Boltz Closes $110M Series A to Accelerate Drug Discovery With Molecular AI
The pharmaceutical AI startup has secured backing from Bessemer Venture Partners and a consortium of investors to expand its small molecule research platform.

Boltz PBC, which builds artificial intelligence systems for drug researchers, announced a $110 million Series A funding round on Wednesday. Bessemer Venture Partners led the investment, with participation from Greycroft, Andreessen Horowitz, Amplify Partners, F-Prime Capital, Hummingbird and additional returning and new investors.
The startup's core challenge addresses a fundamental bottleneck in pharmaceutical development. Drug discovery requires identifying small molecules—chemical compounds that form the basis of most medications—capable of binding effectively to disease-causing pathogens. Researchers typically must screen hundreds of thousands of candidate molecules per project, a process consuming months and millions of dollars.
Boltz's flagship offering, BoltzMol-1, streamlines this workflow by generating binding affinity predictions across only dozens of the most promising molecular candidates rather than the typical hundreds of thousands. This reduction in required screening dramatically compresses timelines and reduces costs. The model produces data points called small molecule hits—preliminary indicators that guide researchers toward the most productive areas of investigation.
In internal testing, Boltz researchers deployed BoltzMol-1 against 10 therapeutically relevant biological targets, successfully identifying small molecule hits for six of them. The model accommodates project-specific constraints, allowing research teams to configure it for compounds they already have access to or to evaluate a broader candidate pool when material sourcing is not a limitation.
The company also offers BoltzProt-1, a complementary model that generates tens of thousands of protein designs derived from small molecule hit data and automatically assesses each design's viability for drug development. Unlike competing systems that rely on indirect evaluation metrics, BoltzProt-1 employs direct measurement approaches. The company reports that BoltzProt-1 demonstrated "higher success" than several competing models during internal benchmarking.
Boltz distributes its algorithms through an application programming interface priced on token consumption. Since the June launch of BoltzMol-1 and BoltzProt-1, customer token usage has grown 50-fold. The platform counts six of the world's 20 largest pharmaceutical companies among its clients.
The company plans to deploy the new capital toward expanding its team and acquiring additional computing resources. This infrastructure investment aims to keep pace with leading AI research organizations including the recently public Anthropic PBC and OpenAI Group, both of which have released models specifically optimized for biological research applications over the past year.
Boltz is already developing a next-generation suite of molecule design algorithms. According to the company, these systems are "showing substantial improvements" relative to the current generation of models.


