Global telecom equipment makers are putting their full weight behind competition in artificial intelligence (AI) network technology, driven by the shift to 5G standalone (SA) mode. As South Korea's three mobile carriers begin building next-generation infrastructure, major equipment makers are moving to secure an early lead in the market.

According to the industry on Feb. 27, major global telecom equipment makers have set "networks for the AI era" as a core theme. They aim to create new markets through faster 5G and AI-powered network technology.

5G SA in particular is emerging as essential infrastructure for the AI era. Unlike the existing non-standalone (NSA) approach, 5G SA uses a 5G-dedicated core rather than relying on a 4G LTE core network. As AI services spread, uplink traffic is increasing as large volumes of data are sent to servers in real time. 5G SA has an edge over NSA in ultra-low latency, network slicing and high-speed data transmission, allowing it to handle more traffic effectively.

Earlier, the Ministry of Science and ICT made conversion to 5G SA mandatory when it announced reallocation conditions for a total of 370 MHz of 3G and LTE spectrum. For equipment makers, that amounts to a new market opening up. Base station technology using AI is also nearing full-scale commercialisation.

Ericsson recently unveiled a roadmap at a press briefing in South Korea combining 5G SA and 6G with AI-based autonomous networks. As AI traffic rises, Ericsson sees it as difficult for existing network technology to meet required performance, and stressed that redesigning infrastructure by combining 5G SA, 6G and autonomous network technology is essential. Ericsson is targeting the market with new technology, including introducing a neural network accelerator into its in-house silicon Isaac (ASIC) and supporting 10 times more AI computing capability.

AI-integrated base station technology is also drawing attention. A leading example is AI-RAN. AI-RAN is a next-generation base station technology that provides communications and AI services simultaneously on a single piece of equipment. Beyond simply processing communications traffic, it uses AI computing to reallocate radio network resources efficiently. That can improve base station performance and support an evolution toward autonomous networks in which the network identifies its own status and carries out optimisation and recovery.

Nokia recently succeeded in an AI-RAN proof-of-concept with SK Telecom. It developed an approach that simultaneously uses a graphics processing unit (GPU) to handle AI services and a communications-dedicated accelerator to handle parts of communications functions. With KT, it validated AI-based beam pattern optimisation technology. Beam pattern optimisation analyses various network data in real time and actively reconstructs base station signals using AI.

Nokia also validated "cloud RAN" technology with LG Uplus. Cloud RAN implements software functions previously installed on dedicated hardware for each base station on a central server virtualised in a cloud environment, securing both equipment efficiency and scalability.

Huawei is also pushing AI-based solutions. Huawei has developed next-generation network solutions in which AI fixes failures on its own and reduces energy use. It plans to secure an early lead in the autonomous network market through a so-called "communications foundation model". Huawei plans to showcase these next-generation solutions at MWC26 next month.

The industry says these moves by telecom equipment makers also need to be viewed as transitional competition ahead of 6G. "Companies that secure an early lead in the 5G SA and AI network market can also take the initiative in the 6G era," an industry official said. "In South Korea, whether 5G SA market leadership is secured will be a variable that will shape the future market landscape," the official said.