Jiushi has declared a strategic enhancement of its 'city-level physical AI' and has positioned itself as the industry's inaugural L4 enterprise to set up a formidable 10,000-GPU computing power cluster. This state-of-the-art cluster underpins the entire R&D pipeline, encompassing model training, and propels the evolution of its proprietary multimodal foundational large model, APEX. The model's parameters have surged from tens of billions to hundreds of billions, empowering the Jiushi brain with pivotal capabilities like autonomous cognition.

Capitalizing on this robust computing infrastructure, Jiushi has devised a vehicle-cloud collaborative operational framework and a vehicle-cloud linkage (signifying seamless interaction between vehicles and the cloud) large model architecture. It has pioneered the Safety Officer Agent to ensure comprehensive scenario coverage. Through a bidirectional iterative process of 'data feedback-foundational model refinement-distillation and deployment', Jiushi has achieved twin breakthroughs in safety and traffic efficiency. This will fuel the iterative development of multimodal foundational large models and vehicle-cloud collaborative frameworks, paving the way for expansion into a broader spectrum of urban logistics scenarios.