KT and Samsung report 29% cut in high-speed rail call drops

KT said on the 5th that a trial with Samsung Electronics cut the communication-dropout rate by about 29% on South Korea’s Gyeongbu High-Speed Railway—equivalent to 100 dropouts before the change and about 71 after it.

The reported result came from a demonstration between Godeok-myeon in Pyeongtaek and Maesong-myeon in Hwaseong. KT, the telecommunications carrier, and Samsung Electronics, the technology developer, ran the trial on that section.

Why did calls fail above 250 km/h?

On trains traveling above 250 km/h, signals from successive base stations change rapidly. The mobile connection must transfer between them more frequently. When handover timing slips, data transmission delays and call quality deteriorate, which matters more as passengers watch video or make calls in transit.

Until the trial, engineers had addressed persistent trouble spots by inspecting individual base stations through on-site measurements. KT and Samsung instead used AI to analyze signal-strength data and connection-failure records collected along high-speed rail sections.

The analysis identified delayed handovers as the main cause of in-cabin dropouts. On sections where the fault repeatedly appeared, the companies demonstrated a change that starts the connection to the next base station earlier. The earlier transition shortens the period in which a moving train can lose service.

What did the trial actually measure?

KT reported a before-and-after comparison of communication-quality data. It found a dropout rate about 29% lower after the handover adjustment on the Gyeongbu demonstration section. The company expressed the change as 100 dropouts reduced to about 71.

The figure applies only to the tested section and conditions captured in the data. The announcement does not disclose:

• the baseline dropout percentage;
• train or passenger-journey counts;
• the test duration or any confidence range;
• results on a second high-speed rail section;
• a regulator, approval process or commercial rollout date.

The result therefore cannot yet support a network-wide service or cost assumption.

What did the executives claim?

Jung Jin-kook, Samsung Electronics executive vice president and head of Samsung Research’s Advanced Communications Research Center, said: “We have proven that we can improve the communication quality customers experience in high-mobility environments.”

Lee Jong-sik, a managing director who heads KT’s Future Network Lab in its Network Division, called it “a meaningful achievement that confirms the potential to enhance customer-perceived quality using AI in vulnerable areas such as high-speed rail.”

The engineering data supports the narrower conclusion that earlier handovers reduced dropouts in the recorded trial. The executives’ language reaches customer-perceived quality more broadly, but the announcement gives no figures for completed calls, throughput, complaint rates or costs.

Can AI-RAN deliver the gain without new equipment?

KT and Samsung present AI-based radio access network, or AI-RAN, as a way to raise customer-perceived quality through software and data analysis alone, without adding separate equipment. For KT, that is the claimed cost advantage: service quality can improve without separate equipment additions. The announcement reports no installation or operating cost, so any saving remains unquantified.

KT and Samsung cite growing demand for reliable communications, with passengers watching video during journeys as a driver. Fewer dropouts would mean more continuous calls and data sessions, while the software-led approach could limit the equipment burden. The companies frame AI-RAN as a new competitive axis for advancing 5G and 6G networks.

The announcement does not define whether “separate equipment” means new base stations, onboard devices or both. That ambiguity matters when checking the claim against fleet plans: a network-side change carries different deployment requirements from an upgrade to train-borne equipment. KT and Samsung have not stated which hardware, if any, each rollout would require.

What happens next?

KT plans to consider how to apply the technology secured in the demonstration to actual high-speed rail sections. The companies also agreed to continue joint development and verification in AI-RAN.

Neither side has named candidate routes, rollout dates, capital budgets or the number of trains and base stations involved. Before wider use, the 29% figure needs checking against train movements, speed profiles and network logs rather than extrapolation from one corridor.

Until KT publishes that evidence, the 29% result should stand as a measured improvement on one South Korean test section, not a network-wide forecast. KT’s route-by-route assessment and joint verification will determine whether early handover becomes standard on high-speed rail and a broader competitive tool for 5G and 6G.