Boltz
Category: AI in Healthcare
An AI biotechnology company developing open-source biomolecular foundation models designed to predict the structure and interactions of proteins, DNA, RNA, and small molecules to accelerate drug discovery. Boltz was founded in 2024. The company is led by Gabriele Corso. Based in Cambridge, USA. Team size: 10-20. Total funding raised: $28.0M. Latest round: Seed round (Jan 2025). Key investors include Amplify Partners, Andreessen Horowitz (a16z) Bio + Health, Zetta Venture Partners, Factorial Capital, Obvious Ventures, Clément Delangue (Angel).
- Founded
- 2024
- Headquarters
- Cambridge, USA
- Team size
- 10-20
- Total funding
- $28.0M
Value proposition
Democratizes state-of-the-art structural biology by providing open-source, high-fidelity foundation models that rival proprietary systems like AlphaFold 3, enabling researchers to predict complex molecular interactions without restrictive licensing.
Products and solutions
Boltz-1 (Open-source biomolecular foundation model, Nov 2024), Boltz-1x (Inference-time steering variant, late 2024), Boltz-2 (Joint structure and affinity prediction model, Jun 2025), BoltzGen (Generative protein design model, Nov 2025), Boltz Lab (Hosted platform with AI agents, Jan 2026), DiffDock (Molecular docking tool)
Unique value
Boltz is the first to provide a fully open-source alternative to Google DeepMind’s AlphaFold 3, offering the community both the model weights and the code for commercial applications, which was previously a major barrier in the field.
Target customer
Pharmaceutical companies, biotechnology startups, academic research institutions, and drug discovery platforms.
Industries served
Biotechnology, Pharmaceuticals, Life Sciences, Healthcare AI, Academic Research
Technology advantage
Leverages proprietary geometric deep learning architectures and generative modeling expertise from the creators of DiffDock. Their 'open-science' business model attracts top-tier talent and rapid industry adoption, validated by a strategic partnership with Pfizer.
How they differentiate
Boltz differentiates through an 'open-science' approach, providing both model weights and training code under a commercial-friendly license. While competitors like DeepMind initially restricted AlphaFold 3's commercial use, Boltz offers a high-fidelity, open-source alternative that matches state-of-the-art performance in predicting complex molecular interactions.
Main competitors
Google DeepMind (AlphaFold 3), EvolutionaryScale (ESM3), Chai Discovery (Chai-1), Profluent
Key partnerships
Pfizer (Strategic partner for biomolecular AI development), Andreessen Horowitz (a16z) Bio + Health (Lead Investor), MIT Jameel Clinic (Academic collaboration and research roots), BoxGroup and GreatPoint Ventures (Investment partners)
Notable customers
Pfizer (Strategic Partner)
Major milestones
Released Boltz-1 as open-source rival to AlphaFold 3 (Nov 2024), Released Boltz-2 joint structure/affinity model with Recursion (Jun 2025), Released BoltzGen generative protein design model (Nov 2025), Launched as Public Benefit Corporation with $28M seed led by Amplify, a16z, Zetta (Jan 2026), Established strategic collaboration with Pfizer for biomolecular AI models (Jan 2026), Previewed Boltz Lab hosted platform with AI agents for drug discovery (Jan 2026)
Growth metrics
Rapid community adoption of Boltz-1; founders' previous research (DiffDock) has over 4,500 citations.
Market positioning
Open-source challenger in the AI-driven structural biology and drug discovery market.
Geographic focus
Global (via open-source distribution) with a primary corporate focus on the North American biotechnology hub (Cambridge, MA).
Patents and IP
No specific registered patents disclosed; the company currently prioritizes an open-source IP strategy to drive industry-wide adoption.
About Gabriele Corso
Gabriele Corso is the CEO and Co-founder of Boltz. He holds a PhD from MIT CSAIL, where he specialized in generative models and geometric deep learning for structural biology. He is the lead developer of the Boltz-1 and Boltz-2 biomolecular foundation models and the creator of DiffDock, a widely used tool for molecular docking. His research has been cited over 4,500 times, and he was recognized in the Forbes 30 Under 30 list for Healthcare & Science.
Official website: https://boltz.bio/