OpenEvidence's $250 Million Round Was a Bet on Physician Attention

OpenEvidence announced a $250 million Series D at a $12 billion valuation on 21 January 2026. The company said physicians ran 18 million clinical consultations in December and that more than 40% of U.S. physicians used the product daily across more than 10,000 hospitals. Those are company-reported figures, and the release does not define how it calculated physician share or daily use.

Even with that caveat, the strategic claim is clear. OpenEvidence was asking investors to value a recurring moment in clinical work: a physician looking for an evidence-backed answer. That moment sits upstream of prescribing, referrals, documentation, and patient communication. A product that becomes habitual there can gain distribution without replacing the electronic health record.

The JAMA Network announced a strategic content agreement with OpenEvidence in June 2025. That relationship matters because medical answers are judged by the quality, currency, and traceability of the underlying evidence. It also shows why this category is not only a model contest. Publishers and professional institutions control assets a general model cannot simply declare authoritative.

OpenEvidence later launched EvidenceGrade, which exposes the strength of cited evidence beneath an answer. That July release arrived after the financing, so it should be read as execution on the trust thesis rather than a pre-round signal. The distinction matters: citations can improve transparency, but they do not by themselves establish clinical accuracy, unbiased source selection, or better patient outcomes.

Mount Sinai's March 2026 announcement said OpenEvidence would be available inside Epic for clinicians across the health system. The timing is post-round, but the integration makes the competitive direction concrete. The closer an evidence tool sits to the patient chart, the less effort a physician spends switching contexts and the more often the tool can become part of routine work.

That puts pressure on clinical reference vendors, health-system decision-support tools, specialty products, and medical publishers in different ways. The winning counter-position may be deeper specialty evidence, tighter local protocol integration, clearer governance, or independently demonstrated clinical value.

The pattern continued in September. On 16 September Memorial Sloan Kettering announced a two-way arrangement: OpenEvidence inside MSK's Epic record, and MSK's OncoKB precision-oncology knowledge base inside OpenEvidence for clinicians at other institutions. That single agreement joins both parts of the thesis, licensed expert content and distribution inside the clinical record.

The funding round validates investor belief in distribution; it does not establish clinical benefit. A Mayo Clinic study listing shows the product being evaluated against nurse-generated patient answers, but a registered study is not a result. Independent evidence on diagnostic quality, treatment decisions, time saved, bias, liability, and patient outcomes remains more important than a valuation.

Trust also depends on how the product is paid for. OpenEvidence is funded by advertising: according to Out-Of-Pocket's March analysis, clinicians see an ad while an answer is being generated, and the physician audience is what advertisers buy. That analysis and other health-industry commentary have argued both sides: ads keep the tool free for clinicians whose employers do not pay for reference subscriptions, while sponsored research and relevant ads at the point of care could shape decisions if they are not clearly separated. How OpenEvidence labels sponsored material and which sources it treats as evidence are questions clinicians are already asking.

Competitors should resist replying with a larger generic knowledge claim. They should make the evidence chain legible: which sources are licensed, how recommendations are graded, when local guidance overrides general literature, what a clinician must review, and where the product has been independently tested.