Huawei's Agentic Core targets the gap between AI hype and network reality

The telecom giant's MWC 2026 release aims to give operators the infrastructure to monetize AI demand–not just carry it.

Huawei launched its Agentic Core solution, a network architecture it says is built to address the practical pressures AI is placing on telecom operators: surging traffic, increasingly fragmented service requirements, and the persistent difficulty of turning connectivity into a revenue stream at Mobile World Congress 2026 in Barcelona.

The solution is structured around three intelligence layers–NE (network element) intelligence, network intelligence, and service intelligence–each targeting a different dimension of the problem. The first layer addresses infrastructure at the device and connection level. As AI agents become embedded in smartphones, embodied robots, and autonomous vehicles, the number of connected entities is projected to grow tenfold.

Huawei argues this demands more than bigger pipes; it requires new capabilities including digital identity management, agent registration and discovery, and Agent-to-Agent (A2A) session management. The intent is to lay the groundwork for physical AI to operate reliably at scale–something that current network architectures were not built to support.

The second layer–network intelligence–deals with how operators respond to the diversity of demands those agents will generate. An AI-powered industrial robot and a consumer streaming an immersive video call have fundamentally different latency and bandwidth requirements.

Huawei's pitch is a shift away from static, rule-based network configuration toward what it describes as intent-driven networks: AI agents within the network itself that interpret the needs of different organisations, dynamically allocate resources, and handle the full cycle of policy generation, configuration, and delivery in a closed loop.

The third layer is where the commercial logic sits. Huawei's AISF (AI Service Framework) is positioned to help operators compete with over-the-top players by moving beyond connectivity into services–personal AI assistants that integrate communication, content, and transaction capabilities; immersive communication experiences that go beyond voice; and infrastructure-level computing power sold as a service to support AI inference and content generation.

For operators across Asia Pacific, the framing is particularly relevant. The region's telcos face structurally low ARPU and increasing network investment demands driven by 5G rollouts. The question of how to generate returns from AI-era infrastructure–rather than simply enabling the AI businesses of others–is one that most carriers have yet to convincingly answer.

Huawei described the solution as part of a "three-layer intelligent collaboration" strategy with operators, with the stated goal of delivering what it calls 7x24-hour inclusive intelligent connectivity. The language is aspirational, but the underlying engineering challenge is real: AI is restructuring demand patterns on telecom networks faster than existing architectures were designed to accommodate.