Modern healthcare runs on data, and there’s too much of it. Thousands of new studies appear every day, each one adding to the mountain of evidence that doctors are expected to know. This gap between research and practice can slow decisions, affect patient safety, and strain already overworked medical teams.
That’s where OpenEvidence comes in. Often described as “ChatGPT for doctors,” the startup gives medical professionals fast, reliable insights backed by peer-reviewed research.
Earlier this week, OpenEvidence closed a $200 million round at a $6 billion valuation, led by Google Ventures with participation from Sequoia, Kleiner Perkins, Thrive Capital, Coatue, and others.
Giving every healthcare worker the ability to make evidence-backed decisions instantly
OpenEvidence launched in 2021, founded by Daniel Nadler and Zachary Ziegler. Nadler holds a PhD from Harvard and previously founded Kensho Technologies, an AI company that S&P Global acquired for $550 million in 2018. Ziegler, the company’s CTO, was a PhD candidate in machine learning at Harvard, where he worked with professor Alexander Rush in natural language processing.
The inspiration for OpenEvidence came from personal experiences with healthcare challenges. Nadler’s grandfather died from a medical error, while Ziegler’s brother-in-law battled leukemia. Both founders recognized that doctors faced the same information overload problem that Nadler had solved in finance with Kensho: too much data, too little time to process it.
At the heart of OpenEvidence is a mix of retrieval-augmented generation (RAG) techniques and structured medical ontologies like SNOMED CT and MeSH. The system is trained exclusively on data from The New England Journal of Medicine and JAMA, and every output links directly to its source material.
Clinicians interact with OpenEvidence much like they would with a colleague: through natural-language questions. But behind that interface sits a high-precision reasoning engine built to filter, rank, and verify responses before they ever reach a user’s screen.
That emphasis on accuracy and transparency sets the startup apart from its competition. Rivals like Google’s Med-PaLM, Nuance’s Dragon Medical One, and Amazon’s HealthScribe use broader language models or speech tools that weren’t specifically trained for clinical interpretation.
What’s next?
OpenEvidence now handles more than 15 million medical consultations every month, nearly doubling its usage since July. The platform is actively used across over 10,000 hospitals and medical centers, with more than 40% of U.S. physicians logging in daily.
With the new funding, the company plans to expand into Europe and Asia, retraining its models on local medical journals and languages to better serve regional clinicians. OpenEvidence is also developing multimodal capabilities to analyze medical imaging alongside text data, and has launched DeepConsult, an AI agent that conducts advanced medical research autonomously while physicians attend to other tasks.
