Overview
Epsilon Health, incorporated as Epsilon Labs, is an AI-native radiology practice that reads medical imaging and reports on it. Board-certified radiologists own every diagnosis while the company's models support throughput, accuracy, and report quality, an arrangement aimed at imaging volume that keeps rising against a shrinking radiologist workforce, leaving patients waiting days for results that should take hours. It serves imaging centers and payers, and its research teams build the models and evaluation tools the practice runs on. Unlike vendors that sell software to hospitals and leave adoption to them, it operates the practice itself, so models reach live reads instead of papers. The company's goal is to shorten turnaround by putting clinicians, engineers, and researchers inside one organization.
Work here?Claim your company
Simplify's take
What believers are saying
- September 2026 funding raised $27.6 million from AlleyCorp, Uncork, Renegade, and SemperVirens.
- By November 2026, Epsilon reported 3,000 daily studies and 300,000 patients served.
- Verda said Epsilon runs five times more experiments weekly with three-times faster training.
What critics are saying
- Outpatient imaging giants internalize Epsilon's workflow, then squeeze pricing at renewals.
- Radiology groups and teleradiology incumbents like vRad and Akumin already own referral relationships.
- Model errors trigger missed diagnoses; one lawsuit ends the service model.
What makes Epsilon Health unique
- It is a radiology practice, not a point-solution vendor, owning service delivery.
- AI augments board-certified U.S. radiologists, who sign every report and retain clinical responsibility.
- Epsilon trains on real corrections from 300,000 patients, building a proprietary workflow dataset.
Help us improve and share your feedback! Did you find this helpful?