A research report from CITIC Construction Investment highlights how Meta Muse and the Yunqi Conference underscore the impact of AI application deployment on computing power requirements. Meta's personal AI assistant, Muse, facilitates email management, travel arrangements, and transaction processing. Within just 12 days of its release, it garnered 2.8 million downloads. Moreover, Muse is integrating with Shopify's payment system to expedite the commercialization of consumer-facing AI agents. At the Yunqi Conference, Alibaba announced that the Qwen 4 model is currently in training, with plans to scale up its parameters to a range of 5 trillion to 10 trillion in the future. The company also introduced the Zhenwu V900 chip, set an ambitious target of surpassing 20GW in global data center capacity by 2032, and advanced enterprise agent applications through Qianwen Office. As AI evolves from conversational interactions to task execution, the demand for inference capabilities, tool invocation, and cloud resources is set to surge. Therefore, it is prudent to keep a close watch on related sectors, including high-speed optical modules, switches, servers, liquid cooling solutions, and power supplies.