[Edaily Reporter Kim Seung-kwon] #Pharos iBio, a South Korean company specializing in AI-driven drug discovery, has joined forces with global pharmaceutical giant Eli Lilly. Pharos iBio has officially signed a strategic cooperation agreement with Lilly Tune Lab, an AI-driven drug discovery federated learning platform. This is considered the first such partnership for a South Korean AI drug discovery company.

Linking

a $1 billion big data and NVIDIA platform… Leaping forward as a global partner following in the footsteps of Schrödinger

Lilly TuneLab is a platform officially launched last September by Eli Lilly to connect its big data from new drug development research—built through investments of over $1 billion (approximately 1.5 trillion won) over the past few decades—with AI and machine learning (ML) models, linking them with global innovative companies. Schrödinger, a leading global AI-driven drug discovery company, is among the participants in this project.

Lilly TuneLab is at the heart of this partnership. As the first pharmaceutical-led global AI federated learning infrastructure, Lilly TuneLab was created to connect the vast big data from new drug development research and AI models—accumulated through Lilly’s massive investment—with external innovative companies.

A key highlight is the strong technological partnership with NVIDIA. LilyTune Lab applies NVIDIA’s FLARE-based federated learning technology to safely share and refine only algorithmic models and predictive results, eliminating concerns about the leakage of proprietary data or copyright infringement for each company. In the future, NVIDIA Clara-based open foundation models for life sciences are also scheduled to be integrated into the workflow. Pharos iBio can now access world-class computing infrastructure at virtually no cost.

Pharos iBio will actively leverage this powerful infrastructure to advance its proprietary Chemiverse platform. The company will integrate the practical pipeline discovery and clinical development experience accumulated through Chemiverse—from target identification to candidate compound derivation—with LillyTune Lab’s advanced AI and ML prediction environment.

In particular, by cross-referencing Eli Lilly’s vast real-world validation data to comprehensively evaluate drug properties and development potential—such as determining whether early-stage compounds possess tolerability suitable for humans—the company will be able to exponentially reduce clinical trial trial-and-error. This is viewed as going beyond simply verifying the accuracy of external AI models.

This is because it serves as a starting point to instantly overcome the limitations faced by small and medium-sized biotech companies—which would otherwise require trillions of won to build large-scale AI infrastructure—and to have the success potential and development feasibility of their own pipelines validated against global standards within the AI and ML collaboration framework of a global big pharma.

Kim Kyu-tae, President of Business Development at Pharos iBio, stated in a recent interview with Pharm iDaily, the premium pharmaceutical and biotech content platform of Edaily, “This agreement with LillyTune Lab is a crucial starting point for connecting the candidate discovery and clinical development know-how accumulated through Chemiverse with the AI and ML collaboration framework of a global big pharma.”

Biomarker-based clinical trials now fully feasible… Green light for global technology transfer

This collaboration is noteworthy not only for securing data but also because it represents a commercialization pipeline that could lead to future investment, joint research, and technology transfer.

Moon Sung-won, Executive Vice President (CFO) of Pharos iBio, explained, “The main point is that Eli Lilly intends to prioritize exploring opportunities for investment, joint R&D, and technology transfer,” adding, “It is a mutual selection.”

Among the compounds currently leading the commercialization effort, Rasmotinib (PHI-101), a treatment for acute myeloid leukemia (AML), is being targeted for global conditional approval. Rasmotinib overcomes the issues of frequent relapse and mutation resistance associated with Astellas’s competing drug, Josapta, through Chemivers’s precision design. In Phase 1b clinical trials, 60% of evaluated patients achieved complete remission (CR/CRi), demonstrating Rasmotinib’s value as a best-in-class new drug.

In addition, Pharos iBio is accelerating the expansion of its indications and pipeline through open innovation. PHI-501, which targets solid tumors such as colorectal cancer, is actively exploring synergistic clinical trials when used in combination with the global blockbuster Keytruda (MSD).

Furthermore, the company recently achieved the distinction of being selected for the Structure-Based AI New Drug Development Support Project, led by the Ministry of Health and Welfare. Through this initiative, Pharos iBio has established a robust domestic and international collaboration framework by formally launching joint research with Kolon Pharmaceutical on the next-generation EGFR mutation inhibitor (PHI-701).

President Kim Kyu-tae emphasized, “Eli Lilly has its AI platform, and we have Chemiverse. The concept of federated learning is that these two platforms communicate with each other to enhance each other’s quality.” He added, “We do not share the raw data itself; we only share the results. The rights to any compounds derived from these results belong to Pharos iBio, and Lilly plans to review the results and prioritize whether to pursue future commercialization.”

Executive Vice President Moon Sung-won also explained, “This can be viewed as a proposal from Eli Lilly: they will encrypt Pharos iBio’s data and conduct research on it using their platform, with all compounds derived from this process belonging to Pharos iBio.” He added, “The key point is that Eli Lilly will only have access to the results and, based on those, will prioritize exploring opportunities for investment, joint research, and technology transfer.”