Recursion is a clinical stage TechBio company leading the space by decoding biology to radically improve lives. Enabling its mission is the Recursion OS, a platform built across diverse technologies that continuously expands one of the world’s largest proprietary biological and chemical datasets. Recursion leverages sophisticated machine-learning algorithms to distill from its dataset a collection of trillions of searchable relationships across biology and chemistry unconstrained by human bias. By commanding massive experimental scale — up to millions of wet lab experiments weekly — and massive computational scale — owning and operating one of the most powerful supercomputers in the world, Recursion is uniting technology, biology, and chemistry to advance the future of medicine. Read more about our mission

Recursion’s Co-founder and CEO is Chris Gibson, Ph.D.

Recursion was founded by Blake Borgeson, Ph.D., Chris Gibson, Ph.D., and Dean Li, M.D., Ph.D.

November 5, 2013

Recursion’s mission is to decode biology to radically improve lives.

Salt Lake City, Utah

Recursion has 4 programs in the clinical stage, spanning oncology and rare disease: 

• Oncology: 
  
  • REC-617: A CDK7 inhibitor being developed for the treatment of multiple advanced solid tumor indications.
  • REC-1245: An RBM39 degrader being developed for the treatment of biomarker-enriched solid tumor indications and lymphoma. 
  • REC-3565: A MALT1 inhibitor being developed for multiple hematology indications.
• Rare disease:
  
  • REC-4881: A MEK1/2 inhibitor being developed to treat familial adenomatous polyposis (FAP).

For more detailed information on each program, please visit our pipeline page.

Information on ongoing Recursion clinical trials and their locations can be found on clinicaltrials.gov

Yes - Recursion focuses on diseases where there is a high unmet need, including rare diseases and oncology. For rare diseases, Recursion has a clinical-stage program in familial adenomatous polyposis (FAP) and a preclinical stage program in hypophosphatasia. See our full pipeline

Yes - Recursion focuses on diseases where there is a high unmet need, including oncology and rare diseases. Recursion has multiple clinical stage and preclinical stage oncology programs ongoing, across solid tumors and blood cancers. See our full pipeline

Recursion collaborates with leading biopharmaceutical and technology companies to broadly explore diverse disease domains and rapidly identify novel therapeutic candidates via AI. Our partners fit into two main categories:

Therapeutic Partners: 

• Bayer Pharmaceuticals
• Roche and Genentech
• Sanofi
• Merck KGaA, Darmstadt, Germany
• The Gates Foundation

Data, Technology, and Capability Partners: 

• NVIDIA
• Google Cloud
• Helix
• Tempus

Visit our Partners page

The Recursion Operating System, or Recursion OS, is our proprietary, AI-enabled end-to-end platform for drug discovery and development. It is an interconnected set of technical and scientific tools that seamlessly unites physical experimentation (wet lab) with computational analysis (in silico), using intelligent machine learning models to generate novel therapeutic insights.

Explore the Recursion OS in more detail

Recursion's core competitive advantage lies in its industrialized, end-to-end platform. Unlike competitors that focus on single-point solutions, our integrated technology creates a comprehensive "World Model"—a virtual, predictive representation of biology and chemistry. This approach allows us to deliver a rapidly advancing portfolio of medicines and significantly accelerate the discovery process from idea to clinic.

Recursion employs an industrialized, AI-first strategy, replacing slow, manual, hypothesis-driven research. We use our integrated, end-to-end platform—which connects automated labs and a massive proprietary dataset with advanced AI/machine learning models—to rapidly and in an unbiased manner discover potential new medicines at an unprecedented scale, achieving speeds up to 3X faster than traditional industry averages.

Recursion applies AI/ML across the entire drug discovery and development process, targeting the three main failure points in traditional R&D. The AI then targets failures in: Efficacy (Maps of Biology), Safety (generative chemistry), and Clinical Success (patient simulations). This industrialized capability is powered by three core pillars of infrastructure:

1. Massive, Fit-for-Purpose Data: We use a hybrid strategy combining our proprietary 65+ petabyte dataset (from automated labs, like phenomics) with external real-world patient data from partners like Tempus.
2. Powerful Compute: Our BioHive-2 supercomputer (the most powerful wholly owned and operated by any biopharma company) trains the models and iterates on the World Model.
3. World-Class Talent: A "bilingual" team of technologists and life scientists integrates the data, compute, and biology across the platform.

A Virtual Cell is an AI-driven "World Model" that represents a complex, predictive simulation of human biology and chemistry. This interconnected system operates around three core functions: Predict (forecasts cell responses), Explain (reveals underlying pathways), and Discover (generates high-confidence hypotheses). This system fundamentally inverts the traditional R&D funnel, using the automated wet lab as a rapid validation tool for the AI's most promising predictions.

Maps of Biology and Chemistry are comprehensive, high-dimensional datasets and searchable digital representations of the vast complexities of biological systems. Recursion's maps are created by systematically generating and integrating large-scale experimental data—such as gene knockouts, compound perturbations, and multi-omic measurements—across living cells, molecules, and organisms. These resulting datasets function as navigational tools, guiding AI-driven exploration and enabling the discovery of:

• New disease mechanisms
• Drug targets
• Potential treatments

Autonomous drug discovery is Recursion's vision for a future state where intelligent AI agents manage and execute the entire drug discovery pipeline—from identifying targets to designing experiments—with minimal human intervention. This vision seeks to industrialize R&D by eliminating waste and vastly accelerating the speed and scale at which new medicines are created.

Recursion has spent over a decade building the foundational components required for this system, including:

• Massive Data: Generating and aggregating high-quality biological and chemical datasets using autonomous laboratory systems.
• Powerful Compute: Leveraging one of the most powerful supercomputers in biopharma for training large foundation models.
• Virtuous Cycles: Developing specialized AI tools that create continuous feedback loops between wet-lab experimentation and computational prediction to accelerate learning.

Automation plays a central, foundational role in Recursion's drug discovery process, transforming a traditional research lab into an advanced, industrialized data factory.

The primary purpose of this automation is to generate the massive, high-quality, and standardized data required to train and validate advanced AI models:

• Industrial Scale: Highly automated laboratories, often resembling advanced manufacturing facilities, execute experiments at an unprecedented volume, running up to 2.2 million experiments every week.
• Data Integrity: Robots and precision systems ensure every experiment follows the exact same protocol, drastically reducing human error and experimental variation (batch effects). This guarantees the data fed into the AI models is clean, reliable, and usable across different batches.
• Continuous Quality Control (QC): The entire process is meticulously monitored. Automated cameras capture every step, allowing the QC systems to perform "instant replays" and ensure data quality, which is critical for high signal-to-noise ratios in the resulting dataset.

In essence, automation provides the scale and standardization necessary to build the world's largest biological datasets, which in turn fuels the predictive power of Recursion's AI.

ClinTech is Recursion's approach to applying AI and technology to the entire domain of clinical research, trials, and drug development, extending its industrialized approach beyond initial drug discovery. The strategy is built on three key pillars:

• Optimizing Trial Design: Using predictive modeling and AI to refine protocols and precisely define patient populations.
• Streamlining Trial Execution: Applying technology to logistics, such as using AI-driven methods for efficient site selection and activation.
• Improving Evidence Generation: Applying machine learning to analyze large, diverse clinical and real-world datasets to produce high-quality, regulatory-compliant evidence.

This overall approach aims to improve the speed and efficiency of bringing new therapies to patients.

Yes. Recursion is committed to advancing science through open-sourcing data and models for non-commercial use. Our current public offerings include:

• Datasets: Six large-scale open-source datasets (Rxrx1, Rxrx2, Rxrx19a, Rxrx19b, Rxrx3, and Rxrx3-core). Note: These represent less than 1% of Recursion’s full, proprietary dataset.
• OpenPhenom-S/16 model: A foundation model available on Google Cloud’s Model Garden and on HuggingFace.
• Boltz-2 Model: Developed with MIT, this is an open-source model that combines structure and binding affinity prediction. It approaches the accuracy of physics-based free energy perturbation (FEP) calculations while being over 1,000 times faster and less computationally expensive. We also released a related open-source SynFlowNet-Boltz trainer, enabling researchers to to find diverse, synthesizable, high-affinity binders.

You can explore Recursion’s open source data and models at rxrx.ai

TechBio is an evolution of the life sciences focused on leveraging advanced technologies like artificial intelligence, machine learning, engineering and automation to understand biology at a systems level to discover new medicines. It relies on the creation of large, robust, highly structured biological and chemical datasets to fuel advanced computational models that can reveal subtle patterns that humans alone might miss. TechBio companies have data, compute and AI models at their core, enabling a new era of data-driven and unbiased drug discovery. 

We believe TechBio companies will create step-function shifts in the healthcare industry. Recursion is both leading this new sector, and working hard to help define it.

The difference lies in the strategic starting point and technological core. Traditional biotech begins with a therapeutic hypothesis developed around a simplified view of disease biology, relying on serial, human-intensive lab work. TechBio, conversely, starts with massive, scaled data generation and supercomputing power to train AI models and identify patterns humans can’t see. This industrialized approach enables companies like Recursion to move beyond limited hypotheses and understand complex biology at a systems level, leading to data-driven, unbiased drug discovery with vastly accelerated speed and scale.

Phenomics measures the morphology (shape and structure) of cultured cells grown in laboratory plates. Morphology is a holistic readout that integrates changes from underlying layers of cell biology—such as gene expression and protein modification—into a single, powerful measure of cellular state.

Phenomics forms the backbone of Recursion’s Maps of Biology by enabling the platform to:

• Identify and analyze biological and chemical relationships.
• Inform on potential pathway biology (e.g., mechanism of action).
• Detect off-target liabilities (unintended drug effects).

Transcriptomics is a high-dimensional measure of cellular biology, distinct from phenomics, that assesses gene expression by measuring RNA levels in the cell.

As a crucial complementary data layer, Recursion uses high-throughput transcriptomics platforms to generate massive amounts of data on chemically or genetically perturbed cells. This non-image-based data is combined with other modalities to:

• Provide confirmation of biological changes detected by the platform.
• Validate hypotheses derived from the predictive AI models.