Boltz PBC 獲 2800 萬美元融資，致力於推動 AI 藥物發現平台的普及化

Boltz PBC Launches with $28M to Democratize AI Platforms for Drug Discovery

MIT researchers are turning the prolific Boltz series of models into an AI research and product company that promotes open science for AI-guided therapeutics

By

        Fay Lin, PhD

In November 2024, Boltz-1 debuted as a fully commercially available AI model to achieve AlphaFold 3-level accuracy in predicting the 3D structure of biomolecular complexes. Since that time, the Massachusetts Institute of Technology (MIT) research team has continued to push a new standard in open-source AI-based drug development.

In less than 18 months, the Boltz series of models have achieved accurate binding affinity prediction (Boltz-2) and therapeutic design across a wide array of drug modalities (BoltzGen) with more than 100,000 scientists across thousands of biotechs implementing Boltz to accelerate discovery.

Led by freshly minted PhD graduates, Gabriele Corso, PhD, and Jeremy Wohlwend, PhD, and research scientist, Saro Passaro, the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) team originates from the lab of Regina Barzilay, PhD, distinguished professor of AI and Health.

The trio has now co-founded Boltz, as a public benefit corporation (PBC), turning their prolific technology into a business centered on open science. Corso will lead the company as CEO.

Boltz launches with a $28 million seed round led by Amplify, a16z, and Zetta Venture Partners, alongside angel investors, including Clement Delangue, CEO of Hugging Face, Factorial Capital, and Obvious Ventures.

Dylan Reid, managing director at Zetta, highlights that Zetta’s investment focus lies in AI native applications, infrastructure, and developer tools. The conventional path for a therapeutics company, which involves raising hundreds of millions of dollars to build a pipeline for a one shot bet on a breakthrough drug, seemed “super disconnected” from how technology and tools evolve.

“When we looked at the first generation of AI-driven drug discovery companies, we had a lot of conviction that the technology would be the real difference maker,” Reid told GEN. “The best tools are the ones that are widely used.”

Rather than design drugs, Boltz will build AI-based infrastructures that “allows every scientist to go from a therapeutic hypothesis to a human-ready molecule without leaving their computer,” the team wrote in the company’s manifesto.

Aligned with the launch, Boltz has unveiled a preview of their platform, Boltz Lab, along with their first agents for small-molecule discovery and protein design. The platform is built to remove the practical bottlenecks prominent in adoption, including compute cost, scalable infrastructure, and intuitive, collaborative interfaces with scientific workflows.

Corso says the conviction that Boltz needed to become a company emerged soon after work began on Boltz-1, with two realizations driving the decision.

“First, continuing to push the frontier of biomolecular AI requires sustained investment in talent, compute, and data at a scale not attainable within academic environments,” he told GEN. “Second, truly democratizing the technology, and maximizing its impact, means going beyond publishing models and building reliable, well-designed products that scientists could integrate directly into their daily work.”

Boltz has already solidified a multi-year collaboration with Pfizer. Under the agreement, Boltz will use Pfizer’s data to build exclusive models for structure prediction, small-molecule affinity, and biologics design. Boltz scientists will also work closely with Pfizer’s discovery teams to advance custom models and workflows for a number of target programs to enhance preclinical decision making.

When asked why open science-oriented companies are still uncommon in AI for biology, Corso points to cultural inheritance, where founders carry forward the norms and incentives of the environments in which they were trained, as a bottleneck. He notes that many startups spin out of pharmaceutical companies or wet-lab biology groups where value is tightly coupled to assets and the default approach is to patent discoveries before sharing them.

“By contrast, we did our PhDs at MIT CSAIL, where patenting is rare and research is typically shared openly, often through open-source releases,” reflected Corso. “We carried that culture into Boltz, which enabled us to build a community that looks meaningfully different from what is typical in this space.”

Looking ahead, Corso expects open source to remain central to progress in biotech, particularly through foundation models that the research community can collectively build upon and fine-tune for specific tasks.

Aligned with Boltz’s mission, he also predicts that scientists working directly on preclinical programs will increasingly “care less” about access to raw models and more about end-to-end systems, notably tools that are reliable, scalable, and require minimal operational overhead.

Touching Base

Touching Base is the dynamic podcast series from the editors of GEN.
Each episode features a rotating case of senior editors—including John Sterling, Kevin Davies, Julianna
LeMieux, Alex Phillippidis, Uduak Thomas, Corinna Singleman, and Fay Lin—who delve into emerging stories,
exchange ideas, and debate the latest trends in biotech. Additionally, they talk to some of the leading
voices in the industry about what's now and next. Start
listening today!

Stay up to date with the lasted episodes of Touching Base by subscribing to the GEN
Podcast Newsletter

GEN Digital Edition

Access the latest issue of GEN and browse the archive of back issues when you subscribe to our digital
edition.

Copyright © 2026 Sage Publications or its affiliates, licensors, or contributors. All rights reserved, including those for text and data mining and training of large language models, artificial intelligence technologies, or similar technologies.

Healthcare Tech–Focused PATH Plans International Summit, Calls for Presentations

Smoking Accelerates Biological Age, Says AI