Analysis
As of November 19, 2026, Boltz PBC has raised a $110 million Series A led by Bessemer, SiliconANGLE reported, with Greycroft, Andreessen Horowitz, Amplify Partners, F-Prime Capital, Hummingbird and others in the round. Boltz builds AI models that help pharmaceutical researchers find and design small-molecule drugs, and it will spend the money on hiring and on more computing infrastructure.
What does Boltz do?
Boltz started as academic work at MIT. Its founders, Gabriele Corso, Jeremy Wohlwend and Saro Passaro, released open-source models for predicting how molecules bind to proteins, and Andreessen Horowitz wrote that pharma partners kept telling its team they all used "the same open-source AI models out of MIT." By the time a16z co-led the seed, those open-source models had reached every top-20 pharmaceutical company, per the firm.
The commercial products launched in June. BoltzMol-1 scores binding affinity, so a chemist reviews a few dozen promising molecules instead of hundreds of thousands. In an internal test across 10 targets, BoltzMol-1 found small-molecule hits for six, SiliconANGLE reported, and researchers can restrict it to compounds they can actually buy. BoltzProt-1 generates tens of thousands of protein designs from small-molecule hit data and screens each one; Boltz says it showed "higher success" than several competing models in internal tests, a claim from the company's own benchmarks rather than an outside study.
“BoltzMol-1 scores binding affinity, so a chemist reviews a few dozen promising molecules instead of hundreds of thousands.”
Boltz sells access through an API billed on token usage, and customer token consumption has grown about 50-fold since the models launched in June, per SiliconANGLE, which also reported that Boltz counts six of the world's 20 largest pharmaceutical companies as clients. A16z puts the same figure as weekly molecular token use up "roughly 50-fold in just the last 4 months," and says the commercial models are deployed at Pfizer, GSK, Takeda and BMS. StartupHub reported that Boltz also offers Boltz Lab for biologists and chemists, and that it partnered with Excelra to train ADMET models on curated GOSTAR data, per a November 10 announcement.
Boltz funding history
The sources we opened confirm two rounds. Andreessen Horowitz co-led a seed round last year; none of the pages we opened state its amount. The $110 million Series A is the second. No source we opened states a Boltz valuation, so we do not report one.
Who invested in Boltz?
Bessemer led the Series A, per SiliconANGLE. Greycroft, Andreessen Horowitz, Amplify Partners, F-Prime Capital and Hummingbird also invested. A16z says it is "doubling down" after co-leading the seed, so an insider is putting in more money at the Series A.
AI drug discovery funding and competitors
Boltz is raising into a crowded and expensive field. Two recent AI drug-design rounds on Pulse:
- Isomorphic Labs -- $5B Series C, 2.4 times its $2.1B May round (Pulse, Nov 17)
- Chai Discovery -- $400M Series C (Pulse)
Next to those, a $110 million Series A is small. StartupHub notes that "frontier labs are chasing the same prize with far larger compute budgets."
Boltz's model is different. Isomorphic runs proprietary pipelines. Boltz built its customer base on open-source models and now charges for frontier versions by usage, so its revenue grows with how much pharma researchers query it, not with how many partnership milestones it hits.
Value Add VC data
- 2025, Seed -- amount not stated in the sources we opened, co-led by Andreessen Horowitz (a16z)
- Nov 2026, Series A -- $110M, valuation not disclosed, led by Bessemer (SiliconANGLE)
Comparable deals: Isomorphic Labs, $5B Series C (Nov 2026); Chai Discovery, $400M Series C. With no disclosed valuation, we cannot give a markup multiple.
The risk in the token model
The 50-fold growth comes from a small base that started in June, and Boltz has disclosed no revenue figure. Token usage shows how much researchers use the models; it does not show that Boltz-designed molecules work in trials. A hit in an internal test on six of 10 targets is still years from a clinical readout. Open source cuts both ways too. The same openness that put Boltz inside every top-20 pharma company lets a well-funded rival, or a pharma company's in-house team, build on the same published methods. The Series A has to turn Boltz's open-source lead into a paid product before competitors with more compute catch up.