1910
1910 provides a multimodal AI platform for modality-agnostic drug discovery that integrates massive biological datasets, frontier AI models, and high-throughput robotics-driven lab automation. The ITO™ platform generates up to 100 billion proxy assay data points per week and includes specialized models like PEGASUS™ for macrocyclic peptides and CANDID-CNS™ for blood-brain barrier penetration prediction.
- Drug Discovery & Therapeutics
- Artificial Intelligence
- Biotechnology
- Robotics
Funding
Funding not disclosed
Founders
Founder details are not available yet.
Product
Problem
Pharmaceutical R&D productivity has declined steadily for decades, with developing a single drug costing $2.5B+ and taking 12–15 years, with less than 10% probability of success from Phase I clinical trials. Data scarcity, AI model training limitations, computing power constraints, and insufficient wet lab validation prevent AI from achieving its full potential in drug discovery.
Solution
1910 offers the ITO™ platform, a multimodal AI system that integrates massive diverse datasets, frontier AI models, and high-throughput robotics-driven lab automation to design small and large molecule therapeutics faster than traditional approaches. The platform combines wet lab ground truth data (in vitro, in vivo, and clinical), ultra-high-throughput proxy biological assays capable of generating up to 100 billion data points per week, and computational simulations including quantum mechanics and molecular dynamics. A Multi-AI Agent System with hundreds of models addresses the multi-parameter optimization problem inherent to drug discovery, while federated learning enables consortium partners to train models on private data without sharing it. The platform outputs in silico molecules with suggested synthetic routes and validates candidates through automated wet lab assays before advancing to clinical testing.
Target Audience
Primary customers are pharmaceutical and biotechnology companies seeking AI-driven drug discovery capabilities, including those interested in co-engineering bespoke small or large molecule discovery modules and participating in federated learning consortia.
Features
- PEGASUS™ AI model unlocks macrocyclic peptide drugs, a previously challenging therapeutic modality
- CANDID-CNS™ model predicts CNS penetration for small molecules, outperforming industry standards for blood-brain barrier permeability
- Robotics-driven automated lab with end-to-end infrastructure that puts biological datasets, AI models, and lab automation in the same iterative loop
- Federated learning technology enabling pre-competitive consortia to fine-tune models on blinded datasets without private data leaving partner data centers
- Proxy assay technologies (mRNA display, droplet microfluidics, phage display) generating up to 100,000,000,000 data points per week versus tens from traditional gold standard assays
- Multi-GPU compute infrastructure deployed through partnerships with Microsoft Azure and NVIDIA for low-latency, high-throughput workflows