O2 and Nokia report a 20% spectral-efficiency gain in live UK 5G trial

By Rebecca Stone

O2 and Nokia say a live 5G trial at O2's Giga Site in London Paddington improved spectral efficiency by approximately 20%, marking what they describe as the UK's first live deployment of an AI-powered radio access network (RAN) feature in a customer environment.

The trial operated on O2's live network at one of the UK's busiest transport hubs and served customers throughout its duration. Nokia's predictive AI, embedded within the RAN, analyzed network conditions, anticipated changes in demand and automatically adjusted spectrum and capacity allocation without manual intervention.

What did the live trial measure?

Traditional mobile sites have relied on largely static algorithms to manage radio resources. Nokia's system instead learns continuously from network conditions and changes how spectrum and capacity are allocated as traffic patterns shift.

The trial's clearest disclosed result was an approximately 20% improvement in spectral efficiency. In practical terms, the site could carry about one-fifth more data using the same spectrum and existing network capacity. O2 and Nokia also reported throughput gains during high-demand periods, while network accessibility and reliability remained stable.

The announcement did not quantify:

• exact peak or average throughput;
• latency, buffering or other customer-experience measures;
• the number of customers served or the trial duration;
• the method used to calculate the spectral-efficiency gain.

O2 said the efficiency gain supports faster downloads, higher-quality video streaming and reduced buffering in busy locations, but the announcement did not quantify those customer outcomes. Only the spectral-efficiency figure is numerical; the claimed service benefits remain directional.

How does this fit O2's £700m plan?

The deployment forms part of O2's £700m Mobile Transformation Plan, which combines targeted investment with smarter technology to improve network reliability, capacity and performance. The announcement did not disclose the trial's cost, Nokia contract value, allocation from the £700m programme or spending timetable. The 20% result therefore cannot be treated as a measured return on the wider investment.

O2's commercial rationale is to absorb rising mobile data demand by using existing spectrum and infrastructure more efficiently. That proposition matters where demand changes through the day, but the trial covered one London location. Replication elsewhere would show whether the gain follows traffic patterns, radio conditions and site design rather than reflecting Paddington alone.

Jeanie York, Chief Technology Officer at O2, said: “Mobile data demand continues to grow, so we need smarter ways to make the most of the spectrum and infrastructure we already have.

“This live trial with Nokia showed how AI can improve network efficiency and help us respond more effectively to changing customer demand. By making better use of existing network resources, we can deliver a better experience for customers while supporting future traffic growth.”

What does the UK-first claim establish?

O2 and Nokia present the trial as evidence that AI-driven RAN control has moved beyond experimentation into a live mobile network. The claim rests on the companies' own deployment, and the announcement provides no independent validation or comparison with another UK trial.

It also does not identify the radio band, AI model, Nokia product or standards release associated with the feature. The announcement names no regulatory approval, policy condition or standards deadline, so the UK-first label describes a deployment first claimed by the two companies rather than a formal industry designation.

Emma Falck, President of Mobile Networks at Nokia, said: “This milestone demonstrates how AI is moving from experimentation to delivering measurable benefits in live mobile networks. By embedding intelligence directly into the RAN, operators can make better use of existing spectrum resources, improve network performance and create a stronger foundation for future AI-driven services.

“Our collaboration with O2 shows the significant potential of AI-native networks to enhance customer experiences while improving operational efficiency.”

What happens next?

The announcement sets no date for moving the feature beyond Paddington, identifies no additional sites and does not say whether it will become a broader commercial service. O2's £700m Mobile Transformation Plan supplies the investment context, but its rollout milestones remain unspecified.

Nokia argues that embedding intelligence in the RAN can create a foundation for future AI-driven services. No committed offering or timetable accompanies that statement. With no rollout schedule attached to the £700m programme, the next proof point is whether O2 can reproduce the roughly 20% spectral-efficiency gain at more sites and disclose the customer and throughput data needed to validate its service claims.