Partner Article: Posted by
Harry Baldock : As AI becomes increasing ubiquitous, the capabilities of 5G-Advanced (5G-A) is growing vital to operator success
Since the release of 3GPP’s Release 18 last year, the mobile industry’s adoption of 5G-A has been steady. Over 30 mobile operators worldwide have already launched the technology commercially, with many more expected to follow in the coming year. As expected, China has taken an early lead in this regard, with China Unicom, China Mobile, and China Telecom collectively making the 5G-A available in 300 cities, in 30 provinces, to over 10 million subscribers.
The advantages of the new technology are undeniable. With downlink rates of up to 10 Gbps, improved upload speeds, lower latency, greater reliability, and precision positioning, 5G-A is unlocking a host of new use cases, from autonomous vehicles to industrial automation. But while these use cases have long been coveted by operators, in fact the greatest benefit of 5G-A is still emerging: its integration with AI.
AI: A catalyst for 5G-A transformation
Since the explosive launch of ChatGPT at the end of 2022, AI usage has grown exponentially both at home and at work. Recent
research from Salesforce showed that AI usage in the workplace is soaring, rising 233% in the last six months alone.
Speaking at this year’s MWC Shanghai, James Chen, President of Huawei’s Carrier Business, noted that the technology was rapidly becoming more affordable, allowing it to be integrated more freely with existing systems.
“As AI adoption continues to rise, the annual decline in AI inference costs exceeds 90%, making AI as ubiquitous as water and electricity,” he said, emphasising that the networks, functioning as the “central nervous system” for AI, must evolve to meet this rising demand.
5G-A is foundational for this evolution, enabling intelligent services for both consumer and B2B sectors. With AI agents now capable of executing real-time instructions, such as uploading a screenshot to the cloud within seconds, network design must pivot toward upstream capacity and low latency.
“In the same way that 5G laid the foundation for the success of mobile internet, 5G-A is the key supporting pillar of the mobile AI era,” said Chen.
New connectivity: IoT and the Edge
To maximise the potential of mobile AI’s integration with 5G-A, operators need to consider three strategic imperatives, described as “new connectivity, new network planning, and new business models” by the VP of Huawei’s Wireless Network Product Line, David Li.
‘New connectivity’ here refers to a focus on connectivity technologies that have a major role to play in the growing AI ecosystem. Edge computing, for example, will be crucial in this context, allowing AI inference to take place closer the end-user. To this end, Huawei is investing in “cell-free” technologies to ensure uniform signal quality from cell center to cell edge, vital for billions of intelligent endpoints.
Another growing focus area will be on the IoT. As the number of IoT devices continues to grow and be increasingly infused with AI, supporting this ecosystem is crucial to industrial and enterprise success. Here, 5G Reduced Capability (RedCap) will play a leading role, allowing 5G network to support intelligent IoT at scale. In future, Enhanced RedCap (eRedCap) will take this one step further, allowing the IoT become more sophisticated and support video surveillance and IoT livestreaming.
“Today, the cost of RedCap modules has dropped below 100 RMB (~$14), and in the next year or two the cost of RedCap will be comparable to or even lower than that of Cat-4 IoT,” said Li. “With IoT technology, our vision is to connect all physical assets in the network and convert them into data simultaneously.”
New network planning: Prioritising uplink and latency
A major challenge presented by the boom in mobile AI usage is the strain it places on networks in terms of uplink. Historically, networks have been deployed with a focus on downstream capacity, with users focussed on downloading content. With AI, however, users are uploading data far more frequently and expecting AI-generated answers in real-time – a challenge that will only grow as AI agents like Zhipu AI begin uploading data on users’ behalf. As a result, high uplink capacity and low latency will become a crucial feature of, representing a significant paradigm shift in network deployment strategy.
In this regard, Li argues that telcos should be aiming for the “double 20” benchmark of 20 Mbps upload and 20 ms air interface latency, to ensure their networks are ready to deliver novel AI use cases effectively.
New business models: An AI opportunity
With these new connectivity technologies and a more robust 5G-A network in place, operators can begin to explore more flexible, user-centric business. Instead of focussing their proposition on data traffic and KPIs, operators can instead create unique packages targeting specific demographics, whether they are latency-sensitive gamers or live streamers. With the integration of AI within the network, these services can be more easily provisioned, offering quality of experience assurance autonomously at the click of a button.
“Traffic monetisation has hit a bottleneck. Operators need to find new monetization methods, namely seizing the entry points of intelligent services,” said Li. “Our networks should be flexible, able to identify what users, scenarios, and applications are running on the network, and adapt accordingly.”
5G-AxAI in action in China
At MWC Shanghai, one did not have to look far for examples of the powerful combination of 5G-A and AI. China Unicom Beijing and Huawei, for example, were showcasing the world’s first 3D smart 5G-A network, deployed in Beijing. China Unicom Beijing is leading the transition from single-band networks to a layered communications system, with high bands for network capacity, mid-bands for continuous coverage, and low bands and space-air-ground coordination for wide-area coverage. At the same time, integrated AI is helping to automate the network, with the partners aiming “to achieve automation across all network settings by 2026, covering site deployment, maintenance, optimization, and complaint handling”, according to a company
press release.
Further to the north of China, in Harbin, China Unicom Heilongjiang deployed 5G-A for the 2025 Asian Winter Games, providing full coverage across all the event’s venues and enabling peak download rates of over 10 Gbps using millimetre wave and three-carrier aggregation. As part of the company’s broader ‘CHARMS’ (Cloud, Hi-tech, AI, Reduction, Metaverse, and Safety) strategy, the 5G-A network was used to support real-time 8K livestreaming, intelligent drone monitoring, and AI-powered network scheduling, which established a solid digital foundation for the “Ice and Snow City”.
A collaborative industry drive to take the next steps
As 5G-A and AI technology become more mature we will see their further integration in deployments worldwide. For now, however, there remains much work to be done, and operators are at risk of missing out on a major opportunity for revenue growth. There is a deep need to foster collaboration across the telecoms industry, build new value chains, and seize the growth opportunities this convergence brings.
“The era of AI agents is here, ready to reshape how we live and work. Grounded in powerful networks, sharpened by relentless innovation, and propelled by thriving ecosystems,” concluded Eric Fang, President of the 5G-A Domain of Huawei Wireless Network Product Line. “Let’s work together to advance 5G-AxAI demands and applications, foster cross-domain collaboration, and drive industrial and economic growth.”
IANS Photo
