KT said on Tuesday it has launched an "AI dedicated line" supporting bandwidth of up to 800Gbps.

With generative AI spreading and investment in AI infrastructure increasing, the importance is growing of networks that can transmit large volumes of data quickly and reliably between data centres. Demand is also rising for ultra-high-speed, ultra-low-latency networks as AI data centres become distributed nationwide.

KT has prepared services that expand dedicated-line bandwidth from a previous maximum of about 100Gbps to up to 400Gbps and 800Gbps. An advantage is that it can design optimal connection routes between multiple data centres and hubs using its 840,000 km fibre-optic infrastructure and diversified communication paths.

It has applied the All Photonic Network (APN), an optical-centric network technology, to strengthen low-latency performance. APN reduces latency by minimising conversion between optical and electrical signals during data transmission.

KT plans to design data transmission routes and equipment to match customers' network environments and support stable transmission of large volumes of data across distributed AI infrastructure. It will expand service adoption, focusing on AI data centre operators and companies building large-scale AI infrastructure, and will also strengthen ultra-high-speed, ultra-low-latency network services for business-to-business customers.

Jeon Myung-joon (전명준), head of KT's Enterprise Service Headquarters in the Enterprise Division, said, "KT will support customers so they can quickly and flexibly build and expand large-scale AI transformation (AX) infrastructure through the AI dedicated line."