Huawei said the “Network for AI” concept focuses on adapting optical infrastructure to support AI applications.
In summary, here’s what you need to know:
Dual AI Strategy – Huawei is building an optical roadmap centered around “AI for Networks” and “Network for AI,” which covers both operational optimization and support for AI workloads.
Improved performance and efficiency – Claimed improvements include 20% longer transmission distance, 20% faster Wi-Fi speeds under interference, and 40% average energy savings.
AI latency goals – The company outlined 5-ms, 3-ms, and 1-ms latency benchmarks for national, regional, and metropolitan optical networks.
At MWC Barcelona 2026, Huawei unveiled a new series of optical network products and solutions designed to address the growing technological demands related to artificial intelligence applications.
The announcement was made by Bob Chen, president of Huawei’s optical business product line, during Huawei’s product and solution launch event.
Chinese vendors say that as AI expands into home use cases such as entertainment, education and healthcare, and as computing scheduling becomes more prevalent, network requirements are increasing in areas such as bandwidth, latency and reliability.
Bob Chen said, “Huawei is promoting next-generation optical network solutions in two directions: AI for Networks and Network for AI. In AI for Networks, AI technology enables intelligent fiber sensing to improve network performance and user experience, improve O&M efficiency, and reduce energy consumption. In Network for AI, enhanced network capabilities enable carriers to build AI-centric all-optical target networks and accelerate AI adoption across homes and enterprises.”
Under the “AI for networks” approach, Huawei described several applications aimed at improving operational performance.
Intelligent fiber sensing uses fiber risk sensing and fault identification models to proactively detect potential problems and locate faults within 10 meters, according to the vendor.
To improve user experience, AI algorithms detect Wi-Fi interference in real-time and automatically adjust power levels to improve data rates by 20% under interference conditions.
Energy management is addressed through real-time traffic analysis that dynamically adjusts ports and boards. When there is no traffic, the components enter a hibernation state, reducing average energy consumption by 40%, Huawei explained.
Operations and maintenance (O&M) applies AI to planning, building, maintaining, and optimizing processes. For example, the Home Broadband O&M Agent identifies and locates more than 60 types of faults and assists network operations center engineers through natural language interactions, reducing the need for on-site visits.
The second direction, “Networks for AI,” focuses on adapting optical infrastructure to support AI applications.
In optical access, Huawei described a targeted network design that provides gigabit-level downlink and 100 Mbit/s-level uplink capacity to support new AI-driven home services.
For optical transmission, the company has outlined latency goals of 5ms for national networks, 3ms for regional networks, and 1ms for metro networks, with the aim of enabling millisecond-level computing access for AI workloads.
Huawei said it has introduced a series of next-generation optical access products including FTTR, OLT, ONT, and ODN, as well as OTN products for backbone, optical layer, and metro deployments. These are aimed at supporting the development of all-optical networks tailored to AI-related traffic requirements, the company added.
