A $110 million Series A led by Bessemer Venture Partners will go into compute, data and hiring at Boltz, the molecular-design company that began as an MIT research project, the firm said on 18 November 2026.
Andreessen Horowitz, which is an existing investor, confirmed the round the same day, congratulating the team “on their $110M Series A” and saying it was proud to “double down.” Neither post discloses a valuation, and TSN is not inferring one. What the two accounts do claim, and what they leave as the company’s own figures, is the rest of the story: Boltz says it has signed partnerships with six of the world’s 20 largest pharmaceutical companies, names Pfizer, GSK, Takeda and BMS among its customers, and says molecular token usage among paying customers has grown 50-fold since June.
Bessemer is leading; existing backers, including a16z, are doubling down
The company’s post puts the use of the money in plain terms: deeper investment in compute and data, capacity to meet demand from partners, and a larger team. The company’s longer-term aim, it says, is “to put better molecular design tools in the hands of all scientists developing the medicines of tomorrow,” and, further out, “a world where a scientist can go from a therapeutic hypothesis to a human-ready molecule without leaving their computer.” That last line is an ambition, not a product date.
Bessemer is leading the round, and the company says Andrew Hedin and Byron Deeter of the firm are joining it on the journey. All existing investors “have doubled down,” the company writes, naming Andreessen Horowitz, Amplify Partners, Zetta Venture Partners and Obvious Ventures. New names on the list are Greycroft, F-Prime Capital, Hummingbird and Benchling, “alongside other new amazing investors.”
Andreessen Horowitz’s note, by Jorge Conde and Zak Doric, adds the earlier chapter. The firm says it saw demand while Boltz “was still a research project at MIT,” and that it co-led the company’s seed round last year. It names the founders as Gabriele Corso, Jeremy Wohlwend and Saro Passaro. The company, for its part, says that in just over a year it has grown from a small group of MIT researchers to a team of 40 scientists and engineers. It dates the company itself to 2025.
Boltz claims six of the world’s 20 largest drugmakers as partners
The commercial claims need to stay attached to their sources. The firm says it launched Boltz Lab in January “to put our latest models in the hands of any scientist or AI agent,” and that in June it released its first proprietary models, BoltzMol-1 and BoltzProt-1. “Today, the platform serves a growing number of pharmaceutical, biotech and industrial companies, including six of the world’s 20 largest pharma companies, among them Pfizer, GSK, Takeda and BMS.” Those four names, and the “six of 20” framing, are the company’s. Andreessen Horowitz repeats the same four customers, writing that Boltz’s “commercial models are already deployed at Pfizer, GSK, Takeda, BMS, and other leading pharmaceutical and multinational companies.” Among the named groups is GSK, the UK-headquartered drugs company.
The 50-fold figure is also a claim, not a public ledger. In its short summary, the company says “molecular token usage among paying customers has grown 50-fold since June.” Later in the same post it says “the volume of molecular tokens processed for our users has grown 50-fold.” Andreessen Horowitz, describing the unit as “the units of molecular information processed by its models,” says that after the June model launch “weekly molecular token consumption by Boltz customers has risen exponentially, roughly 50-fold in just the last 4 months.” The direction is the same; the wording is not identical, and neither side publishes the underlying counts.
That usage-based pitch is the contrast the company wants to draw with the rest of the field. It is also the context in which TSN has been watching other AI drug-design efforts, from UCLA’s MAESTRO system for ranking compounds to pair with a cancer drug to Isomorphic Labs’ $5 billion Series C and the Deep Origin and BioAge collaboration that reported first hits in the lab. Those are different stages of the same broad attempt to move more of discovery onto models. They are not a way to back-solve a price for this company.
The 50-fold token claim is a company figure, not a public ledger
A Series A of $110 million, confirmed by the company and by a well-known existing investor, is a hard fact as far as it goes. It tells readers that Bessemer led, that earlier backers put in more, and that the company wants the cash for compute, data and people. It leaves unsaid what the firm is worth, how much of the new money is already spent, or how durable the token-usage growth will be. The absence of a valuation in both primary posts is itself information: any number attached to this round would be someone else’s guess.
The partner and usage claims should be read in the same spirit. “Six of the world’s 20 largest pharma companies” is a striking phrase, and four global names sit underneath it, one of them British. It is still a company count of “partnerships” and deployments, not a set of independently published contracts. The 50-fold rise, dated from June or “the last 4 months,” begins at the moment Boltz says it shipped its first proprietary models. Starting the clock at a product launch is a reasonable way for a company to describe its own traction; it is also the kind of window that can make a growth rate look its sharpest. Until the company or a customer publishes the raw token volumes, the figure remains an attribute of the announcement.
None of that makes the round uninteresting. A team that, by its own account, was a handful of MIT researchers in 2025 and is 40 people now is raising a nine-figure Series A to sell molecular design as a metered tool rather than as a closed pipeline. That is a distinct bet from a drug company building a proprietary model for its own programmes, and from a university screen that ranks compounds for one disease. The next evidence that would change the picture is not another adjective about the models. It is a customer, or Boltz itself, putting a number on what those tokens have actually produced in the lab, and any later filing that states a price for the shares.
Sources
- Boltz, “Announcing our Series A,” 18 November 2026
- Andreessen Horowitz, Jorge Conde and Zak Doric, “Investing in Boltz,” 18 November 2026
Get TSN’s biggest tech stories, explained.
News and insights across AI, tech, crypto and cybersecurity. Every source linked. Unsubscribe anytime.
Get the newsletterRelated stories
UCLA’s MAESTRO AI Ranked 3,592 Compounds to Pair With a Cancer Drug is the closest earlier TSN neighbour on the science side: another attempt to put models to work on which molecules to try next.
- Isomorphic Labs Announces a $5 Billion Series C Funding Round is the recent large funding comparison in AI molecular design, and a reminder that a round size is not the same thing as a valuation for a different company.
- Deep Origin and BioAge Labs Team Up on AI Drug Discovery, With First Hits Confirmed in the Lab is the nearer example of a claim that has already been tied to laboratory hits, which Boltz has not published here.
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X