-
2026.07.08
LG Showcases EXAONE's Real-World Industry Impact and Latest AI Research at ICML 2026
LG unveiled its latest AI research achievements and demonstrated how its flagship AI model, EXAONE, is driving tangible innovation across industries at ICML 2026, one of the world's premier conferences on machine learning and artificial intelligence.
Held in Korea for the first time, ICML is widely recognized as one of the world's top three AI conferences, bringing together leading researchers, developers, and industry pioneers to showcase the latest advances in machine learning and artificial intelligence.
_8d57d9861.png)
■ EXAONE Demonstrates Innovation Across Materials Science, Finance, and Data Intelligence
At the conference, LG AI Research showcased how EXAONE is creating real-world value across a wide range of industries, from materials discovery and financial intelligence to AI-powered data engineering.
LG AI Research presented a live demonstration of EXAONE Discovery, its AI-powered materials discovery platform, alongside two tangible innovations enabled by AI: Rhamsydil, a novel ingredient for hair-loss care currently being prepared for commercialization, and an immersion cooling fluid developed for next-generation AI data centers.
Rhamsydil was discovered by LG Household & Health Care in collaboration with LG AI Research, using AI to screen more than 420,000 candidate compounds in just one day. The ingredient has demonstrated hair-loss prevention efficacy without steroid-derived compounds, with the research presented at the World Congress for Hair Research. It is currently being prepared for commercialization.
The immersion cooling fluid for next-generation AI data centers was jointly developed by GS Caltex and LG AI Research. Building on this achievement, the two organizations plan to further expand their collaboration in AI-driven materials discovery.
EXAONE Discovery is powered by the core technologies behind LG AI Research's AI Co-Scientist, a platform designed to accelerate the discovery of advanced materials and drug candidates.
_19120c621.png)
Earlier this year, LG AI Research secured patents covering an end-to-end AI-driven R&D workflow that automatically interprets molecular structures from scientific literature and designs new materials, further strengthening its technological leadership in AI-enabled scientific discovery.
The technology is already being applied across a broad range of industries, including hair-care ingredients, immersion cooling fluids, cosmetics, battery materials, semiconductor materials, and pharmaceutical research.
In finance, LG AI Research demonstrated EXAONE Business Intelligence (EXAONE BI), its AI-powered financial intelligence agent, which is expanding from global markets into the Korean market.
EXAONE BI analyzes approximately 8,000 publicly listed companies across the Korean and U.S. stock markets each day, providing predictive scores and institutional-grade analytical commentary to help investors make more informed decisions.
Earlier this year, LG AI Research officially launched its U.S. equity forecasting service in partnership with the London Stock Exchange Group (LSEG). This week, the company further expanded its financial AI business by signing an agreement with KOSCOM to launch AI-powered equity forecasting services for the Korean stock market.
Finally, LG AI Research demonstrated EXAONE Data Foundry, its AI-powered data factory platform, which automatically generates high-quality datasets and builds domain-specific AI models.
EXAONE Data Foundry has increased data productivity by more than 1,000 times while improving data quality by an average of over 20%. In a pilot project with the National Pension Service of Korea, the platform successfully automated the creation of more than 10,000 domain-specific data records per day.
"EXAONE has already moved beyond the experimental stage and is delivering meaningful results across real-world industries," said an LG AI Research spokesperson. "We will continue identifying new opportunities where AI can help solve complex industrial challenges."
■ LG AI Research Presents 14 Papers at ICML 2026, with AI Materials Generation Technology Ranked No. 2 Globally
_237efa131.png)
LG AI Research demonstrated its research leadership at ICML 2026 by presenting 14 technical papers spanning a wide range of AI disciplines.
Among the highlights, its AI-powered materials generation technology ranked second globally on LeMat-GenBench, a leading benchmark for evaluating generative AI models for materials discovery.
LeMat-GenBench rigorously evaluates AI-generated crystal structures based on their physical stability, novelty compared with existing materials, and diversity of candidate generation, providing a comprehensive assessment of an AI model's capability in materials discovery.
"Our results demonstrate that LG's AI is among the world's best at discovering practical new materials capable of addressing real-world industrial challenges," said Sehee Han, Head of the Materials Intelligence (MI) Lab at LG AI Research.
Since its establishment in December 2020, LG AI Research has published 363 papers at the world's leading AI conferences, including NeurIPS, ICLR, CVPR, AAAI, and ACL. The institute has also built a strong innovation portfolio with 838 patent applications filed to date, including 371 domestic patents, 243 international patents, and 224 PCT applications, reinforcing its ability to translate cutting-edge research into real-world industrial impact.
"EXAONE has evolved beyond a laboratory AI into an expert AI that delivers solutions in real-world industrial environments," said Woohyung Lim, Co-President of LG AI Research. "Through our first-mover strategy for solving complex industrial challenges, we will continue strengthening LG's leadership in the global AI landscape."
■ LG AI Day Brings Together Global AI Talent to Share LG's AI Vision
LG AI Research is also accelerating its efforts to attract and cultivate world-class AI talent.
On July 7, the institute hosted LG AI Day, inviting more than 100 master's and Ph.D. students attending ICML 2026 to a networking event.
Senior researchers, including Honglak Lee and Woohyung Lim, Co-Presidents of LG AI Research, shared the latest progress in LG's AI technologies, outlined the company's talent development strategy, and engaged participants in discussions about future research directions.
"Leadership in AI will ultimately be determined not by technology alone, but by people," said Honglak Lee, Co-President of LG AI Research. "We are committed to providing researchers with world-class computing infrastructure, high-quality data, and mentorship, enabling LG AI Research to become a global hub for AI innovation."
LG continues to invest in developing next-generation AI talent through initiatives including the LG Graduate School of AI—Korea's first officially accredited in-house graduate school—expanded joint research programs with leading global universities, and industry-focused internship opportunities.
By fostering a virtuous cycle connecting research, industry, and talent, LG aims to strengthen its leadership in AI while advancing its vision of becoming a global hub for AI innovation.