Speaking during a keynote at the Mobile Broadband Forum (MBBF) 2024, the corporate SVP opined that the convergence of mobile technology and AI will transform telecom carriers' network infrastructure, operations, and business models.
The Huawei executive predicted that by 2030, most people will rely on personal AI agents for home, work, and travel assistance. These AI assistants will operate continuously, generating and processing what Li claims will be ten times more data than current levels.
“AI agents will change how we work and live. They will bring intelligent services everywhere and drive explosive growth in data,” Li said.
Li explained that the surge in AI-driven data will require complete overhauls of network architectures.
He suggested that traditional one-way traffic flows from data centres to devices will evolve into multidirectional streams of personalised data flowing between cloud services, edge computing facilities, and devices.
“We’re going to see a rise in east-west traffic, and even mesh connections between multiple types of devices and hosts,” Li said. “With the structural changes in traffic models, network optimisation will be more critical than ever.
“We need to provide larger uplink bandwidth, deterministic latency, and more reliable network capabilities to meet these new service requirements.”
To address network challenges faced by AI, Li outlined four areas requiring transformation: network services, infrastructure, operations and maintenance (O&M), and business models.
Li pointed to the early adoption of consumer services in China, where operators have launched '5G New Calling' services enabling AI-powered features such as digital avatars and real-time translation during voice calls — with more than 24 million users already using such features.
The enterprise market represents another significant opportunity, with Li suggesting carriers can provide affordable AI services to small and medium-sized enterprises while offering larger industrial customers combined connectivity and AI capabilities.
However, supporting these services will require substantial network improvements. Li noted that AI assistants need less than 300 milliseconds of end-to-end latency for natural interaction, with air interface latency under 20 milliseconds.
“Carriers can build AI-centric networks that support deterministic access, elastic scheduling, and lossless WAN. This can help deliver on-demand, reliable connections between cloud, edge, and devices,” he said.
Network maintenance will also become more complex, with Li suggesting, however, that AI-powered automation could help address maintenance challenges.
He described systems where AI agents could handle real-time simulation of network data and automate task planning, while AI copilots could assist field engineers with hardware maintenance.
Li outlined: “For service operations, AI agents can support real-time simulation of multi-modal data. This helps carriers to evaluate network resources more efficiently, and provision new services more quickly.
“For network maintenance, AI agents can automate task planning and orchestration, solving problems caused by software. At the same time, copilots can help field engineers to quickly locate and fix any hardware problems.”
On the business front, Li highlighted that more than 30 European carriers have introduced speed-based mobile plans, indicating consumers' willingness to pay premium rates for guaranteed performance.
He suggested that Chinese carriers are exploring multiple revenue streams through AI-based services, including cloud computing power, storage, and VIP services.
During his keynote, he outlined those carriers in China, the Middle East, Europe, and Asia Pacific have already begun testing AI service capabilities on live 5G networks across various scenarios including individual, home, travel, and business use cases.
“They are setting the benchmark for industry innovation, and we can learn from their experience,” Li said.
Li used his keynote to emphasise two key priorities for the industry: preparing networks to support AI through improved uplink speeds, latency and capacity; and leveraging AI to manage increasingly complex networks through automated maintenance and optimisation.
“The opportunities are great. And the best time to act is now,” Li said. “Huawei is ready to work with carriers and industry partners to build networks for AI, and AI for networks.”
