vertical integration, targeting per-mile or subscription-based revenue models rather than hardware sales. While Wayve focuses on global scalability starting from the hardest domain—urban driving—Waabi prioritizes robust physical AI for trucks first, with plans to expand into robotaxis using a unified 'AI brain.' Regulatory progress outside the US and China, OEM hardware readiness, and advances in generative AI have converged to make 2026 a realistic inflection point for commercial deployment. Crucially, both firms reject premature claims of 'solved' autonomy, stressing that engineering rigor, safety validation, and real-world integration—not just algorithmic breakthroughs—remain decisive.