Huawei proposes AI-native core network for agent services

Huawei has detailed its vision for a new core network architecture that natively supports artificial intelligence, aiming to manage the growing number of autonomous AI agents. Traditional mobile networks, primarily focused on voice, text, and internet access, will need to evolve to coordinate vast numbers of these agents effectively.

The company believes integrating AI directly into the network's interfaces, functions, and procedures is crucial for operators to move beyond providing basic connectivity and deliver intelligent network functions for next-generation digital services. Huawei projects that by 2035, there could be as many as 900 billion AI agents globally, necessitating a network infrastructure capable of supporting substantial growth in connections and services.

To meet these demands, Huawei has identified three key scenarios: Agent Over New Radio (AoNR) for low-latency task completion, Wide-Area Computing Network (WCN) for efficient utilization of computing resources, and Agent Computing Network (ACN) for secure collaboration among multiple agents. These technologies are designed to ensure reliable service operation and optimized resource management.

The proposed next-generation core network architecture includes intent-based interfaces for direct task requirement transmission, a new Data & AI Plane (DAP) alongside traditional control and user planes, and deeply integrated AI-driven processes. Huawei anticipates this evolution in two phases: the first focusing on AI-enhanced customer experiences and the second on delivering "enhanced connectivity" and "connectivity beyond boundaries."

This transition to AI-native networks requires comprehensive collaboration across the ICT ecosystem. Huawei suggests that this new architecture will unlock new business opportunities in the era of intelligence by enabling novel connectivity and services.