The rapid integration of AI agents—ranging from automated coding assistants to marketing tools—is fundamentally altering business operations. Gu noted that this transition moves execution from human-led application usage toward collaborative agent-to-agent workflows. This shift demands a departure from traditional siloed resource management, favoring an integrated approach where compute, data, and models are packaged for agent-driven tasks.
To accommodate these needs, Huawei is upgrading its Cloud Stack platform across four pillars: infrastructure, data, AI, and ecosystem. The new infrastructure utilizes A5 SuperPoDs and NPU pooling to push compute utilization from 30% to 70%. By decoupling storage and compute, the company aims to transform enterprise data into a reliable source of productivity. Furthermore, Huawei is deploying domain-specific foundation models, including Vision-Language and scientific computing tools, to facilitate high-performance inference. Through its Industry AI Foundry, the company intends to standardize these solutions across sectors like finance and government, ensuring that AI moves beyond experimental deployment into core production cycles.

Comments (0)
No comments yet. Be the first!