AbbVie, Iambic in multi-year AI drug discovery pact

Small molecule drug discovery has traditionally required scientists to assess each candidate against dozens of properties one at a time. AbbVie and Iambic say their collaboration aims to change that by applying Iambic's platform to optimize the full range of candidate properties simultaneously, which they hope will reduce timelines and increase the chance of making better molecules.

Iambic's molecular superintelligence platform is powered by Enchant and NeuralPLexer, and the collaboration will use Enchant v3, described as a next-generation multimodal transformer model that combines biomedical data modalities and enables many drug design hypotheses to be tested in parallel and at scale. It is trained on over 6,000 molecular properties.

The partnership's emphasis on first-in-class and best-in-class potential suggests the partners are betting on differentiated molecules rather than fast-following. Whether that translates into better candidates hinges on validation in the clinic and on how well the AI predictions hold up against real-world biology — a bar the collaboration has set for itself.