Boston, Massachusetts, United States, October 21, 2026
AI Proteins, Inc. and Astellas Pharma Inc. have entered into a research collaboration and licensing agreement to discover and develop novel AI-designed miniprotein therapeutics, in a deal potentially worth more than $400 million, in addition to royalties on sales of resulting medicines. The agreement brings together AI Proteins’ computational protein-design platform and Astellas’ expertise in pharmaceutical research and drug development. The collaboration will focus on multiple therapeutic targets selected by Astellas, with the goal of creating new protein-based drug candidates. The partnership reflects growing pharmaceutical industry interest in applying artificial intelligence to early-stage drug discovery. Rather than relying solely on naturally occurring proteins as starting points, the companies aim to design miniproteins from scratch, engineering their structures and properties around specific therapeutic objectives.
AI-Driven Design Targets Novel Therapeutic Miniproteins
Under the agreement, AI Proteins will use its proprietary AI-driven discovery platform to design and optimise de novo miniproteins against multiple targets identified by Astellas. These small, engineered proteins are designed to bind selected molecular targets while incorporating properties intended to support their potential use as medicines. Traditional biologic medicines often originate from proteins found in nature and are subsequently adapted for therapeutic purposes. In contrast, de novo protein design begins with a defined objective and computationally generates candidate structures intended to achieve it. AI Proteins combines this approach with high-throughput experimentation and optimisation to identify and refine promising candidates. The company aims to engineer miniproteins with precise target-binding activity and tailored molecular properties. Depending on the molecule and its intended application, such properties may influence stability, manufacturability, and potential administration routes. These characteristics must be established experimentally for each candidate, and computational design alone cannot demonstrate clinical effectiveness. The collaboration will explore multiple targets that have not been publicly disclosed. The companies have not announced specific disease indications or named drug candidates arising from the agreement.
Astellas to Lead Development and Commercialisation
The agreement establishes a division of responsibilities between the two companies. AI Proteins will design and optimise candidate miniproteins against Astellas-selected targets, while Astellas will hold exclusive worldwide licences to resulting therapeutic candidates and assume responsibility for subsequent development, regulatory activities, and commercialisation. AI Proteins will receive an upfront cash payment for its research work and may earn additional milestone payments as candidates achieve specified development and commercial objectives. The company is also eligible for royalties on net sales of any medicines that reach the market. The potential agreement value exceeds $400 million in upfront and milestone payments, with royalties payable separately under the agreement. The headline value represents potential payments tied to contractual terms and future achievements, rather than an immediate cash payment of that amount. The upfront payment itself was not disclosed in the announcement. For Astellas, the collaboration supports its strategy of strengthening its early-stage research pipeline through partnerships that provide access to emerging technologies and new therapeutic approaches. For AI Proteins, the arrangement provides a pharmaceutical partner with the resources and experience to advance successful research candidates through later development stages.
Advertisement
Expanding the Role of Computational Protein Design
AI-enabled drug discovery is increasingly being explored as a way to generate and evaluate molecular candidates more efficiently. Protein design presents a particularly relevant opportunity because therapeutic performance can depend on a complex combination of molecular structure, target interaction, stability, and other biological properties. AI Proteins is developing its platform to create purpose-built miniproteins rather than simply modifying existing natural proteins. The company combines computational design with experimental optimisation, with the aim of moving selected candidates toward the studies required before clinical testing. However, the development pathway remains demanding. Candidate molecules must undergo appropriate laboratory evaluation, preclinical safety and efficacy studies, and clinical trials before regulatory authorities can determine whether they are suitable for patients. No clinical-stage candidate, regulatory approval, or patient benefit was announced as part of this collaboration. The agreement with Astellas illustrates how established pharmaceutical companies are partnering with specialist biotechnology firms to access AI-driven molecular design capabilities. If the research generates promising candidates, the partnership could contribute to future biologic drug development. Its ultimate impact will depend on the quality of the molecules discovered and their ability to demonstrate safety and therapeutic benefit during subsequent development.
Source: AI Proteins, Astellas Pharma press release
