AgentAI Launches Next-Generation AI Drug Discovery Platform 'MolForge'... Six Candidate Compounds Enter Validation in Europe

KOSDAQ-listed AgentAI announced on April 30 that its subsidiary, AgentAI Labs, has officially launched full-scale operations of its next-generation, AI-based drug discovery platform "MolForge" and has initiated external validation procedures for six key candidate compounds through a European contract research organization (CRO).

MolForge is a platform that integrates over ten of the latest AI drug discovery engines released by leading global research institutions such as MIT, AstraZeneca, and Stanford University between 2025 and 2026 into a single pipeline. Unlike conventional drug development companies that rely on a single AI model, this platform employs a "5-way Consensus" structure, allowing multiple independent AI engines to sequentially evaluate a single molecule. This design significantly reduces potential errors that can arise from using only one model.

The platform includes Boltz-2, co-developed by MIT and Recursion Pharmaceuticals, as well as tools such as AiZynthFinder 4.0 and ADMET-AI v2.0, covering the entire process from molecular structure prediction and synthetic accessibility to toxicity validation in an integrated manner.

Another distinctive feature of MolForge is its approach of having AI directly design molecules optimized for target proteins, rather than selecting candidates from existing compound libraries. AgentAI Labs has identified a total of 72,387 candidate compounds focusing on major targets such as autoimmune diseases (TYK2), cancer (CDK4/6), obesity, and fibrosis (TNIK). Of these, 1,974 were selected as core candidates. In particular, for the TNIK target, the company has secured a new chemical space with low similarity to existing drugs, thereby lowering the risk of patent conflicts.

Currently, the company is conducting synthesis and biochemical experiments for a total of six candidate compounds—three for TYK2 and three for TNIK—at laboratories under Eurofins Group, Europe's largest CRO. After securing preclinical binding affinity data in the second half of 2026, AgentAI plans to actively pursue global licensing-out (L/O) agreements. In addition, the company is enhancing technical sophistication by establishing a system that automatically recalibrates its AI prediction models based on experimental results.

The company has also differentiated itself in terms of efficiency. By using its proprietary "ROBOGATE" algorithm, it has implemented a structure that can deliver optimal results with only 12% of total computational resources. Furthermore, by recording and releasing data on error domains within AI predictions, AgentAI has raised trust levels for collaboration with global pharmaceutical companies.

An AgentAI representative stated, "MolForge is a platform that integratively manages various global AI engines, dramatically shortening both the time and cost required for drug development. By publicly sharing not only successes but also limitations through a data-driven approach, we will continue to strengthen our global competitiveness."

Meanwhile, the government is supporting the AI drug industry as a key growth sector by increasing the budget for the medical AI field to 247.8 billion won in 2026, 2.5 times more than the previous year. AgentAI Labs plans to strengthen its technological and intellectual property competitiveness by filing a provisional patent in the United States and registering with additional global leaderboards within the first half of the year.

In summary, AgentAI is accelerating its entry into the global AI drug discovery market by simultaneously advancing candidate validation and technology transfer through its multi-AI engine-based drug discovery platform, MolForge.

This content was produced with the assistance of AI translation services.