LG unveils K-EXAONE global push at MWC 2026

LG AI Research Co-President Lim Woo-hyung, right, and LG Uplus Chief Technology Officer Lee Sang-yeob hold a press conference in Barcelona, Spain, Sunday. Joint Press Corps

LG Group unveiled a road map Sunday (local time) in Barcelona to elevate its artificial intelligence (AI) model K-EXAONE into a globally competitive platform by pairing advanced model development with large-scale infrastructure.

The announcement, made by LG AI Research Co-President Lim Woo-hyung and LG Uplus Chief Technology Officer Lee Sang-yeob, came a day ahead of the opening of Mobile World Congress 2026.

Lim emphasized that LG’s AI vision goes beyond building the most powerful model.

“The AI that LG pursues is not about competing in the height of intelligence, but about creating a partner that helps people and solves problems in the real world,” Lim said. “What ultimately matters is what we create with AI. We will focus on delivering tangible value in physical spaces.”

LG outlined four priorities for K-EXAONE: securing leadership in AI foundation models, developing expert-level AI, expanding adoption across industrial sites and strengthening trust and safety.

During the second phase of the national AI foundation model project in the first half of this year, LG aims to develop what it calls the world’s top-performing language model among global open-weight models.

Infrastructure investment will proceed in parallel.

“To implement AI that can truly support the real world, robust infrastructure and connectivity are essential,” Lee said. “As K-EXAONE reaches completion next year, One Team LG is preparing the largest AI data center in the Seoul metropolitan area.”

LG revealed plans for a 200-megawatt AI data center in Paju, Gyeonggi Province, scheduled for completion next year. The facility will house up to 120,000 graphics processing units and serve as a strategic hub integrating the capabilities of LG Electronics, LG Energy Solution and LG CNS. LG Uplus plans to provide end-to-end customized services based on K-EXAONE and the new center.

LG AI Research will also unveil EXAONE 4.5, a vision-language model capable of understanding text and images simultaneously. Lim noted that LG developed EXAONE 1.0 in 2021 as Korea’s first multimodal AI model.

EXAONE 4.5 will be released as an open-weight model and is expected to serve as the brain of Korea’s humanoid project, KAPEX, positioning LG in the emerging physical AI era, where intelligence is directly embedded into machines and production systems.

LG Uplus also introduced its agentic AI strategy, built around a cycle of planning, execution, evaluation and revision. The company highlighted four core pillars: self-evolving systems, a model-data foundry, trust-based integrated control and hybrid AI infrastructure.

“In the AI era, competitiveness is not just about bigger models or more computing power,” Lee said. “The key will be designing architectures that evolve through continuous feedback loops.”

He added that close collaboration between LG Uplus and LG AI Research will help set a new standard for personalized, trustworthy AI services.