Skip to content

Funding Alert|Boltz Raises $110M Series A (Bessemer leads, $138M total raised)

Nov 20, 20261 min
TL;DRBoltz raised a $110M Series A led by Bessemer Venture Partners, bringing total funding to $138M. The signal: AI agents are expanding beyond text and code into scientific discovery that requires real-world experimental validation.

Look up & save

Loading dictionary…

中文版

Funding details

ItemValue
CompanyBoltz (MIT spinout; research team now based in London's King's Cross)
RoundSeries A
Amount$110M
LeadBessemer Venture Partners
ParticipantsAndreessen Horowitz, Amplify Partners, Zetta Venture Partners, Obvious Ventures (existing investors); Greycroft, F-Prime Capital, Hummingbird, Benchling (new); Nvidia invested earlier in the year
ValuationUndisclosed
Total raised$138M (Seed $28M → Series A $110M)
Founded2025
HeadcountAbout 40 (The Pharma Letter)

What the company does

Boltz builds AI molecular-design models — predicting how proteins and candidate drug molecules bind, estimating binding strength, and proposing new candidate molecules, helping pharma and biotech researchers decide which candidates are worth testing in the lab.

The core technology originated at MIT, where the co-founders released open-source models (Boltz-1, Boltz-2) that matched the performance of DeepMind's AlphaFold3 — now a default tool for many drug-discovery teams. Rather than running its own drug pipeline, Boltz sells computational capacity through "molecular tokens," priced by usage, while offering both open-source models and its own platform API. Agent features launched in early 2026 let scientists run agentic workflows for tasks like de novo protein design and small-molecule hit discovery, handing results off to wet-lab partners for validation.

Boltz has announced partnerships with Pfizer, GSK, Takeda and Bristol Myers Squibb — including a multi-year partnership with Pfizer that gives its entire scientist base access to Boltz Lab and its agents. The company says AI usage has grown 50-fold over the past year, though it hasn't disclosed revenue figures.

What this funding signals

What it means for the agent ecosystem

This round is as much a signal about where agents are headed as it is about molecular-design models: AI agents are expanding from text and code into scientific discovery that requires physical-world validation — not just searching literature or predicting structures, but designing experiments, handing them off to real labs, and learning from the results to refine the next design cycle.

What investors are betting on

Bessemer is leading this round, while seed-stage investors Andreessen Horowitz and Amplify Partners all returned to add more — suggesting these investors are betting less on a single valuation jump and more on whether the "open-source models set the technical standard → commercial platform monetizes" path can keep compounding. Nvidia also invested earlier this year during Jensen Huang's UK visit, signaling that compute providers themselves are betting AI-driven drug discovery will be the next compute-hungry customer segment.

Boltz's two funding rounds compared
Boltz's two funding rounds comparedBoltz started with a $28M seed round in early 2026; less than a year later, its Series A reached $110M — nearly 4x growth. Round size ($M) Seed $28M (early 2026) Series A $110M (2026-11)

Round size grew nearly 4x in under a year. Conversion: k≈3.18px/$M. Figures compiled from Tech.eu (2026-11-19) and European Biotechnology's reporting on the official announcement.

Numbers worth watching

  • Round size jumped from $28M to $110M in under a year — nearly 4x — while valuation remains undisclosed, unlike most AI companies in this cycle that readily publish it.
  • The company's self-reported 50x growth in AI usage has no corresponding revenue or customer-count figure to back it up — a single self-reported metric.
  • A team of only about 40 people has already landed partnerships with 4 of the top 20 pharma companies — a notably high partnership-per-headcount ratio for the sector.

Watchlist status

Boltz is not yet on the watchlist. Recommend adding it under a new section (AI for Science / drug-discovery agents), tracking: agentic workflows for protein design and small-molecule discovery, deep partnerships with Pfizer and other major pharma companies, Series A led by Bessemer.

Today's takeaway

Boltz's business-model choice is worth noting: instead of running its own drug pipeline to collect milestone payments and royalties like most AI drug-discovery companies, it sells "compute plus models" — a more infrastructure-like business. That suggests AI drug discovery is splitting into two different business models — "who has the best model" versus "who actually gets the drug made" — and Boltz is betting the former scales faster.

References