"Reviewing 40 Million Substances in a Single Day"... LG AI Research Registers Patent for New Materials and Drug Development

Unveiling Solutions at Europe's "ISE 2026"
Showcasing in a 1,184㎡ Store-Like Space
Utilizing Display Operation and Management Platform

LG AI Research is accelerating its efforts to secure technological leadership in AI-based new materials and drug development by registering patents for its core technologies.

On February 3, LG AI Research announced that it has recently completed the patent registration for "ExaOne Discovery," the core technology behind its "AI Co-Scientist" that supports new materials and drug development. ExaOne Discovery is an AI-based platform for new materials and drug development, designed to analyze various forms of multimodal data-including research papers, patents, molecular structures, and images-to identify promising candidate substances at speeds dozens of times faster than conventional methods. This, the company explained, enables a dramatic reduction in research and development timelines.

ExaOne Discovery is characterized by its ability to directly select molecular structures embedded in unstructured documents or to derive experimental results through conversations using labels and tags such as "m3" or "m5." Researchers do not need to manually input complex chemical formulas or structures; the AI interprets documents and data to support experimental design and prediction.

Actual operating screen of ExaOne Discovery. LG Corporation.

LG AI Research emphasized that this patent allows it to protect the entire research and development process for new substances as intellectual property, thereby solidifying its technological leadership. The patent specifies in its claims the entire sequence of methods and systems, including extracting molecular structures from unstructured documents, assigning numbers, predicting specific labels based on researchers’ queries, designing experiments, and predicting new substances.

As a result, it is considered a "blocking patent" that cannot be easily circumvented by simple algorithm improvements. Unlike conventional AI patents that focus solely on formulas or algorithms, this patent is differentiated by protecting the entire research process-from data analysis to experimental design and new substance prediction.

LG explained that even if competitors develop AI models with similar performance, they would still face limitations, such as requiring researchers to manually input or connect molecular structures and chemical formulas. Therefore, it would be difficult to match the speed and convenience offered by ExaOne Discovery. Consequently, any system that enables researchers to use an AI co-scientist to make natural language queries and receive results would require a patent licensing agreement with LG AI Research.

LG is applying ExaOne Discovery to various industries, including cosmetics materials, battery materials, and drug development. A representative example is the development of cosmetic ingredients, which requires repeated cycles of structural design, laboratory synthesis, and property testing. By utilizing ExaOne Discovery, LG explained that it was able to complete the review process-which previously took 22 months-for more than 40 million synthesized substances in just one day, checking whether the resulting materials met the required properties for cosmetics, were easy to synthesize, and did not generate harmful substances.

LG plans to further develop ExaOne Discovery into a leading "chemical agentic AI" that will play a key role in discovering new substances capable of transforming industries such as batteries, semiconductors, and pharmaceuticals. A representative from LG AI Research stated, "This patent is a prime example of LG not only delivering innovation that surpasses existing technologies, but also establishing an exclusive rights barrier to protect those innovations."