Boltz raises $110 million Series A led by Bessemer for AI drug-design models
London-based Boltz raised a $110M Series A led by Bessemer, bringing disclosed funding to $138M, to expand usage-priced AI molecular design models.
Boltz raises $110 million Series A led by Bessemer for AI drug-design models
Boltz, a London company, raised a $110 million Series A led by Bessemer Venture Partners. Andreessen Horowitz, Amplify Partners, Zetta Venture Partners and Obvious Ventures returned as investors. Greycroft, F-Prime Capital, Hummingbird and Benchling joined as new investors. Total disclosed funding is now $138 million, which includes a $28 million seed round unveiled in January. Boltz says it will spend on computing infrastructure, data, model development and hiring.
Boltz started with open-source models, Boltz-1 and Boltz-2, that predict molecular structures, interactions and binding affinities. In June it launched two proprietary models: BoltzMol-1 for small molecules and BoltzProt-1 for proteins, including antibody-like binders. It sells access through usage-based fees counted in molecular tokens, not through stakes in individual drug programs. The company says it works with six of the world's 20 largest pharmaceutical companies and names Pfizer, GSK, Takeda and Bristol Myers Squibb. Paid-customer token usage has risen 50-fold since June, which measures work processed, not drug candidates. In June experiments, Boltz reported confirmed hits against six of 10 challenging targets, testing 28 to 51 compounds per target. These are early research leads, not evidence of safety or efficacy.
For investors, the round tests whether usage-based pricing for molecular design can scale across large pharma buyers. Per the AI Market Watch index, which covers about 5,000 companies and is not a census, Boltz was listed with $28.0M in total funding before this round. For builders in drug discovery, the API and the Boltz Lab platform offer a way to run design experiments without building models in-house. Treat the usage growth and hit rates as company-reported until outside validation appears.

