Core Viewpoint - Shanghai Telecom is leveraging Huawei's AI WAN solution to create a new service model for computing power rental, addressing the needs of various industries while ensuring data security and efficient resource allocation [1][2]. Group 1: Infrastructure Development - Shanghai Telecom is developing a 400GE IP metropolitan intelligent computing POD to connect enterprise parks with operator computing centers, providing integrated computing network services [2]. - The new infrastructure aims to facilitate the efficient flow of computing resources between operators and enterprises, promoting the widespread adoption of AI technology [1]. Group 2: Business Model and Demand - Enterprises, particularly state-owned and government sectors, are increasingly opting to rent computing resources from operators to meet their training and inference needs while keeping sensitive data within their premises [1]. - A growing trend among enterprises is to adopt a light-asset operation model, combining self-built computing servers with rented resources from operator intelligent computing centers to reduce procurement and operational costs [1]. Group 3: Data Security and Flexibility - The distributed training and inference model allows enterprises to deploy sensitive data models locally while utilizing operator resources for other parts, ensuring data security during transmission [2]. - The use of RDMA lossless transmission technology enables efficient inference over distances of up to 240 kilometers without performance degradation [2]. Group 4: Collaboration and Future Goals - Shanghai Telecom is collaborating with Huawei and various enterprises to tackle key technological challenges in intelligent computing networks and validate solutions in vertical industry scenarios [2]. - The ultimate goal is to achieve commercial implementation of integrated computing network services, contributing to the high-quality development of computing infrastructure in China [2].