Ataraxis AI

Ataraxis AI is a New York-based AI precision medicine company that develops multi-modal machine learning models to predict cancer patient outcomes and treatment response. Its technology integrates digital pathology slide images with clinical and molecular data to identify morphological and molecular features associated with disease aggressiveness, recurrence risk and benefit from specific therapies. The company positions its work as building "AI world models" for disease, with research directions in causal inference of treatment effects, representation learning (self-supervised and multimodal foundation models) and time-to-event survival analysis.

The company's first commercial offering is Ataraxis Breast, described by the company as an AI-native prognostic and predictive platform for breast cancer covering recurrence risk, chemotherapy benefit and neoadjuvant treatment response. The company states the test was clinically validated to be 30% more accurate than the current standard of care and that it works across breast cancer subtypes, including subtypes without currently guideline-recommended diagnostic tools. Because the analysis uses existing pathology specimens, Ataraxis states no additional histology preparation or tissue consumption is required, with results returned within 24 hours of receiving patient data (company technology page cites delivery as fast as one hour after specimen receipt).

Ataraxis tests are built on the company's pan-cancer digital pathology foundation models — Kestrel, pretrained on hundreds of millions of pathology images from cancer tissue specimens using vision transformers and self-supervised learning, and Falcon, also described as trained on hundreds of millions of pan-cancer pathology slide images. The company says its models are indication-agnostic and intended to be extended to additional cancer types and treatment-selection questions.

Founding story

The company was co-founded by Jan Witowski, MD, PhD, a physician-scientist with more than a decade of medical image processing experience who previously worked at Harvard Medical School, Massachusetts General Hospital and NYU Langone Health, and Krzysztof Geras, PhD, an assistant professor at NYU Grossman School of Medicine's Department of Radiology affiliated with the NYU Center for Data Science and the Courant Institute, with prior industry experience at JP Morgan, Microsoft and Amazon. The company spun out of New York University in 2023. Witowski serves as CEO and Geras as Chief Scientific Officer.

Business model

Ataraxis operates as a diagnostics/testing company: physicians order laboratory-style prognostic and predictive tests (beginning with Ataraxis Breast) through an online ordering portal, with results generated by the company's AI models from submitted patient pathology and clinical data rather than from additional tissue processing.

Revenue derives from physician-ordered AI-based prognostic and predictive tests, beginning with Ataraxis Breast for breast cancer patients.

Traction

The company reports clinical validation of Ataraxis Breast showing 30% greater accuracy than the current standard of care, a stated plan to launch the test for U.S. oncologists in 2025, an active ordering portal, and collaborations with international cancer centers and health systems on development, validation and implementation. Research output includes presentations at NeurIPS (2024), ESMO Breast (2025), SABCS (2025) and ASCO (2026, first Phase III randomized controlled trial validation readout), and a clinical validation publication in Nature Communications (July 2026).

Latest developments

The company reported presenting the first Phase III randomized controlled trial validation study readout at ASCO in June 2026 and publication of the Ataraxis Breast clinical validation in Nature Communications in July 2026, following SABCS 2025 presentations on survival calibration and pan-cancer foundation model biomarker and survival prediction.