Samsung and telecom operator KT have proven AI-driven RAN (AI‑RAN) optimisation on a live commercial network.
The two companies demonstrated that AI can preserve consistent service quality in real-world conditions, even as users move through areas with fluctuating signal quality. This live validation follows a simulation phase completed in June and marks the first time this particular AI‑RAN architecture has been verified outside of a laboratory environment.
As mobile data demand grows, applying a single configuration to entire cell sectors is increasingly inadequate. The trial indicates future networks will adapt to individual devices and their context, rather than requiring devices to adapt to a one-size-fits-all network configuration.
AI‑RAN optimisation: from cell‑centric to user‑centric networks
Traditional networks use broad parameters. A base station typically enforces the same settings for every connected device, whether a user is stationary in a strong-signal area or moving quickly through interference. That rigidity can cause dropped connections when conditions change.
The AI‑RAN model tested by Samsung and KT reverses that approach. Instead of optimising for an average user, the system applies configurations tailored to each device based on real‑time wireless conditions.
This process uses predictive analytics. AI learns from a user’s mobility patterns and usage history to identify recurring circumstances that lead to connectivity degradation. By forecasting likely disruptions, the network can adjust radio parameters proactively to avoid service interruptions before they occur.
Field validation results
To validate the AI‑RAN architecture under real conditions, Samsung Research and KT’s Future Network Laboratory deployed the solution in Seongnam, Gyeonggi Province. The trial covered roughly 18,000 users and concentrated on areas with varied environmental and mobility conditions.
The trial specifically targeted users who had experienced frequent service interruptions. Applying AI‑RAN to those devices enabled the system to detect the patterns causing failures and apply customised radio configurations.
Comparison of performance metrics before and after the deployment showed a reduction in connection failures among the targeted users. The results also showed fewer issues for other users in the same cells, indicating that resolving edge‑case problems can improve overall cell performance.
Samsung and KT lay groundwork for 6G
Although 5G is still being rolled out, this successful AI‑RAN validation helps prepare networks for 6G. Future generations will need to handle far greater data volumes and more complex mobility without increasing failure rates, and predictive, software‑defined mechanisms like AI‑RAN will be central to that effort.
“This is a major milestone that shows how AI can improve user experience on real‑world commercial networks,” said JinGuk Jeong, Executive VP and Head of the Advanced Communications Research Center at Samsung Research. “Through close collaboration with KT, we will continue shaping and validating the next generation of AI‑driven communication technologies.”
For KT, the trial demonstrates that network operations can become more adaptive. “This achievement demonstrates that AI can transform network operations around users,” said Jong‑Sik Lee, Executive VP and Head of the Future Network Laboratory at KT. “Together with Samsung, we will continue developing customised optimisation technologies to deliver stable, uninterrupted service and drive core 6G capabilities that create new value for customers.”
The commercial validation of AI‑RAN by Samsung and KT shows wireless infrastructure is shifting toward software‑defined, predictive operation.
Enterprises responsible for mobile workforces or distributed IoT assets can expect improved service-level agreements as operators adopt dynamic optimisation like AI‑RAN. When selecting connectivity providers, assess their roadmap for AI-integrated radio access networks—this capability will influence reliability in high‑mobility and weak‑signal scenarios.
See also: Global connectivity slows as enterprise infrastructure gaps widen
Interested in learning more about IoT from industry leaders? The IoT Tech Expo runs in Amsterdam, California, and London as part of TechEx and is co‑located with other major technology events.
Telecoms coverage is produced by TechForge Media. Explore upcoming enterprise technology events and webinars through event organisers and industry channels.