Pfizer and Boltz have announced a collaboration aimed at expanding the use of advanced biomolecular AI across Pfizer’s drug discovery and preclinical research programs.
The partnership brings together Pfizer’s scientific data and therapeutic expertise with Boltz’s open-source AI foundation models and generative design workflows, signaling a deeper integration of AI into pharmaceutical R&D.
Gabriele Corso, CEO of Boltz, said, “Pfizer scientists have been some of the earliest adopters of our open-source models and members of our community – with users across modalities and disease areas. This partnership helps us take our platform to a new level in terms of accuracy, performance, and integration.”
The agreement, revealed today (Jan. 8), reflects a broader industry shift toward using large-scale AI models to shorten discovery timelines, improve decision-making, and reduce the cost and risk associated with early-stage drug development.
Along with that news, Boltz received the backing of Zetta, Amplify, a16z, and angels, including Clement Delangue (CEO of Hugging Face), Factorial Capital, and Obvious Ventures with a $28 million seed round.
It’s a busy day as it also launched Boltz Lab and its first agents for small-molecule and protein design. Boltz Lab combines models and agents with compute, scalable, infrastructure, and collaborative interfaces. The firm offers guarantees: “you own what you build, your data stays secure, and we do not train on your data.”
Deploying AI foundation models at scale
At the core of the collaboration are Boltz’s biomolecular foundation models, including Boltz-2 and BoltzGen. These models are used in the pharmaceutical and biotechnology sectors for tasks such as protein structure prediction, biomolecular design, and estimating binding affinity between molecules.
Boltz has combined these models with proprietary generative AI workflows, “user-friendly” interfaces, and high-performance computing infrastructure, enabling their direct use in real-world preclinical discovery programs. For Pfizer scientists, this means access to AI tools that can analyze molecular interactions, generate candidate compounds, and evaluate design options before committing to costly laboratory experiments.
By embedding these capabilities across Pfizer’s research organization, the collaboration aims to make AI-assisted design a routine part of discovery rather than a specialized or experimental capability.
Exclusive models built on Pfizer data
Under the terms of the partnership, Boltz will further refine its latest foundation models using Pfizer’s historical datasets. These refinements will result in exclusive models tailored to Pfizer’s needs in areas such as biomolecular structure prediction, small-molecule affinity estimation, and biologics design.
Boltz scientists will also work directly with Pfizer’s discovery teams to develop custom AI models and workflows for specific target programs. This collaboration is intended to improve the speed and quality of decision-making in preclinical research, where early insights can determine whether a program advances or is deprioritized.
The focus on exclusivity underscores the competitive importance of proprietary data in AI-driven drug discovery. While the underlying Boltz models are open-source, the Pfizer-specific versions are designed to capture insights unique to Pfizer’s research history and therapeutic focus.
Industry context and implications
The partnership comes amid growing interest in so-called AI foundation models for biology, which mirror the role of large language models in text and code. These biomolecular models aim to learn general principles of biological structure and interaction from vast datasets, allowing them to generalize across targets and disease areas.
For Pfizer, the collaboration aligns with a broader strategy to modernize discovery through digital and computational approaches. For Boltz, working with a major pharmaceutical company provides both validation and an opportunity to push the limits of its models using real-world, high-quality data.
The deal also highlights the increasing role of open-source AI in life sciences. Boltz’s models have gained traction precisely because they are accessible to a broad community, creating a user base that includes academic researchers, biotech startups, and large pharmaceutical companies.
In terms of funding elsewhere, global investment in financial technology rebounded sharply in 2025, marking a potential turning point after several years of contraction.
