LG Group has announced its ambitions to build a domestic AI ecosystem based on its own AI model EXAONE and advance to the top three globally in the future. The group plans to commercialize AI models previously applied primarily to LG affiliates, increasing the number of commercialized models, including next-generation precision medicine and AI agents for employees and businesses. The group also introduces progress in collaboration with major national and international partners, including leading Korean AI semiconductor startup Furiosa AI, suggesting the potential for expansion of its business (B2B) business model from AI-based businesses.
The LG AI research held the LG AI Talk Concert 2025 at LG Science Park Convergence Hall in Magok, Seoul on July 22nd, revealing the blueprint for EXAONE, which encompasses AI models in its services. We have introduced Exaone Data Foundry, ExaONE on-premises, and Exaone 4.0 VL. While LG AI research focuses on proof of the performance of previous models, this presentation highlighted cases of real industry applications and various commercialization scenarios. “If we can lead technology through LG’s own investment and national support, Korea could become one of the “top three” AI countries in the world,” said Lee Hwa-Young, head of AI business development.
Exaone Data Foundry is a platform technology that acts as an AI factory that generates the high quality data needed to build enterprise exclusive AI. It helps 60 experts that take three months to produce and complete it in one day.
The demonstration projects were carried out by LG affiliates and national institutions, and data productivity has increased by at least 1,000 times, and data quality has increased by an average of over 20% compared to existing methods. EXAONE ON-PREMISE is an AI solution that can be installed on internal enterprise systems and operate independently without external connections. Starting today, Chatexaone has launched beta services for businesses, public institutions and research institutions.
The company also highlighted the results of collaborations with global partners such as Furiosa AI, Friendly AI and Rondon Stock Exchange Group (LSEG) during the development process. EXAONE On-Premises was implemented by combining Furiosa AI’s latest neural machining unit (NPU) “Renegade” with EXAONE. Emphasizing three times more power efficient than Nvidia’s graphics processing units (GPUs) with the same performance, he also expressed his intention to challenge Nvidia’s advantage with AI solutions made with domestic technology.
With LSEG, we plan to jointly develop and launch AI services in the third quarter to help investors make decisions.
“We’ve got a lot of effort to bring you the Collaboration results,” said Baek Jun-Ho, CEO of Furiosa AI, who directly introduced Collaboration results.
Lim Woo-Hyung, co-head of LG AI research, said “If we ignore the development of internalized technologies while the latest AI models are being poured into the world, we are facing a crisis of subordinateship,” adding that he is “creating AI models that can compete in global markets beyond technological independence.”
LG AI Research plans to establish a variety of B2B business scenarios through discussions with affiliates such as LG CNS and LG Uplus. The order amount has already exceeded 100 billion won, and new businesses are being discussed in the biosector centered around EXAONE Pass.
“We will accelerate the development of physical AI technologies that recognize reality, make judgments and change the real world,” explains Lee Hong-Rak, co-head of LG AI research, and “we are primarily looking at physical AI that supports robots, factory equipment, manufacturing, and discuss with affiliates.”
It was also expressed that building infrastructure is urgent to build a domestic AI ecosystem. LG AI Research recently joined the consortium of the “Independent AI Foundation Model Development Project” promoted by the government. “If the country supports two things, it would be useful if it supports two things: data-related GPUs and infrastructure, including deregulation,” Prime Minister Lee added, “When you set up an infrastructure, domestic companies can freely create and commercialize AI-based applications and services.”
