Recursion
AI-driven drug discovery platform decoding biology to radically improve lives
Enterprise·Technical·Powered by Proprietary (Recursion OS)
Key strengths
Largest proprietary biological/chemical dataset (>50 petabytes)Automated wet lab with robotics capturing millions of cell experiments per weekBioHive-2 supercomputer built with NVIDIA for biopharma computeIntegrated Recursion OS platform covering hit ID to IND-enabling studiesActive clinical pipeline across oncology and rare disease indications
Enterprise pricing
Salt Lake City, USA
Founded 2013
No ratings yet
Technical Overview: Recursion OS & Platform Architecture
Recursion's platform is built around three foundational pillars — Data, Models, and Compute — tightly integrated into the Recursion Operating System (Recursion OS):
Data
- >50 petabytes of proprietary biological and chemical datasets spanning phenomics, transcriptomics, proteomics, ADME profiling, and de-identified patient data.
- Generated continuously via an automated wet lab utilizing robotics and computer vision, capturing millions of cell experiments per week.
Models
- Intelligent ML models trained on the above datasets to perform target identification, molecule optimization, and disease mechanism mapping.
- Models are fed back into the OS in a continuous improvement loop, enabling rapid iteration from hit identification through IND-enabling studies.
Compute
- BioHive-2 — built in partnership with NVIDIA — serves as biopharma's most powerful supercomputer, enabling large-scale model training and inference on petabyte-scale biological data.
Note: Recursion does not offer public API access or self-service developer tooling. Technical engagement is through strategic partnership agreements.