Recursion Pharmaceuticals

Learning Objectives

• Understand how Recursion uses automated experiments and machine learning to find drug targets and design molecules
• Explain what the Exscientia combination and the Boltz-2 model added to the platform
• Judge an AI drug developer by its clinical progress, not by the size of its dataset

What Is Recursion Pharmaceuticals?

Recursion Pharmaceuticals (NASDAQ: RXRX) is a clinical-stage drug developer in Salt Lake City, Utah, founded in November 2013. It does not sell software. It uses its own data and models to find and design medicines, then develops them itself or with pharmaceutical partners. Najat Khan, previously Chief Data Science Officer at Johnson & Johnson's pharmaceutical arm, became chief executive in January 2026, succeeding co-founder Chris Gibson.

The company's approach is to industrialize biology. Instead of starting from one hypothesis about a disease, it runs automated experiments at very large scale, photographing human cells after thousands of genetic and chemical changes, and trains models on the results to build what it calls maps of biology. Those maps suggest which genes and pathways matter in a disease and which molecules might change them.

In November 2024 Recursion completed an all-stock combination with Exscientia, the British AI drug-design company, which became a wholly owned subsidiary. Exscientia brought automated precision chemistry and its partnerships with Sanofi and Merck KGaA.

The Recursion OS

Recursion calls its platform the Recursion OS, and describes it as covering biology, chemistry and clinical development.

Maps of Biology

Automated labs generate images and other measurements from human cells, including cells grown from stem cells into neurons. In 2026 the company reported a whole-genome CRISPR knockout map built in those neurons. The maps are reusable: one experiment can feed many programs, including partners' programs.

Compute and Open Models

The work runs on BioHive-2, a supercomputer built with NVIDIA, which also invested in Recursion in July 2023. In June 2025 Recursion and MIT released Boltz-2, which predicts both the shape of a protein bound to a drug molecule and how tightly the two bind. It is open source under the MIT license for academic and commercial use. The MIT researchers behind it went on to found Boltz, which maintains the open models.

AI Agents

In 2026 Recursion added agents on top of the platform: a target discovery agent that reasons over its biology maps, drug design agents that the company says cut structural analysis from four hours to thirty minutes, and a clinical strategy agent that it credits with raising trial enrollment rates by 1.3 to 1.6 times against historical benchmarks. These are the company's own internal figures.

Pipeline

In May 2025 Recursion narrowed its pipeline, discontinuing programs in neurofibromatosis type 2, cerebral cavernous malformation and C. difficile infection.

As of August 2026, its most advanced program is REC-4881, for familial adenomatous polyposis, a rare inherited condition that causes large numbers of colon polyps. It is in a mid-stage trial called TUPELO, and Recursion has begun talks with the FDA about a registrational study. Behind it are several early-stage cancer trials, including an LSD1 inhibitor that the company describes as AI-designed. Its PI3K-alpha inhibitor, REC-7735, has cleared the FDA to begin a first trial; Recursion says the molecule took 10 months and 242 synthesized compounds to design.

Partnerships

Partnerships supply much of the funding.

• Roche and Genentech work with Recursion on neuroscience. In August 2026 Genentech exercised its first option on a validated target, and Recursion reports 216 million dollars in upfront and milestone payments from the collaboration.
• Sanofi, a partnership inherited from Exscientia, covers immunology and oncology; Recursion reports 134 million dollars in upfront and milestone payments.
• Bayer works with Recursion on precision oncology.
• Merck KGaA, also inherited from Exscientia, works with it on finding new targets in oncology and immunology.

Recursion also licenses patient oncology data from Tempus AI.

Strengths

• A very large in-house dataset of cell experiments, built for machine learning from the start
• Chemistry and biology under one roof since the Exscientia combination
• A real open-model contribution in Boltz-2, usable by anyone under the MIT license
• Partnerships with several large drugmakers that pay for access to its maps and designs

Limitations and Considerations

• No approved drug, and its most advanced program is in a mid-stage trial
• The programs discontinued in 2025 show that better target-finding does not remove clinical risk
• Speed and efficiency figures for its agents and design programs are the company's own, not independently measured
• There is nothing to buy or try: learners can run the open Boltz-2 model, but not the platform itself

How It Compares

Recursion's bet is data first: generate biology at scale, then let models find the patterns. Insilico Medicine is further along clinically, with its lead drug rentosertib in Phase III. Isomorphic Labs licenses its models to large drugmakers while building its own pipeline. Boltz took the modeling work Recursion co-developed and turned it into an open-core product anyone can use.

Key Takeaways

• Recursion is an AI-native drug developer that builds maps of biology from automated cell experiments; it sells no AI product
• The November 2024 Exscientia combination added precision chemistry and the Sanofi and Merck KGaA partnerships
• Boltz-2, released with MIT in June 2025, is its most widely usable contribution: open source for commercial use
• Its most advanced program, REC-4881, is in a mid-stage trial, and the company has discontinued three clinical programs, so its value rests on trial results