Domestic pharmaceutical and biotechnology companies are increasingly developing their own artificial intelligence (AI) platforms and engaging in global collaborative research. This strategy aims to enhance their competitiveness in new drug development by combining proprietary data and research capabilities with AI-driven candidate substance discovery. However, the industry is still in its early stages, as few companies have secured sufficient specialized personnel and high-quality research data for AI-driven drug development.
According to industry sources, Hanmi Pharmaceutical has designed an obesity drug candidate, HM17321, using its proprietary AI platform, HARP-pSAR. HM17321 aims to reduce weight while maintaining or increasing muscle mass and is currently undergoing Phase 1 clinical trials in the United States.
The company reports that AI can predict pharmacological effects based on subtle changes in protein structures and suggest optimal design directions, enabling ultra-fast screening with minimal experimental data. Recently, Hanmi signed an exclusive licensing agreement with Genentech, a subsidiary of Roche Group, for HM17321, with a deal value of up to $2.3 billion (approximately 3.2 trillion won).
The trend of pharmaceutical companies building their own platforms is on the rise. JW Pharmaceutical operates an AI drug development platform called J-Wave, which integrates its existing drug discovery systems, Jewel and Clover. The platform aims to reduce drug development time and costs by over 25-50% by analyzing more than 400 genomic datasets and over 45,000 compound data.
The company plans to utilize its proprietary cell lines, organoids, disease animal models, and synthetic compound data for AI training to discover and optimize new drug candidates.
Daewoong Pharmaceutical has created a database of molecular information for 800 million compounds, which serves as the foundation for its AI drug development system, Daisy. Meanwhile, Dong-A Socio Group established its own AI drug development platform, AIDDP, in July in collaboration with its IT subsidiary DAI and Dong-A ST. AIDDP is designed to perform tasks such as new compound design, predicting binding affinity between candidate substances and target proteins, molecular simulations, and managing research results within a single environment.
Additionally, SK Biopharm entered into a joint research and development agreement with Insilico Medicine in June for drug candidates in the central nervous system (CNS) and neuroimmunology fields. The two companies plan to conduct joint research targeting three CNS-related targets using Insilico's AI platform, with a total contract value of up to $2.5725 billion (approximately 3.58 trillion won) depending on research, development, regulatory, and commercialization outcomes.
The competition among AI drug development companies is also intensifying. Pharos AI Bio is using its Chemiverse platform to discover new drug candidates, while OncoCross focuses on drug repurposing based on its RaptorAI platform to find new indications for existing drugs. Syntekabio is advancing large-scale virtual screening technology, and Galax is leveraging protein and antibody design technology to stay competitive.
However, some industry experts point out that while announcements of AI platform development and implementation are increasing, there is a lack of institutional support to back these initiatives. The shortage of specialized personnel is also a significant concern. A survey conducted by the Korea Pharmaceutical and Bio Association's AI Drug Research Institute among 55 respondents from domestic pharmaceutical and biotech companies, AI firms, academia, and research institutions found that 89.1% are either utilizing or considering the use of AI in drug development. However, 67.3% reported having no AI specialists in their drug development departments. While many companies are interested in adopting AI, there is a shortage of personnel to apply it in actual research.
Industry insiders emphasize that AI platforms should not be approached merely as software implementations. An AI drug development expert stated, "High-quality experimental and clinical data organized by disease, researchers capable of validating these results, and organizations that can connect AI outcomes to actual candidate substances are all necessary simultaneously." They added, "In the future, the amount of data accumulated and the success of clinical transitions will determine competitiveness among companies more than just having a platform."
Amid this backdrop, a domestically developed bio AI model called K-Fold has emerged, which predicts not only the three-dimensional structure of proteins but also how drug candidates will bind to proteins. The Korea Pharmaceutical and Bio Association plans to promote the industrial application of K-Fold through its AI Drug Research Institute.
Noh Yeon-hong, president of the Korea Pharmaceutical and Bio Association, remarked, "The development of K-Fold signifies our capability to secure world-class core technology beyond merely utilizing foreign bio AI technologies. It is expected to lower the entry barriers for researchers to utilize AI and enhance the speed and efficiency of new drug candidate discovery."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.