OpenEvidence, an AI-powered medical search engine, says ~65% of US physicians, or ~650K doctors, and another 1.2M doctors internationally use its services
OpenEvidence, an AI-powered medical search tool, has become a fast friend to America's doctors and is now used by nearly two-thirds of physicians.
Context & Ripple Effects
OpenEvidence’s reported reach follows a rapid financing arc: it raised its first outside round at a $1B valuation in early 2025, was valued at $6B by October, and was later reported to be seeking capital at $12B. Its model has been described in related coverage as an ad-supported AI chatbot for medical professionals.
The adoption claim also lands as AI is already becoming routine in physician work. An AMA survey cited in related coverage found that more than 80% of surveyed US physicians use AI professionally, particularly for research summarization and clinical documentation.
First-order effects
- OpenEvidence becomes a much more consequential distribution channel for clinical information if its reported physician user base is accurate, giving the company greater leverage with advertisers and prospective investors.
- Physicians using the service gain a dedicated AI layer for finding and summarizing medical information, while OpenEvidence must sustain trust and usefulness across a far larger clinical audience.
Second-order effects
- Medical-information publishers, clinical reference products, and other AI tools for clinicians face a stronger incumbent in the research-summarization workflow, one of the physician AI use cases identified by the AMA survey.
- An ad-supported clinical AI product at this scale could make advertiser relationships and the handling of commercial influence more central competitive questions for physician-facing AI services.
Third-order effects
- If physician adoption continues to concentrate around a few AI search interfaces, clinical-information discovery may shift from standalone reference destinations toward AI-mediated answers, increasing the importance of evidence quality and trust controls.
- The pattern points to healthcare AI’s near-term commercialization centering first on workflow and information access rather than autonomous diagnosis; the cited OpenAI emergency-room study shows diagnostic capability is advancing too, but it does not establish clinical deployment.
The trend: Physician AI is moving from isolated experimentation toward broadly adopted, workflow-specific platforms, with medical search and summarization emerging as an early distribution wedge.