[Barcelona, Spain = DigitalToday reporter Jin-ho Lee] LG is accelerating expansion of artificial intelligence (AI) technology and infrastructure under its "One Team LG" strategy. It aims to create synergies by pooling group capabilities to build a comprehensive AI ecosystem spanning infrastructure, platforms and services.

LG held a media briefing in Barcelona, Spain, on March 1 local time, a day ahead of the opening of MWC 2026. It unveiled its AI technology blueprint, including an AI data centre (DC) and EXAONE.

Lee Sang-yeop (이상엽), chief technology officer (CTO) of LG Uplus, said AI that can help in the real world requires powerful infrastructure and connectivity technology. He said One Team LG, including LG Uplus, is preparing South Korea's largest AI data centre (AIDC).

◆Strengthening 'One Team LG' AI capabilities with AIDC and EXAONE

LG is seeking group-wide synergies with AI infrastructure as the محور. LG Uplus' Paju AIDC under construction will have a capacity of 200 MW, the largest in the Seoul metropolitan area, and is scheduled for completion in 2027. It is designed to accommodate up to 120,000 graphics processing units (GPUs) and will serve as an execution base for the "One LG strategy" that brings together core capabilities across LG affiliates including LG Electronics, LG Energy Solution and LG CNS.

LG also plans to present infrastructure that combines EXAONE from LG AI Research, LG Uplus' enterprise AI platform, and neural processing units (NPUs) from FuriosaAI to deliver synergies between K-AI and K-semiconductors.

LG is also speeding up development of its AI model, EXAONE. Lim Woo-hyeong (임우형), co-lead researcher at LG AI Research, said it aims to elevate K-EXAONE, being developed as a national representative AI model, to the global top level.

The goal is to advance K-EXAONE, which has progressed to the second-stage evaluation of the government's "independent AI foundation model" programme, into a global top-level open-weight language model. It will focus first on improving large language model (LLM) performance through the second evaluation, then fully pursue a multimodal expansion strategy including vision language models (VLMs).

LG is pushing EXAONE development on four pillars: securing leadership in AI foundation model development, pursuing expert AI, expanding applications centred on industrial sites, and securing trust and safety for sustainable AI. LG AI Research plans to unveil the EXAONE 4.5 model in the first half of this year.

Lim said the ultimate goal is to create a global top-level AI model. He said it is continuing tailored upgrades for specialised fields.

◆LG Uplus links to physical AI with agentic AI

LG Uplus, which last year put forward "secure AI" as its theme, this year is moving to differentiate AI in telecommunications with "agentic AI" and "voice AI". Lee said a structure of systemic evolution that can solve real problems is important, and securing agentic architecture to implement it is key.

LG Uplus will build a hybrid AI infrastructure architecture to run agentic AI in real time and will work with LG AI Research to accelerate field applications of foundation models. It plans to advance AI services centred on voice, where telecom operators have their strongest advantage, and ultimately expand to physical AI that leads to action.

Lee said LG Uplus will move as one team with LG AI Research and present a standard for agentic AI. He said it will provide experiences that let customers feel "LG's AI really understands me" anytime and anywhere and change the market landscape.