Shanghai — At its first Data Center Innovation Summit held during HUAWEI CONNECT 2025, Huawei introduced a suite of new solutions to accelerate and improve the construction, efficiency, and reliability of AI data centers globally. Over 700 industry leaders and experts gathered to discuss how to push DC infrastructure toward higher performance.
Huawei emphasized that data centers are at a “turning point” — evolving into AI DCs that demand superior computing efficiency, seamless data-AI convergence, resilience, and reliability. Key innovations span six major areas: large-scale compute clusters; agile compute resource pools; vector databases; AI data lakes; multi-site active-active architectures; and elastic facility design.
One highlighted solution is the Xinghe AI Fabric 2.0, a network architecture built around three layers: AI Brain, AI Connectivity, and AI Network Elements. It integrates components such as the AI network agent (NetMaster), scheduling engines for training and inference, and automation tools to ensure “always-on” networks with full compute capability.
In addition, Huawei announced advancements in its AI data lake platform, combining high-performance all-flash storage, data resilience tools, and unified data management to support industries facing growing demands for data processing.
Looking ahead, Huawei said it will continue collaborating with partners worldwide to foster sustainable and high-quality infrastructure, delivering faster, more robust AI-modernized data centers for the intelligent future.
