Boltz: building towards programmable molecular design
Bessemer Venture Partners leads Boltz's $110M Series A to back the team behind the state-of-the-art protein and small molecule foundation models for AI drug discovery.
Andrew Hedin, Byron Deeter, Marla Jalbut, M.D., Sam Bondy & Grace Dai
Published on Nov 18, 2026
Two years ago, a small team of researchers out of MIT’s CSAIL and Jameel Clinic released a suite of open-source models. Matching the performance of their closed-source counterparts like AlphaFold3, these models have now become part of the default toolset for early-stage drug discovery and are used by more than 100,000 scientists across thousands of biotechs, labs, and top-20 pharma companies.
These models are just the beginning of a much bigger vision to make drug design a programmable problem. The feedback cycle of drug development has traditionally run through physical experimentation and trial-and-error, bounded by human reasoning. With AI-driven property prediction and design, scientific reasoning capabilities can drastically expand, allowing researchers to unlock previously undruggable targets and generate novel therapies at a fraction of the historical time and cost.
Boltz is driving this paradigm shift with their foundation models for biology and chemistry, and Bessemer is excited to announce that we’re leading their $110M Series A.
The team setting the standard in end-to-end molecular design and discovery
Boltz is a London-based AI research lab building foundation models for biomolecular modeling, small-molecule and protein design. The three co-founders were among the core developers of the open-source models that came out of Professor Regina Barzilay's lab. CEO Gabriele Corso won the Best Student Paper Award at the NeurIPS 2022 SBM workshop for DiffDock, one of the first models to treat molecular docking as a generative problem and predict how a small molecule binds to a protein. CTO Jeremy Wohlwend led Boltz-1 for structure prediction. Head of Research Saro Passaro built Boltz-2 for joint structure and affinity co-folding. And last but not least, Boltz founding scientist Hannes Stärk led the development of BoltzGen, which is now one of the most widely used models for generative protein design.
What stood out to us from the start is how much scientists trusted the models and the team long before Boltz spun out of MIT. Their open-source models have over 5 million downloads, and every partnership Boltz has established to date began as inbound interest from computational chemists and biologists who were excited to share the model’s capabilities with the rest of their teams.
Scaling open-source models into a platform
Since launching in January, the pace of Boltz’s progress has been remarkable. In June, they launched BoltzMol-1, their small-molecule design model, which cuts the time to a validated hit from months to weeks and cost from hundreds of thousands of dollars to as low as $10,000. The team also launched BoltzProt-1, their protein design model, which has nearly triple the hit rate of BoltzGen, outperforms other leading models head-to-head, and produces designs that pass strict criteria around developability, demonstrating their comparability to real therapeutics in the clinic.
From day one, Boltz has built the platform to make models accessible and easy-to-use for scientists. Beyond its open-source offerings, the entire model suite shares a single lineage, letting scientists move from a structure to a binding prediction to a designed molecule in one workflow. Additionally, the company’s Boltz API, hosted inference offerings, fine-tuning services, and integrations with agents like Claude Code, Codex, and Gemini allow scientists to run Boltz inside their existing and increasingly productive AI workbenches. This ecosystem of tools is currently deployed at a number of top-20 pharma companies including Pfizer, Takeda, GSK, and Bristol Myers Squibb, along with DSM Firmenich, one of the largest fragrance and ingredient companies in the world.
Backing foundational AI with Boltz
In software, AI tools like Cursor and Claude Code have become the default for where code is written. We think a similar phenomenon will occur in preclinical R&D, where AI tools like Boltz will move a significant amount of screening and design work from the benchtop to the computer. We’re seeing this happen today, as computational leads have described Boltz as indispensable infrastructure they run thousands of daily predictions on. These anecdotes align closely with our thesis on how AI is reshaping drug discovery from early research through clinical trials, as well as how molecule identification and structure design sit in the high adoption/high value quadrant of our pharma ROI scorecard.
As these models continue to improve, the ROI they deliver will accelerate, and with it the design of novel medicines and enhanced chemicals that foster healthier people and a healthier planet. We could not be more thrilled to partner with Gabriele, Jeremy, Saro, and the rest of the Boltz team, the people who pioneered this field, as they continue to build and innovate on the infrastructure for the next generation of scientific discovery.
Learn more about Boltz or join their team.
Contributors

Andrew Hedin
Partner
Andrew Hedin is a partner in Bessemer’s Boston office where he focuses on investments broadly across the healthcare ecosystem, including new biotech therapeutics as well as software and services sold to healthcare verticals.
Prior to joining Bessemer in 2015, Andrew worked at F-Prime Capital, Fidelity’s healthcare-focused venture capital fund where he invested in early-stage biotech and digital health technologies, as well as Leerink Partners as an advisor to the biopharma industry.
He earned an MBA with honors from The Wharton School, where he majored in health care management and finance, as well as a degree in biological basis of behavior from the University of Pennsylvania.

Byron Deeter
Partner
Byron Deeter is a leading investor in AI, Cloud, Frontier Technology, and the Business of Sports. He co-authored Bessemer’s iconic AI 100, the Bessemer Forbes Cloud 100, the BVP Nasdaq Emerging Cloud Index, and Bessemer’s STRIVE program for executive health and wellness. Byron works closely alongside many of the best founders in the AI and cloud worlds, with 26 of Byron’s investments currently valued above $1 billion each, including 13 IPOs and counting.
Byron first raised a Series A with Bessemer back in 2000, as CEO and founder of Trigo Technologies. His company grew to be one of the first global SaaS companies, reaching profitability and was successfully sold to IBM. Byron foresaw that cloud computing would not only change the way people built new technologies but also how The Cloud would systematically run the world. In 2005, Byron returned to Bessemer, this time to help lead the firm’s global cloud practice.
Byron graduated with honors from The University of California, Berkeley, where he met his then college sweetheart and now wife. Byron is a perennial Midas List investor, past Chairman of the National Venture Capital Association, partner/owner of the San Francisco 49ers and Leeds United, as well as an active board member/advisor to numerous causes including the U.S. Olympic & Paralympic Foundation, U.C. Berkeley Foundation, Pledge 1% and Cal Rugby. Although he’s a four-time Rugby Collegiate National Champion and serial Ironman finisher, he’s largely converted to leisure sports of golf, wake surfing, and skiing with his wife and three children.

Marla Jalbut, M.D.
Investor
Dr. Marla Jalbut is an investor in Bessemer’s Boston office, where she focuses on biotech and healthcare investments.
In addition to her role at Bessemer, Marla serves as an attending physician in the Department of Medicine at Massachusetts General Hospital and is on the faculty at Harvard Medical School. She continues to see patients in the inpatient setting while teaching and mentoring medical trainees.
Before joining Bessemer, Marla was an investor at Omega Funds, a biotech-focused investment firm based in Boston. Her work spanned early-stage company creation to late clinical-stage biotech companies, and she served as a Board Observer at Aviceda Therapeutics, Ousia Pharma, and Rectify Pharma.
Marla earned an A.B. from Princeton University, an M.D. from Tufts University, and an M.B.A. from Harvard Business School. She completed her clinical training in internal medicine at Yale.
Outside of work, Marla enjoys spending time with friends and family, exploring new restaurants (she’s a retired food blogger), attending concerts, and cheering on the Baltimore Ravens.

Sam Bondy
Vice President
Sam is a vice president based out of San Francisco and primarily focuses on early and growth stage cloud and AI.
Prior to venture capital, Sam began his career in Technology, Media, and Telecommunications investment banking at J.P. Morgan where he focused on M&A and capital markets transactions for both private and public companies. He then joined Harvest Partners, a private equity firm in New York, focusing on software and business services buyout. Most recently, Sam moved into earlier stage investing where he worked at Soma Capital primarily focusing on software and fintech.
Sam grew up in New York City and first attended McGill University in Montreal, before graduating summa cum laude from SUNY Binghamton where he received a B.A. in Economics. Outside of work, he is an aspiring outdoorsman, a (very) novice golfer, skier, surfer, and loves to run, read, travel, and listen to podcasts.

Grace Dai
Investor
Grace Dai is an investor in Bessemer’s New York City office.
Prior to joining Bessemer full time, Grace led Critical VC’s University of Pennsylvania team, where she oversaw a group of undergraduate and MBA students investing in startups across health care, climate and economic mobility. She was also a summer analyst at Bessemer.
Grace holds a bachelor’s degree and a master’s degree in computational biology from the University of Pennsylvania. She also led the data analytics team at the Daily Pennsylvanian, the podcast team for the Penn Biotech Society, and founded the university’s Alternative Protein Project in 2023 under The Good Food Institute.
Outside of work, she enjoys hiking, diving, fermenting foods and making jewelry.



