The pioneer of RNA interference, Alnylam Pharmaceuticals, just committed up to $2B to put a specialized AI foundation model, and the talent behind it, to work on its pipeline.

Yesterday, Alnylam and Inceptive announced a strategic collaboration valued at up to $2B, with $30M upfront in cash and equity. The remainder is tied to preclinical, regulatory, and commercial sales milestones.

Speaking about science, Inceptive contributes foundation models for sequence-based medicines and its AI team, led by CEO Jakob Uszkoreit, a co-inventor of the Transformer architecture. Alnylam contributes its RNAi platform, six approved drugs, more than 20 years of proprietary siRNA data, and its clinical and regulatory operations.

The collaboration is specifically targets siRNA design.

This is a quite revolutionary modality I have to say. siRNAs (small interfering RNAs) is a class of double-stranded non-coding RNA that regulates gene expression by "silencing" specific target genes. By degrading messenger RNA (mRNA) before it can be translated into proteins, siRNA stops disease-causing or problematic genes from functioning.

The companies will model target mRNAs, explore sequence space and chemical modifications to improve potency, and predict candidates for Alnylam to advance through preclinical development.

In joint exploratory work, Inceptive's model characterized siRNA molecules from relatively small datasets within weeks, as the release says.

On a broader context, Inceptive, founded in 2021, is based in Palo Alto with offices in Berlin and Zurich, and is backed by a16z, NVIDIA, S32, and Obvious. Alnylam is a publicly traded biotech that has medicines available in more than 70 countries.