The rapid proliferation of artificial intelligence across consumer and industrial sectors has necessitated a fundamental transformation in how global telecommunications infrastructure handles unprecedented data volumes and real-time processing requirements. During the recent MWC Barcelona gathering, industry leaders examined the critical role of optical networking in supporting the next generation of digital services. Huawei introduced a strategic framework for an AI-centric all-optical target network, a move designed to address the surging demands for high-speed connectivity and lower latency. Presented by Kim Jin, Vice President of Huawei’s Optical Business Product Line, this proposal shifts the focus toward four primary enhancements: bandwidth, reliability, coverage, and latency. The ultimate objective is to facilitate a transition for telecommunications operators toward Agentic Ultra-Broadband (UBB) networks. These networks are no longer just passive pipes but are becoming active participants in the delivery of high-performance AI experiences, ensuring that the underlying infrastructure does not become a bottleneck for modern technological innovation.
By prioritizing a shift toward deterministic networking, the industry aims to provide a more consistent user experience regardless of the physical distance between data centers and end-users. This strategic pivot recognizes that the AI era requires more than just raw speed; it demands a level of network intelligence that can anticipate and adapt to changing traffic patterns in real time. The framework advocates for a holistic upgrade of existing systems to support the massive throughput required by large language models and distributed computing tasks. For operators, this represents a significant shift from traditional connectivity models to a more integrated, service-aware architecture. The discussion at the event highlighted how these improvements in optical foundations could redefine the competitive landscape, allowing carriers to offer premium, low-latency tiers of service that were previously impossible. This evolution is seen as a necessary response to the growing complexity of the global digital ecosystem, where the ability to move data efficiently is as important as the ability to process it.
Redefining Connectivity: Home and Metro Network Optimization
A central pillar of the newly proposed strategy involves the optimization of connectivity at the most granular levels, specifically within home and metropolitan environments. To address the increasing density of smart devices and high-bandwidth applications, the focus has shifted toward the deployment of Wi-Fi 7 tri-band FTTR solutions. These systems are engineered to provide stable 4 Gbps speeds and seamless roaming, ensuring that users maintain high-performance access as they move through different areas of a building. This technological leap is complemented by the integration of next-generation PON technology, which utilizes application-level slicing and intelligent scheduling to prioritize critical data streams. Such granular control allows the network to distinguish between high-priority tasks, like real-time AI processing or video conferencing, and standard background traffic. By managing resources at this level of specificity, operators can guarantee a higher quality of service and reduce the friction that often plagues shared residential or commercial broadband connections.
Beyond the home, the strategy extends into the metropolitan layer by introducing Optical Cross-Connect (OXC) solutions to metro networks. This extension is vital for achieving the millisecond-level computing latency required for edge-based AI services to function effectively. By reducing the number of hops and moving optical switching closer to the user, the network can minimize the delays associated with traditional electronic processing. This approach effectively creates a low-latency “computing circle” within metropolitan areas, enabling a wide range of new applications from autonomous vehicle coordination to real-time industrial automation. The convergence of high-speed fiber access and intelligent metro switching forms a robust foundation for the Agentic UBB vision. It allows the network to function as a unified, high-performance fabric that bridges the gap between local processing and remote data centers. This localized efficiency is essential for the widespread adoption of AI tools that rely on instantaneous feedback and massive data synchronization across diverse geographical points.
Strengthening the Backbone: High-Capacity Transmission and Reliability
On a broader scale, the vision for an AI-centric network necessitates a significant overhaul of backbone infrastructure to accommodate the vast quantities of data moving between regions. The strategy advocates for the adoption of a 3D-mesh architecture paired with 400G and 800G transmission capabilities, which significantly increases the capacity of long-haul fiber routes. This evolution is further supported by Super C+L band technology and high-capacity submarine cables, which are designed to eliminate data bottlenecks in traffic-heavy corridors. By expanding the usable spectrum of the fiber, operators can dramatically increase the amount of information transmitted over existing physical infrastructure. This approach not only improves efficiency but also provides a scalable path for future growth as data demands continue to climb. The focus on high-capacity transmission ensures that the core of the network can handle the surge in traffic generated by centralized AI training and the subsequent distribution of large-scale models.
Maintaining service continuity in such a high-stakes environment requires a focus on reliability and self-healing capabilities. To this end, the proposal incorporates “zero-outage” features, such as wavelength switched optical network (WSON) technology, which facilitates ultra-fast 50ms switching to protect against fiber cuts or equipment failures. This level of resilience is non-negotiable for industries that rely on 24/7 connectivity for critical operations. Furthermore, the shift toward intelligent operations and maintenance (O&M) is managed through the deployment of AI agents like FANSpirit for access domains and OTNSpirit for transmission. These agents allow operators to automate complex tasks, such as path optimization and fault detection, based on specific latency and bandwidth requirements. By utilizing machine learning to manage the network, carriers can respond to issues before they affect the end-user, significantly improving the overall stability of the system. This proactive management style represents a departure from traditional reactive maintenance, setting a new standard for network uptime.
Future Horizons: Unlocking Value Through Intelligent Infrastructure
The consensus among the nearly two hundred industry leaders present at the announcement was that the AI era represents a decade-defining opportunity for growth and structural evolution. Representatives from various global telecommunications organizations emphasized that building a unified, high-capacity, and intelligent optical foundation is the only way to unlock significant network value. By investing in these deterministic, high-speed networking strategies, operators can position themselves as essential providers of the infrastructure necessary for the next generation of smart home services and AI innovations. The discussion moved toward how these technologies would be integrated into existing business models, with many experts suggesting that the ability to offer guaranteed latency and bandwidth would become a primary differentiator in the market. This strategy reflects a broader industry trend toward networks that are not only faster but also significantly more aware of the content and requirements of the data they are carrying across the globe.
To capitalize on these advancements, stakeholders identified several actionable steps that must be taken to ensure the successful rollout of AI-centric optical networks. First, the industry must prioritize the standardization of 800G and beyond to ensure interoperability between different equipment vendors and regional networks. Second, operators were encouraged to accelerate the decommissioning of legacy copper and lower-speed fiber systems in favor of FTTR and OXC-enabled architectures. This physical modernization is the prerequisite for implementing the AI-driven management layers that provide real-time optimization. Additionally, there was a clear call for increased collaboration between cloud service providers and network operators to align data center locations with low-latency optical hubs. By synchronizing infrastructure investments, the ecosystem can more effectively support the heavy computational loads of modern AI applications. Ultimately, the move toward an intelligent optical foundation was viewed as the critical path toward a more responsive and efficient digital future, where the network and AI work in tandem to drive innovation.
