ai

Aignostics is the Berlin pathology AI company that built the Atlas foundation model with Mayo Clinic.

The Arc Institute is a well-funded nonprofit research institute building 'virtual cell' AI models and the datasets to train them.

French startup behind H-optimus-0, an open 1.1-billion-parameter pathology foundation model.

Makes Chai-1 and Chai-2, open-weight structure models used for antibody and binder design.

Funds and builds the Human Cell Atlas infrastructure and the CZI virtual cell models (TranscriptFormer), plus one of the largest nonprofit GPU clusters for biology.

Generative chemistry platform behind the first AI-discovered drug to reach phase 2, plus oncology candidates.

Breast radiologist who co-developed the Mirai AI model that predicts breast cancer risk from a mammogram.

Cradle is a protein-engineering AI platform used by pharma to optimise antibodies and enzymes.

Radiation oncologist who built AI models from trial slides to decide which prostate cancer patients need hormone therapy.

Spun out of Meta's protein team; makes the ESM3 protein language model that generated a new fluorescent protein.

Built UNI and CONCH, the pathology foundation models that let AI read whole-slide images across cancer types.

Genesis Molecular AI does small-molecule discovery on its GEMS platform, which combines physics and deep learning, with oncology programmes in the clinic pipeline.

Google DeepMind built AlphaFold, AlphaMissense, AlphaGenome and Med-Gemini, the reference models for structure, variants, and medical multimodal reasoning.

HistAI, of Miami, released Hibou-B and Hibou-L in 2024, open vision transformer models for pathology trained on more than a million slides and published under the Apache 2.0 licence, which allows commercial use. Whether permissively licensed open models reach regulated clinical products or stay research tools is the open question.

AI-discovered oncology molecules in the clinic, including a HER2 inhibitor and a CDK2/4 inhibitor.

David Baker's institute, where de novo protein design and the RFdiffusion models come from (Nobel Prize in Chemistry 2024).

Showed that deep learning can read genetic features like microsatellite instability directly from routine pathology slides.

Dutch-Swiss company releasing open pathology foundation models (Midnight) and evaluation tooling (eva).

Led MASAI, the first randomised trial of AI-supported mammography screening, which found more cancers with half the radiologist workload.

Founded by an AlphaFold co-developer to make protein design programmable; released Latent-X in 2025.

Lila Sciences is Flagship Pioneering's 'scientific superintelligence' company, combining AI with autonomous labs, with life science as a first domain.

Microsoft co-developed Prov-GigaPath and BiomedCLIP and is the compute partner for Paige's Virchow2.

Supplies the GPUs and the BioNeMo framework most biological foundation models are trained on, and co-developed Evo 2 with Arc.

Profluent uses protein language models to design new gene editors (OpenCRISPR-1) and antibodies.

Providence is a 51-hospital US health system whose 1.3 billion pathology tiles trained Prov-GigaPath with Microsoft.

Computer scientist who built Mirai and Sybil, AI models that predict breast and lung cancer risk years before diagnosis.

Trained the first clinical-grade AI on tens of thousands of slides, leading to the first FDA-authorised AI pathology product.

Produced Tahoe-100M, the largest single-cell drug-perturbation atlas, and trains models on it.

Launched in 2024 with over $1 billion to build AI-native drug discovery from Baker-lab protein design.

Home of the Cell2Sentence single-cell language models built with Google, and an NCI-designated comprehensive cancer centre.