Huawei anticipates a seismic shift in mobile network dynamics with the surge of AI agent applications. These applications are slated to generate significantly more uplink traffic compared to traditional mobile services, necessitating an evolution in network architecture. Notably, Huawei predicts the number of AI agent connections could skyrocket to a staggering 100 billion, leading to unprecedented demands on mobile networks, cloud infrastructure, and edge computing systems.
To adapt to these impending changes, Huawei is introducing new metrics and protocols. Among them, the AI MOS is a pioneering method to evaluate AI service quality. This benchmark resembles the Mean Opinion Score, traditionally used for measuring voice service quality. Additionally, a new protocol, A2A-T, is proposed to facilitate efficient agent-to-agent communications across telecom networks.
Zhao Dong, vice president and chief marketing officer of the wireless network product division at Huawei, elaborated on these developments during a media and analyst roundtable. He described the onset of the “agent era,” where AI agents will function in various environments, interacting with users, devices, and other agents. Zhao highlighted, “We already stand at the dawn of the agent era. Agents go to your mobile phone, go to cars, go to robots.”
The transition to an AI-driven ecosystem also implies a shift in traffic patterns. Traditional mobile networks emphasized downlink traffic, but as Zhao explained, “This time agents need more uplink.” In an example from a Shanghai museum guide trial, AI agents used uplink traffic for analyzing artworks, contrasting with the typical downlink-heavy nature of conventional networks.
The changes heralded by AI agents align with Huawei’s broader vision of an “agent-oriented network,” promoting extensive AI agent interactions. Zhao stressed the need for “collaboration across the broader technology ecosystem,” involving terminal providers, chipmakers, vendors, and service providers.
Ultimately, Huawei’s proposed solutions aim to redefine how network capacity, service quality, and communication protocols are structured to accommodate AI’s growing influence in telecommunications. As such, these advancements mark a pivotal moment in tech evolution, compelling mobile networks to adapt proactively in the AI-driven future.
