OpenEvidence says nearly two-thirds of ~1M US physicians use its AI search engine; a source says it is on track to generate $400M-$500M in 2026 ad revenue (Rafe Rosner-Uddin/Financial Times)

OpenEvidence is an AI-powered search tool for physicians. Doctors can ask medical questions in ordinary language instead of typing only keywords. The system then searches and summarizes relevant medical information into an answer. Its importance comes from speed: clinicians may use it while considering a diagnosis or treatment, when time matters.

For example, a physician could ask about evidence linked to a symptom, condition, or therapy. An AI search engine interprets the question, retrieves material, and produces a concise response. Unlike a simple list of links, it aims to organize information into a usable answer. The source describes OpenEvidence as ad-powered, meaning advertising supports the service.

The reported reach is unusually large. OpenEvidence says nearly two-thirds of about one million US physicians use it. That makes the tool influential in clinical information seeking. Its growth also raises questions about whether answers are accurate, transparent, and appropriately separated from advertising. The source does not explain OpenEvidence's exact search technology or evidence sources.

The source says nearly two-thirds of approximately one million US physicians use OpenEvidence. Two-thirds of one million is about 666,667, so the claim implies roughly 667,000 doctors. Because both “nearly” and “roughly” signal approximation, this should not be treated as a precise user count.

The scale is striking because it suggests OpenEvidence has reached a large share of the US physician population. A doctor using the tool might seek medical information during routine work, diagnosis, or treatment planning. The source does not say whether every user is active, how often physicians use it, or whether use is independently verified.

This level of adoption gives OpenEvidence substantial influence over how doctors encounter medical information. It also makes the quality of its answers and the visibility of advertising important concerns. A large user base can support a powerful business, but reach alone does not prove that the tool improves clinical decisions or patient outcomes.

OpenEvidence makes money through advertising built into its physician-focused search experience. The source does not describe the ad formats, pricing, or targeting rules. It does report that the company is on track to generate $400 million to $500 million in advertising revenue during 2026, according to a source.

The basic mechanism is familiar: advertisers pay to reach a valuable audience, while the service uses that revenue to support a free or widely available tool. Here, the audience is physicians. A doctor searching for medical information may encounter sponsored material alongside, before, or within an AI-generated answer, depending on the product's design. Those details are not provided.

The forecast indicates that OpenEvidence's advertising model could be commercially substantial, not a minor side business. It also increases the importance of clear disclosure and safeguards. If advertising revenue grows with physician usage, commercial incentives may influence product design, audience targeting, or the placement of information. The forecast is not a guarantee of actual 2026 revenue.

Pharmaceutical and healthcare companies may value OpenEvidence because it reaches a large professional audience. The company says nearly two-thirds of roughly one million US physicians use the tool. That gives advertisers access to doctors in a setting closely connected with medical questions and clinical information.

For example, a company could want a physician to notice information about a medicine, device, or healthcare service while researching a condition. The key mechanism is targeted attention: advertising appears within a tool doctors already use for work. The source confirms the tool is ad-powered, but it does not state which companies advertise, what products they promote, or how ads appear.

This business model can be attractive because physician attention may influence later conversations or decisions about care. It also creates a need to distinguish paid messages from independent evidence. The more widely doctors use OpenEvidence, the more valuable its advertising space may become. Commercial reach, however, does not establish that an advertisement is medically appropriate or beneficial.

If doctors use OpenEvidence during diagnosis or treatment, its answers could shape clinical decisions. Reliable, current information might help physicians compare options or answer questions faster. But an AI-generated response can also contain errors, omit important context, or misunderstand a patient's situation. The source raises a debate about patient care but provides no outcome data.

For example, a doctor might ask the tool about a treatment and use its summary to guide a discussion. The mechanism is reliance: the answer may affect what evidence the physician notices, which options receive attention, or how confidently a decision is made. Advertising could add another influence if sponsored content is not clearly separated from medical evidence.

The current reality is reported widespread use, not proven clinical benefit. Doctors remain responsible for checking information against appropriate evidence and the patient's circumstances. As adoption grows, OpenEvidence's accuracy, transparency, and handling of commercial material may become increasingly important. The article does not report specific harms, benefits, or safeguards.

An AI-generated medical answer is a summary produced in response to a question. Medical journals publish studies, while clinical guidelines present recommendations developed through formal review and expert processes. Traditional sources usually let readers inspect the original evidence, methods, limitations, citations, and publication context. OpenEvidence may make information easier to consume, but the source does not explain exactly what material it uses.

For example, a physician could ask OpenEvidence about a treatment and receive a concise response. The mechanism is synthesis: the system interprets the prompt and turns information into an answer rather than simply displaying documents. A journal search may require more time, but it allows the doctor to read the study. A guideline may offer structured recommendations and conditions for applying them.

AI tools and traditional sources can serve different purposes. AI may help locate or summarize information, while journals and guidelines support verification and deeper judgment. OpenEvidence's reported scale makes that distinction important. The source does not establish that its answers are better, faster, or less reliable than conventional medical resources.

Verification matters because an AI answer may be incomplete, mistaken, or poorly matched to a patient's situation. Health decisions can affect diagnosis, treatment, and safety. Checking the answer against medical journals, clinical guidelines, and patient-specific facts helps prevent an appealing summary from becoming unquestioned advice. The source does not document specific OpenEvidence errors, but it does describe a debate about patient care.

Advertising disclosure matters because OpenEvidence is ad-powered and a source projects $400 million–$500 million in 2026 ad revenue. If paid content is not clearly identified, doctors may confuse promotion with independent evidence. Bias management matters for the same reason: the system's outputs or presentation could favor certain products, sources, or viewpoints.

These safeguards become more important as usage expands. OpenEvidence says nearly two-thirds of roughly one million US physicians use the tool, giving it broad influence. Clear labeling, evidence checking, and attention to competing recommendations can support better judgment. The article does not state which safeguards OpenEvidence currently uses or whether advertising changes its answers.