[K-Bio on the AI Test Bench]②Galaxy: Creating Proteins Beyond Data Analysis
Building a 'Generative AI' Platform
Accelerating Global Partnerships and Technology Transfer
The axis of competition in Korea's AI-driven drug development is split between 'data observation' and 'protein design.' The Samsung-Protina alliance, which accumulates large-scale experimental data to identify antibody rules, has adopted an 'analytical AI'-centric strategy, while Galaxy, in partnership with Celltrion, is at the forefront of a 'generative AI' approach that directly designs protein structures not found in nature. Both stand at the very forefront of the global AI drug development battlefield. In the midst of this intense competition, we spoke with these two companies, each of which is redrawing the map of Korea's AI new drug industry.
<1> Galaxy, Rewriting the Rules with Direct Protein Structure Design
On March 5 at the Galaxy headquarters in Gwanak-gu, Seoul, Vice President Taeyong Park told The Asia Business Daily, "The competition to design from scratch the protein structures needed for disease treatment—so-called de novo design—will become the core battleground in AI drug development," adding, "Galaxy is taking a leading position in this global race." Galaxy is developing an AI platform that designs candidate drugs based on their interactions with protein structures. When a pharmaceutical company presents a desired function, the AI designs a protein structure capable of performing that function.
Vice President Taeyong Park of Galaxy is being interviewed by The Asia Business Daily at the Galaxy headquarters in Gwanak-gu, Seoul. 2026.3.5 Photo by Hyunmin Kim
View original imageDe novo antibody design is a technology that involves designing protein structures for therapeutic purposes from scratch, fundamentally different from searching for naturally occurring proteins. Currently, only about six companies worldwide, including Galaxy, are considered to have achieved this technology. In 2024, Galaxy announced the results of de novo antibody designs for several therapeutic targets using its proprietary AI platform, GaluxDesign. The company successfully designed antibodies for six targets, including proteins without structural information, thereby demonstrating the platform's applicability.
Vice President Park explained, "De novo antibody design is the most challenging field in antibody drug development," and added, "If you can obtain the desired antibody by designing only a few candidates, it proves the high precision of the platform."
Galaxy's Approach: Not Just Computing, but Physics and Chemistry
Vice President Park cited Galaxy's "physics- and chemistry-based approach" as its core competitive edge. Unlike many AI drug development companies that focus on computer science-based modeling, Galaxy interprets protein binding from the perspective of physical forces—such as charge distribution, three-dimensional structure, and molecular interactions—and incorporates these principles into AI training. "Ultimately, protein binding is determined by physical interactions," he said. "How accurately you understand and model these physical and chemical principles determines the precision of your designs."
There are still technical barriers to 'generative AI-based drug design.' One of the main challenges is the 'flexibility' of protein structures. Proteins do not have a single fixed shape but instead transition between multiple conformations, making antibody design more difficult. Among the industry's most challenging targets are GPCRs and ion channel proteins. These are membrane proteins found in cell membranes, have low structural stability, and are difficult to purify for experiments, making them hard to approach using traditional antibody discovery methods.
The story changes with AI-based protein design. Even without actual protein samples, it is possible to predict binding potential and design antibody structures through structure-based computation. In particular, AI can simultaneously search for bindable structures even in proteins that exist in multiple conformations, enabling the targeting of proteins that are difficult to approach with conventional methods. For this reason, generative AI-based protein design is considered to have a competitive edge for targets that are inaccessible using traditional, experiment-centered antibody discovery techniques.
Vice President Taeyong Park of Galaxy poses for a photo with The Asia Business Daily at Galaxy headquarters in Gwanak-gu, Seoul. 2026.3.5 Photo by Hyunmin Kim
View original imageIn fact, although GPCRs and ion channels are important therapeutic targets in drug development, there are few cases of successful development as antibody therapeutics. Galaxy is also focusing on these highly challenging proteins as key targets. Vice President Park explained, "Proteins like GPCRs and ion channels, which are flexible and exist in multiple states, are difficult to discover antibodies for, but with AI-based design, we can design bindable antibodies based on structural information without actually purifying the proteins."
Galaxy is accelerating its business achievements. Its core strategies include joint research with global pharmaceutical companies and technology transfer of candidate molecules, while also pursuing commercialization with the aim of an initial public offering (IPO) next year. Galaxy believes it is difficult to convince the market with just a technological vision. This is because many AI drug development companies have so far focused solely on highlighting their technological potential.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
Vice President Park said, "While many AI drug development companies have focused only on their technological vision, it is now necessary to show that you can actually create candidate molecules and achieve business results," and added, "We intend to prove the value of generative AI-based drug design through partnership agreements and technology transfers of candidate molecules."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.