Two-Thirds of American Doctors Now Use OpenEvidence as Its Ad Revenue Heads for 500 Million Dollars

OpenEvidence, the AI search engine built for doctors, says nearly two-thirds of the roughly one million physicians in the United States now use it, according to the Financial Times. A source told the paper the Miami-based startup is on track to generate between 400 and 500 million dollars in advertising revenue in 2026, up from 100 million dollars in ad commitments booked for 2025. The company itself has not published a revenue figure, and the projection comes from a single unnamed source, so treat it as directional rather than audited.

The adoption claim, if it holds, would make OpenEvidence one of the most widely used AI tools within any single profession. Clinicians are famously cautious adopters, and the product remains free for verified doctors in the US and Canada. The business model is advertising: pharmaceutical companies pay to place ads alongside answers drawn from peer-reviewed literature including the New England Journal of Medicine and JAMA. The platform now processes more than 42 million prompts per month, and founder Daniel Nadler calls it the first viral consumer internet app in healthcare.

Investors have chased that growth. OpenEvidence raised 250 million dollars in September at a 15 billion dollar valuation from Andreessen Horowitz, Blackstone, Thrive Capital and Sequoia Capital, roughly triple the 12 billion dollar valuation set in January, which itself followed a 6 billion dollar round in October 2025 and 3.5 billion dollars in July 2025.

Hospitals are moving from informal use to formal plumbing. AdventHealth wired OpenEvidence into its electronic health record across more than 2,000 care sites on October 1, Penn Medicine deployed it to over 10,000 clinicians across seven hospitals in late September, Northwestern Medicine made it accessible through Epic with hospital credentials on October 8, and NYU Langone, Mount Sinai and Memorial Sloan Kettering have formal integrations. NYU Langone chief health informatics officer Paul Testa said clinicians were already using the tool as an independent resource before the hospital stepped in.

The accuracy record is mixed. A Mayo Clinic study, with the caveat that Mayo is also an investor, found accurate evidence-based recommendations in sampled cases, but research published in Nature found OpenEvidence underperforming Google's Gemini and OpenAI's GPT on standard clinical benchmarks. A JAMA adoption study documented 556,200 verified physicians using the tool cumulatively over 30 months and 271 million queries from 952,008 verified clinicians, though it was NIH-funded, used company data and included three OpenEvidence employees among the authors, and did not evaluate answer accuracy or patient outcomes.

Regulatory reach is another constraint: OpenEvidence withdrew from the EU and UK earlier this year, citing the compliance costs and restrictions of the AI Act, leaving it dependent on the US healthcare system. It now competes with subscription incumbents like Wolters Kluwer's UpToDate and Doximity, while OpenAI and Anthropic push their own clinical offerings.

If the ad revenue lands anywhere inside the reported band, OpenEvidence will have built the first durable ad-funded business model in vertical clinical AI, and every seat-based medical software vendor will be recalculating. The open question is whether pharma money sitting next to clinical answers survives contact with regulators and with doctors who ask exactly where those answers come from.