According to Zhitong Finance APP, China Securities Co., Ltd. (ASX: 601066) released a research report stating that Meta Muse and the Yunqi Conference together reinforce the logic that AI application deployment drives computing power demand. Meta launched its personal AI assistant Muse, which supports email, travel booking, and transaction execution. It reached 2.8 million downloads within 12 days of launch and is advancing integration with Shopify's payment ecosystem, promoting the commercialization of consumer-grade agents. At the Yunqi Conference, Alibaba disclosed that Qwen 4 is being trained, released the Zhenwu V900 chip, proposed a goal of global data center capacity exceeding 20GW by 2032, and promoted enterprise agent deployment through Qianwen Office. The bank believes that consumer-side agents expanding use scenarios and enterprise-side agents deepening process participation are expected to increase demand for multi-turn reasoning, tool invocation, and continuous operation, driving consumption of computing, storage, and network resources. Attention should be paid to high-speed optical modules, switches, servers, liquid cooling, and power supply segments.
Key viewpoints of China Securities Co., Ltd. are as follows:
Meta Muse drives AI upgrading from conversational interaction to task execution, marking new progress in consumer-grade agent commercialization. On September 8, Meta launched its personal AI assistant Muse, which can help users send emails, book travel, and complete transactions. It adopts a free basic version and paid subscription models at USD 20 and USD 100 per month. According to Apptopia data cited by Reuters, cumulative downloads reached 2.8 million in the first 12 days after launch, indicating strong consumer demand for task-execution AI. On the ecosystem side, Shopify announced it is advancing integration of Muse with Shop Pay checkout capabilities, connecting agent shopping with merchant transactions; Amazon, meanwhile, restricted Muse's access to its platform, reflecting that during agent development, the importance of traffic gateways, transaction control rights, and data authorization has further increased. The bank believes that Muse's industrial significance lies in expanding the ways AI value is realized: as agents participate in product screening, service booking, and transaction execution, the business model is expected to extend from subscription payment further to transaction services, and application ecosystem and execution capability will become key to competition.
The Yunqi Conference focused on collaboration among models, chips, and cloud infrastructure, promoting domestic AI toward deeper large-scale application. At the 2026 Yunqi Conference that opened on September 22, Alibaba disclosed that Qwen 4 is being trained, with subsequent Qwen 4.5 and Qwen 5 planned to evolve toward a parameter scale of 5 trillion to 10 trillion, while continuing to advance model self-iteration and multimodal capability building. On computing power, the company released the new-generation Zhenwu V900 chip, disclosed that its performance reaches three times that of the previous-generation M890, and said that a cluster based on this chip is planned to support up to 500,000 cards. It also proposed a goal of global data center capacity exceeding 20GW by 2032. The above indicators are product capabilities and construction plans disclosed by the company. On applications, Qianwen Office launched an enterprise context product and connected Salesforce customer management capabilities, promoting the embedding of agents into actual enterprise business processes. The bank believes that the two developments together reinforce the industrial logic that application deployment drives computing power demand. Consumer-side agents expanding use scenarios and enterprise-side agents deepening process participation are expected to increase demand for multi-turn reasoning, tool invocation, and continuous operation, driving consumption of computing, storage, and network resources. In the future, focus should be placed on user retention and paid conversion, enterprise deployment results, and fulfillment of cloud vendors' construction plans, while tracking the demand transmission in infrastructure segments such as high-speed optical modules, switches, servers, liquid cooling, and power supply.
Risk warnings: changes in the international environment may affect the security and stability of supply chains; tariff impacts may exceed expectations; the development of the artificial intelligence industry may fall short of expectations; market competition may intensify; exchange rate fluctuations may affect exchange gains and gross margins of export-oriented enterprises; telecom operators' cloud computing business development may fall short of expectations; operators' capital expenditure may fall short of expectations; cloud vendors' capital expenditure may fall short of expectations; demand in the communications module and intelligent controller industries may fall short of expectations.
Comments