Japan's pharmaceutical industry is making a coordinated push into AI-driven drug discovery, with major companies deploying advanced AI platforms to accelerate research and streamline molecular design. The Tokyo-1 AI drug discovery consortium, operated by Xeureka, has expanded to include Eisai alongside existing members Astellas, Daiichi Sankyo, and Ono Pharmaceuticals, all leveraging NVIDIA's BioNeMo platform to reshape how medicines are discovered.

How Are Japanese Pharma Companies Using AI to Speed Up Drug Discovery?

Each member of the Tokyo-1 consortium is deploying different AI tools tailored to their research needs. Astellas has integrated nearly all BioNeMo microservices, which are specialized software components that handle specific life sciences tasks, and is running the BioNeMo Agent Toolkit, an open platform that turns AI agents into autonomous life sciences researchers. This toolkit gives AI systems immediate access to NVIDIA's full life sciences stack, enabling them to work independently on discovery tasks.

Ono Pharmaceuticals is using the Boltz-2 microservice to streamline internal drug discovery workflows. Daiichi Sankyo is conducting ultralarge-scale virtual screening on the Tokyo-1 platform and leveraging NVIDIA RAPIDS, a data processing framework, to accelerate analysis of massive datasets. Xeureka itself is using BioNeMo to power its AI-driven efforts, giving researchers flexibility to choose the most appropriate models and tools for their specific discovery programs.

What New AI Models Are Emerging From This Ecosystem?

Beyond the consortium members, Japanese AI companies are developing breakthrough molecular models. SyntheticGestalt announced two products designed to transform how researchers approach drug design. ZAO is a foundation model that converts small molecules into data that AI can understand through a "4D" representation, capturing the multiple three-dimensional shapes a molecule actually adopts in biological systems. On nine public drug-discovery benchmark tasks, ZAO ranked first, achieving the world's best performance.

KOYA, SyntheticGestalt's second product, is a molecular generative model that designs novel, high-affinity ligands for target proteins while closely reflecting what researchers intend. Both products integrate with the BioNeMo Agent Toolkit, enabling AI agents to evaluate and design molecules while collaborating with human researchers to accelerate discovery.

Biomy is pioneering a different approach with a virtual cell foundation model built on a massive clinical dataset from the Japanese Foundation for Cancer Research. Using NVIDIA single-cell RAPIDS, Biomy achieved 90% faster spatial transcriptomics analysis, a technique that maps gene expression across tissue samples. Biomy plans to use NVIDIA Nemotron-powered agents to autonomously propose and orchestrate complex virtual experiments for drug development.

Steps to Integrate AI Into Pharmaceutical Research Workflows

• Assess Current Data Infrastructure: Evaluate whether your organization's data pipelines can handle large-scale molecular datasets and whether you have the computing resources to run AI models efficiently.
• Select Appropriate AI Models: Choose foundation models and generative tools that match your specific discovery needs, whether that's molecular design, protein structure prediction, or virtual screening.
• Deploy Autonomous Agent Systems: Implement AI agent toolkits that can work independently on routine discovery tasks while maintaining integration with existing research workflows and human oversight.
• Establish Cross-Team Collaboration: Create workflows where AI agents, software platforms, and human researchers work together, with clear handoff points between automated analysis and human decision-making.

Takeda, another major Japanese pharmaceutical company, recently announced a collaboration with Boltz to deploy the BoltzMol-1 and BoltzProt-1 biomolecular models across its research organization. These models provide scientists with tools for structure prediction, affinity estimation, and generative design that integrate seamlessly into existing discovery workflows.

The convergence of these efforts signals a fundamental shift in how Japanese pharmaceutical research operates. Rather than treating AI as an experimental add-on, companies are embedding it into the core infrastructure of drug discovery. The consortium model allows competing companies to share a common platform while maintaining proprietary research, reducing duplication and accelerating the pace of innovation across the industry.

This coordinated approach reflects a broader recognition that AI drug discovery requires not just powerful models, but integrated ecosystems where molecular design, virtual screening, protein structure prediction, and data analysis work together seamlessly. By uniting around shared platforms like Tokyo-1 and BioNeMo, Japan's pharmaceutical leaders are positioning themselves at the forefront of a global shift toward AI-accelerated medicine development.