At MWC Barcelona 2025, Li Peng, Huawei’s Corporate Senior Vice President and President of ICT Sales & Service, delivered a keynote on how carriers can make the most of AI to fully unleash the value of their networks. Peng predicts that symbiosis between 5G-A and AI technologies will stimulate double-digit growth in both DOU (data of usage) and ARPU (average revenue per user) from mobile subscribers.
“We’re rapidly entering a fully intelligent world,” said Peng. “Intelligent applications are spreading everywhere, placing new demands on networks. By embracing and evolving 5G, we can unlock the infinite potential of mobile networks. Huawei is willing and ready to work with carriers and industry partners around the world to promote digital enablement, reinforce network foundations and bring AI to all. Together, we can shape the DNA for an intelligent world.”
AI is changing human-machine interaction, driving different requirements for latency
With advancements in AI, HMI (human-machine interaction) is evolving from simple text-based communications to voice, gestures and more multi-modal interactions. As a result, HMI is more real-time and convenient than ever, giving rise to a new wave of innovative applications. For example, people can interact more naturally with their devices using AI-powered voice assistants. On cloud phones, AI-powered avatars can provide visual feedback as well, creating a more personal experience for services like health monitoring, making the mobile experience far more accessible and productive for different groups of users.
To support applications like these, however, networks need to be able to provide guaranteed latency, which will require ongoing evolution from 5G NSA, to 5G SA and eventually 5G-A. Carriers can also adopt innovative technologies like CUPS (Control and User Plane Separation) and GBR (Guaranteed Bit Rate) to reduce basic latency and ensure differentiated, deterministic latency for specific scenarios.
AI-enabled content production and distribution is raising the bar for upload download speeds
Peng noted that AI will transform how content is produced and distributed. For example, AIGC technology makes it possible to generate hour-long 2D and 3D videos with a single click. Meanwhile, AI recommendations are more targeted than ever, allowing the distribution of more personalised content to broader audiences across the Internet. Both trends will cause network traffic to surge over the next five years, placing unprecedented demands on networks. To keep up, carriers will need more spectrum, greater network capacity and much larger uplink and downlink bandwidth.
Diverse AI services will need experience-centric network coverage
Both AI-powered cloud and mobile devices are making intelligent services more accessible and the industry will see growing demand for experience-centric network coverage. According to third-party data, cloud phones and cloud drives will be used by over 1 billion people by 2030, each of whom will need fast access to cloud computing power. In addition, intelligent in-vehicle applications will require full coverage across cities, highways and the countryside to provide a continuous and reliable mobility experience.
Moving forward, meeting these demands will require ongoing progress in network deployment, from rapid expansion of 5G NSA networks to 5G SA networks for a more seamless indoor/outdoor experience and eventually to experience-centric 5G-A networks. This will help carriers expand network coverage and ensure a smooth experience for tens of billions of new connections for people and hundreds of billions of new IoT connections.
Growing network complexity will drive evolution towards application-oriented O&M
AI will bring more complex application scenarios and a more diverse range of experience requirements. From a networking perspective, this will drive a shift from traditional, resource-oriented O&M to a more application-oriented approach.
Some carriers are already developing O&M systems based on AI agents. For operations enablement, these AI agents can use digital twins to predict personalised needs for individual users, helping shorten service time-to-market from days to minutes. For network maintenance, AI agents with self-learning capabilities can predict and locate faults in seconds, increasing troubleshooting efficiency by 30%. And for network optimisation, digital sandboxes can simulate the traffic of real-world applications, allowing AI agents to analyse traffic patterns and optimise networks 24/7 based on application needs.
Early-movers are scaling up 5G-A deployment to boost monetisation in the age of AI
“New network capabilities will give rise to new business models,” said Peng. “Carriers can go beyond monetising traffic and start monetising experience itself.”
For example, Chinese carriers are working with over 100 industries, including insurance and catering companies, to provide AI New Calling services through Open APIs. This has helped them increase income from industry customers by a factor of 10.
“The opportunities are huge,” concluded Peng. “And the time to act is now. Pioneers are already scaling up fast in over 200 cities around the world. They’re taking solid steps forward, unlocking incredible new value.”