Cloud · September 25, 2026
Huawei Cloud launches global AI Cluster Service and expands Agentic Cloud ecosystem
Huawei Cloud announced the global launch of its latest AI Cluster Service, a move that marks a significant step in strengthening its cloud computing infrastructure and laying the foundation for Agentic AI. The announcement was made during a keynote speech at the HUAWEI CONNECT 2026 event, where the company also detailed the progress of its broader Agentic Cloud strategy aimed at serving enterprise needs.
The latest version of the AI Cluster Service introduces a five-level rapid recovery mechanism and comprehensive system status monitoring capabilities. These features are designed to support stable and continuous model training on the cloud for periods exceeding 40 days. The system is capable of recovering from failures within 10 minutes. Through optimizations in resource allocation, caching, and algorithms, the service achieves a token processing rate that is 20 percent higher than previous generations of computing services. Commercial availability for this service is scheduled to begin in China on September 30, with markets outside China following on November 30.
Huawei Cloud also highlighted the development of its Context Memory Storage solution, which is specifically designed for Agentic AI tasks that require long-duration continuous operation. This solution supports petabyte-level memory capacity, which is double the capacity of comparable products in the industry. It also enables high-speed memory reads at the terabyte level, delivering performance that is 50 percent higher than similar industry offerings. These capabilities address the specific storage and access efficiency requirements of modern AI agents.
The Agentic Model as a Service platform aggregates diverse AI models on a single interface, allowing developers to access state-of-the-art models from leading providers with a single click. This approach eliminates the need for manual installation or deployment of models. During the event, MiniMax demonstrated an advanced multimodal model that, when integrated with Huawei Cloud, expands access to cutting-edge multimodal capabilities for a wider range of users.
Huawei Cloud is advancing its enterprise-level Agentic AI platform through both commercial and open-source channels. The platform is built on the foundation of AgentArts and the open-source openJiuwen. It provides access to more than 5,000 general-purpose Model Context Protocol resources and over 1,000 industry-specific MCP resources. This infrastructure helps organizations develop, deploy, and manage AI agents more easily. The platform currently serves more than 100 customers, including government bodies, business organizations, research institutions, and educational establishments.
Among the customers utilizing the platform are the government of the Longgang district in Shenzhen, China Southern Power Grid, the Guangzhou Laboratory, the University of Science and Technology of China, Kingsoft Office, Sichuan Yingu Carbon Sink Renewable Resources, Changsha Thunder Cloud Network Technology, and KingMed Diagnostics. Kingsoft Office shared how it integrates the Huawei Cloud agent platform with its WPS 365 Document Center and WPS Comate to develop specialized office AI agents. These agents are currently in use across various industries, including the financial sector and government agencies.
AgentArts is set to launch commercially in markets outside China on December 30. The open-source openJiuwen community has accumulated more than 50,000 stars and over 3.29 million downloads. Huawei Cloud has also partnered with Chinasoft International, iSoftStone Group, and Beiming Software to launch a commercial version for partners, aiming to expand market opportunities and support partner growth on Huawei Cloud's computing infrastructure.
The Industry AI Foundry aggregates industry-specific assets, including use cases, models, data, knowledge, and AI agents. It transforms project-specific experience into reusable capabilities for the industry. The Foundry currently includes five specialized zones: Smart Healthcare, Embodied AI, AI for Science, Smart Manufacturing, and Smart Finance. It contains more than 1,000 industry-specific assets and supports over 1,000 real-world deployment projects. Two new zones, the Smart Government Zone and the AI Hardware Zone, were recently added to the Foundry. The Smart Government Zone already has 24 co-founding partners covering various use cases from government operations to urban management.