1910 Genetics applies AI to drug discovery across both small molecules and protein therapeutics, an unusual combination since most computational platforms specialise in one modality. The company’s argument is that modality should be an output of the discovery process rather than an assumption made before the biology is understood: for a given target, the right answer may be a small molecule, a protein, or neither, and a platform locked to one chemistry cannot make that call honestly.

Making that work requires wet-lab capability alongside computation, so that model predictions feed automated experiments and the resulting data improves the models, rather than relying on public datasets collected for other purposes. Backed since 2020 and working in partnership with Microsoft, 1910 Genetics is an agentic applications investment in a domain where the loop between prediction and physical validation matters more than model architecture. The measure of the platform is whether its programs advance, not how it benchmarks. Partnership validation from a company like Microsoft also lowers the cost of proving the platform to pharmaceutical buyers.