Huawei Launches Next-Gen Optical Nets

Why it matters

The massive data processing required by AI is creating a network bottleneck, with some experts forecasting that AI could push annual bandwidth growth rates from an average of 20-30% to nearly 40%. The total volume of data generated each year has already increased eightfold between 2016 and 2024, creating immense strain on existing infrastructure. At MWC Barcelona 2026, Bob Chen, President of Huawei's Optical Business Product Line, framed the company's strategy in two directions: "AI for Networks" and "Networks for AI." This dual approach uses artificial intelligence to improve network management while simultaneously enhancing network architecture to better support AI-driven services. The "AI for Networks" pillar focuses on operational efficiency. It employs AI for intelligent fiber sensing that can pinpoint fault locations to within 10 meters and for real-time traffic analysis that can reduce board and port energy consumption by an average of 40%. An optical performance simulation model, built on thousands of parameters, also extends transmission distances by 20%. To build "Networks for AI," Huawei is aiming for specific latency targets crucial for distributed AI applications. The goal is to create "latency circles" of 1 millisecond for metro areas, 3 ms for regional networks, and 5 ms for national networks, ensuring millisecond-level access to computing power. This new generation of Optical Transport Network (OTN) products is designed for an evolution toward 400G and 800G transmission speeds to handle the increased traffic. The architecture utilizes end-to-end all-optical switching (OXC) to create a mesh network that reduces latency and boosts reliability for always-on AI services. These advancements align with the ITU-T's ION-2030 vision, a global standardization roadmap for the next generation of optical networks. Huawei is a major player in this market, where it, along with Ciena and Nokia, controls a significant portion of the sales for core optical transport systems.

Key numbers

• The massive data processing required by AI is creating a network bottleneck, with some experts forecasting that AI could push annual bandwidth growth rates from an average of 20-30% to nearly 40%.
• The total volume of data generated each year has already increased eightfold between 2016 and 2024, creating immense strain on existing infrastructure.
• It employs AI for intelligent fiber sensing that can pinpoint fault locations to within 10 meters and for real-time traffic analysis that can reduce board and port energy consumption by an average of 40%.
• An optical performance simulation model, built on thousands of parameters, also extends transmission distances by 20%.

What happens next

• The massive data processing required by AI is creating a network bottleneck, with some experts forecasting that AI could push annual bandwidth growth rates from an average of 20-30% to nearly 40%.
• To build "Networks for AI," Huawei is aiming for specific latency targets crucial for distributed AI applications.
• These advancements align with the ITU-T's ION-2030 vision, a global standardization roadmap for the next generation of optical networks.

Sources

• Huawei unveiled
• The massive data processing
• At MWC Barcelona 2026
• This new generation of
• Huawei is a major player

Quick answers