Skip to content

Invest2 publishers3 min readPublished

Anthropic's president says market incentives alone would not have produced its 100-country medical AI rollout

Anthropic and OpenEvidence are giving doctors in about 100 low- and middle-income countries free access to a clinical search tool that is already free in the US and Europe, where physicians ran 42 million queries in August.

The Investor · Invest desk

Illustration accompanying Anthropic's president says market incentives alone would not have produced its 100-country medical AI rollout

What happened

  • Anthropic and OpenEvidence are launching free clinical search for healthcare providers in around 100 countries, among them Angola, Haiti, Mongolia, Sudan and Uganda, on a country list OpenEvidence supplied to Reuters.
  • Anthropic is providing the underlying models and OpenEvidence is adapting them to local medicine, for physicians who lack easy access to medical journals, specialists and continuing education.
  • Founder Daniel Nadler estimated that several hundred million Americans will have been treated in 2026 by doctors who used the platform.
  • The Wall Street Journal reported in August that Anthropic believes its healthcare and biology work will improve public sentiment about AI, one of the strategies it has shared with potential investors as it prepares to go public.

Compiled by The InvestorSomething wrong?How this is made

Why it matters

  • cost The query cost sits with the two companies for a user base that pays nothing on either side of the income divide, so the program runs for as long as someone keeps paying for the compute.
  • capability Anthropic gets its models into clinical question-answering across about 100 countries without hiring a sales force in any of them, and what it pays for that is inference.
  • exposure If the regional adaptation work falls short, the visible failure is a treatment answer given to a patient in Sudan or Mongolia, and it attaches to Anthropic's models while healthcare is part of its pitch to investors.
  • contradiction The country count is the company's own, and one of the two published accounts calls the rollout dozens of countries in one line and around 100 in another. Nobody outside the companies has counted.

Anthropic supplies the models and OpenEvidence adapts them to local medicine [1]. Inference is the recurring cost of that split, and it rises with use. In the US, doctors ran 42 million queries through OpenEvidence in August alone, founder Daniel Nadler told Reuters [7], which works out at about 1.35 million a day [8]. Neither published account says what the new program costs or who pays for the compute [22].

OpenEvidence has been free to clinicians in the US and Europe for some time [4], and the specialized version for low- and middle-income countries is free as well [3]. The change is where it is offered. Anthropic president Daniela Amodei told Reuters that the technology had become capable enough for the initiative but that "market incentives by themselves would not cause it to happen" without a philanthropic push [10]. An officer of the company is saying the usage does not pay for itself.

Last week Anthropic built a lab for physical biology work aimed at drug development [17]. What it gets back for the clinical giveaway is sentiment, and it has already put that argument in front of investors [18].

Checking the distribution case is harder. More than a million queries a day inside clinical workflow is a position of some value, and PYMNTS Intelligence has found that AI adoption in healthcare is not as extensive as in sectors such as financial services [20], so getting in early on how doctors look things up may be worth money later.

The first thing that can go wrong is clinical. Reuters reported the criticism that systems trained primarily on data from wealthier countries may not align with "local practice conditions" [14]. "One hundred percent of what we are developing for these areas is designed to be context adaptive," Nadler said [13], and the company worked with healthcare organizations in Rwanda and Botswana earlier this year to adapt its tools [15]. Delivery is a separate constraint: Nadler said most doctors still have smartphones even where clinics have inconsistent power [16]. Ahmed Bendary, a cardiologist at Benha University in Egypt, told Reuters that the expansion is especially important for doctors in places where access to medical research is limited [19].

I think the philanthropic reading is right, because Anthropic's president gave it on the record and dated it to a push that market pricing would not have funded [10]. The counter-thesis is that free access across about 100 countries is a cheap way into clinical care, and that a platform absorbing 42 million US queries a month will charge somebody for them eventually [3][7]. Two developments would settle it against the first reading: a paid tier in any of the new markets, or a renewal of the program that needs no philanthropic money.

What to watch

  • Whether any Anthropic listing disclosure puts a number on the donated compute and names the philanthropic funder behind the free access.
  • Whether an independent count of participating countries and physicians appears, against the list OpenEvidence provided to Reuters.
  • Whether the Rwanda and Botswana adaptation work produces documented changes to the guidance the tool gives in low-resource clinics.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories