AI Finds Hair-Loss Compound in a Day as LG Showcases EXAONE Innovations

LG (003550.KS) unveiled industrial innovation cases for its proprietary AI model "EXAONE" at "ICML 2026," the world's most prestigious conference in machine learning and artificial intelligence (AI). The company presented AI applications leading to actual commercialization in new materials, finance and data, drawing assessments that it is accelerating its "field-based expert AI" strategy.

LG AI Research said Wednesday that it participated in the International Conference on Machine Learning (ICML) 2026, held at COEX in Seoul from Wednesday through the 11th, to introduce EXAONE-based AI technology and its industrial application results. ICML is one of the world's top three AI conferences, where the latest research trends in machine learning and artificial intelligence can be examined, and was held in Korea for the first time this year.

The area LG AI Research emphasized most at the event was the discovery of new materials. The institute presented a real-time demonstration of "EXAONE Discovery," an AI platform for discovering new materials, and exhibited actual samples of "Rhamsydil," a new hair-loss management material discovered by AI, and an "immersion cooling oil material" for next-generation AI data centers.

EXAONE Discovery is a platform that applies AI scientist technology supporting research into new materials and new drugs. The AI independently recognizes molecular structures from literature and designs candidate compounds with desired performance. LG AI Research completed patent registration for this entire research and development process earlier this year, also building a technological barrier.

The flagship case, Rhamsydil, is a hair-loss management material jointly discovered by LG Household & Health Care and LG AI Research. It is a compound the AI found in a single day among more than 420,000 candidate substances, characterized by demonstrating hair-loss prevention effects without steroid-derived ingredients. LG presented this achievement at the World Congress for Hair Research and is currently preparing for product commercialization.

An immersion cooling oil material jointly developed with GS Caltex was also unveiled. It is a next-generation cooling solution to solve the heat problem of AI data centers. The two companies plan to expand the scope of their cooperation in discovering new materials in the future.

In finance, the finance-specialized AI agent "EXAONE Business Intelligence (EXAONE BI)" was introduced. EXAONE BI is a service that analyzes about 8,000 stocks listed on the Korean and U.S. markets every day to provide prediction scores and analytical commentary at the level of financial experts. After launching a U.S. stock market prediction AI service with the London Stock Exchange Group (LSEG) of the U.K. earlier this year, LG AI Research recently signed a contract with Koscom to enter the domestic stock market prediction AI service business.

In the data field, "EXAONE Data Foundry" was unveiled. This platform is a technology that generates high-quality data with AI and automatically builds AI models specialized for professional fields. According to LG AI Research, it produced results that raise data productivity by at least 1,000 times and improve quality by an average of more than 20%. In a pilot project conducted with the National Pension Service, it implemented a system that automatically builds more than 10,000 pieces of professional data per day.

Research achievements were also disclosed. LG AI Research presented a total of 14 papers at this year's ICML 2026. In particular, its new-material generation AI technology ranked second in the world in overall standing on the global benchmark "LeMat-GenBench." This evaluation is a performance metric for new-material development that comprehensively verifies the stability of crystal structures generated by AI, their differentiation from existing materials, and the diversity of candidate compound proposals.

Since its launch in December 2020, LG AI Research has presented a total of 363 papers at the world's top AI conferences. It has also filed a total of 838 patents, including 371 domestic, 243 overseas and 224 international (PCT). The company explained that it is strengthening its technology commercialization capabilities to apply research achievements to actual industrial sites rather than leaving them confined to papers.

Lim Woo-hyung, co-president of LG AI Research, said, "EXAONE is no longer an expert AI that finds answers in the laboratory but in the field," adding, "We will seize global AI leadership through a first-mover strategy to solve industrial challenges."

Meanwhile, LG is also accelerating efforts to secure global AI talent. On Tuesday, LG AI Research invited about 60 master's- and doctoral-level AI talents who attended ICML 2026 to the "LG AI Day" networking event. There, key researchers including co-presidents Lee Hong-lak and Lim Woo-hyung shared LG's status of AI technology development and its talent development plans.

Lee Hong-lak, co-president of LG AI Research, said, "AI leadership comes from the talents who create the best technology," adding, "We will build a system that can provide researchers with the best experimental environments, data and mentoring to leap forward as the world's top research hub."