Telecom Giants Deploy New Infrastructure for Large-Scale AI Computing
On 10 March 2026, China Mobile and ZTE Corporation announced the launch of their Scale-Across (GSE-DCI) solution, a new high-bandwidth infrastructure technology designed to coordinate computing power across multiple data centers. The system was unveiled at MWC Barcelona 2026, one of the world’s largest telecommunications technology events.
The deployment marks a significant step in building next-generation intelligent computing networks, where artificial intelligence workloads must move efficiently between distributed computing facilities.
What Technology Was Deployed
The Scale-Across (GSE-DCI) platform is a data center interconnect (DCI) solution designed to link large-scale computing infrastructure across geographic locations.
It integrates several core technologies, including:
- SuperPipe packet spraying, which distributes traffic across multiple network paths to maximize throughput
- Fast CNP (Congestion Notification Protocol) to rapidly control network congestion
- 800G optical networking hardware, providing ultra-high bandwidth connectivity
The system was validated on an 800-gigabit platform, enabling long-distance, high-speed interconnection between computing clusters.
This architecture allows cloud providers and telecom operators to synchronize computing power across multiple facilities, ensuring that AI training and inference workloads can be distributed dynamically.
Why the Deployment Matters
AI systems increasingly rely on large clusters of GPUs and specialized processors, often spread across multiple data centers. Traditional networking systems struggle to coordinate these distributed workloads efficiently.
The new solution directly addresses three major infrastructure challenges:
- Low transmission efficiency between computing centers
- Poor resource utilization across distributed clusters
- System instability when scaling high-performance AI workloads
By enabling high-bandwidth, low-latency, and high-reliability data center interconnection, the platform creates the networking backbone required for large-scale AI computing environments.
Building the Infrastructure for AI-Native Networks
The launch also signals the telecom industry’s shift toward AI-native network architecture, where telecommunications infrastructure is designed not only for data transmission but also for distributed computing coordination.
As operators expand AI services, hyperscale data centers and telecom networks are increasingly merging into a single infrastructure layer.
China Mobile and ZTE’s deployment demonstrates how telecom providers are evolving into providers of computing power infrastructure, supporting industries ranging from autonomous systems and financial modeling to advanced AI training platforms.
A New Phase of Global Computing Infrastructure
The Scale-Across solution positions telecom networks as a core layer of the global AI economy.
As AI workloads grow exponentially, technologies capable of coordinating computing power across large geographic networks will become essential infrastructure.
The launch at MWC Barcelona 2026 signals that the race to build AI-scale networking systems has already entered its deployment phase.
