Alnylam and Inceptive Partner in $2B Deal to Advance AI-Driven siRNA Drug Design
Alnylam Pharmaceuticals has announced a multi-year collaboration with Inceptive, an AI company focused on nucleic acid design, in a deal that could exceed $2 billion in value. Under the agreement, Alnylam will gain access to Inceptive's proprietary AI tools to accelerate the discovery and optimization of siRNA (small interfering RNA) drug candidates.
Why This Matters
siRNA therapeutics work by targeting specific genes to silence disease-causing proteins, a modality where Alnylam is a leading developer. However, designing effective siRNA molecules with the right sequence, chemistry, and delivery properties is complex and time-consuming. Inceptive applies deep learning to model and predict nucleic acid behavior, potentially reducing the trial-and-error traditionally involved in early-stage drug design.
Inceptive was founded by Jakub Pistолю, who previously co-invented the Transformer architecture at Google Brain and went on to lead deep learning research at Google's Isomorphic Labs. His team brings expertise in high-throughput experimentation combined with AI-driven sequence modeling—a combination the biotech industry has begun to embrace more broadly.
Deal Structure
Financial details include an upfront payment and near-term milestones, with additional development and commercial milestones that could bring the total deal value above $2 billion. Alnylam will retain rights to develop and commercialize any resulting candidates, while Inceptive will contribute its computational platform and expertise.
Broader Context
This collaboration is the latest example of AI companies and established biopharma firms joining forces in the competitive landscape of genetic medicine. Over the past several years, major players including Eli Lilly, Roche, and AstraZeneca have signed similar deals with AI-native biotech companies. The integration of AI into siRNA design specifically targets a step that has historically been a bottleneck in therapeutic development.