Every Cure, OpenEvidence launch AI drug-repurposing partnership
Every Cure and OpenEvidence will pair AI-assisted drug-repurposing hypotheses with rare-disease physicians, clinical-trial referrals and point-of-care evidence distribution.

Why it matters
- Every Cure and OpenEvidence announced the partnership on November 16, 2026.
- Every Cure estimates roughly 18,000 recognized diseases, about 4,000 approved treatments, and more than 300 million people without an FDA-approved treatment.
- OpenEvidence works with NORD on more than 3,000 expert-reviewed rare-disease summaries.
- David Fajgenbaum has been in remission for more than 11 years and advanced 14 additional repurposed drugs.
- New-drug development can require billions of dollars and more than 10 years, according to the announcement.
By Marcus Bennett
Every Cure and OpenEvidence announced on November 16, 2026, a partnership that combines artificial intelligence (AI), specialist input, clinical-trial recruitment and evidence distribution for drug repurposing (new uses for existing medicines) where commercial incentives are scarce.
Every Cure estimates that roughly 18,000 diseases are recognized but only about 4,000 have approved treatments. The nonprofit frames that broader recognized-disease set as lacking commercial incentives for new treatments.
The nonprofit also estimates that more than 300 million people globally have a disease with no treatment approved by the U.S. Food and Drug Administration (FDA). The partnership targets approved drugs, including generics, that Every Cure can pursue without a conventional commercial return.
The research stakes are concrete because Every Cure’s estimate covers hundreds of millions of people without an approved treatment for their disease. Every Cure and OpenEvidence say their collaboration will test whether an organized pipeline can turn existing medicines and scarce specialist knowledge into credible trial evidence.
Repurposing also changes the starting point. Existing safety data and manufacturing capacity support a candidate before testing a new indication, while new-drug development can require billions of dollars and more than 10 years, according to the announcement. The announcement explicitly says this rationale does not establish that any particular candidate will work in a new disease.
What does the partnership actually connect?
Every Cure applies data science and AI to evaluate approved drugs against diseases systematically. OpenEvidence will contribute proprietary models and biomedical sources and data, including content from its strategic partnership with Nature journals, to help assess repurposing hypotheses.
The announced workflow has four linked functions:
- Every Cure evaluates approved medicines, including generics, against diseases where the nonprofit can pursue development without a conventional commercial return.
- OpenEvidence contributes proprietary models, biomedical sources and Nature journals content for evaluating repurposing hypotheses.
- OpenEvidence matches individual programs with physicians who have cared for patients with the relevant disease.
- During clinical trials, OpenEvidence alerts treating physicians when a patient may qualify and distributes trial findings at the point of care.
The announcement does not report an AI benchmark score or a clinical result. It describes a planned pipeline for testing repurposing ideas, not measured model performance or an established treatment.
Why do specialist physicians matter?
OpenEvidence will match individual repurposing programs with physicians who have cared for patients with the relevant disease. For many rare and underserved diseases, only a small number of these physicians exist, making that access a central part of the collaboration, according to the announcement.
“OpenEvidence can help us reach the physicians who care for these patients,” said Matt Goddeeris, Every Cure’s chief scientific officer.
Goddeeris said the physicians’ firsthand experience would help Every Cure ask better research questions and design studies that address what matters most to patients and physicians. In the announcement’s sequence, AI screens possibilities while physician experience shapes research questions and study design.
That combination addresses two different constraints. Every Cure needs a systematic way to identify candidate uses, while the specialists can test whether a proposed question reflects the disease and the practical problems physicians encounter. Neither role supplies clinical proof on its own.
What reach does OpenEvidence bring?
OpenEvidence already works with the National Organization for Rare Disorders (NORD) on more than 3,000 expert-reviewed rare-disease summaries. The company says its platform is used daily, on average, by a majority of U.S. physicians and is free for all verified U.S. clinicians.
The announcement gives no percentage or denominator for that usage claim, so the partnership’s distribution advantage cannot be quantified from the release. The company nevertheless identifies three specific uses for that clinical network: specialist matching, trial-referral alerts and point-of-care evidence distribution.
The partnership intends to use this national clinical network for three tasks: matching programs to specialists, alerting physicians about possible trial eligibility, and distributing findings at the point of care. The announcement identifies recruitment as a central constraint for rare-disease trials, where eligible patients can be difficult to locate.
How does a hypothesis become a trial?
Once Every Cure identifies a promising opportunity, it remains responsible for generating evidence on whether the treatment is safe and effective in the new disease and for addressing barriers to patient access. OpenEvidence’s role continues as programs move from research into the clinic.
For programs entering clinical trials, OpenEvidence will alert treating physicians when a patient may qualify. The physician, not OpenEvidence, decides whether to pursue a referral. That leaves referral decisions with clinicians while giving rare-disease programs access to a wider pool of potential participants.
As trial results emerge, OpenEvidence will disseminate the findings to physicians at the point of care. The process does not turn a model-generated hypothesis into an accepted treatment without studies designed to test safety and effectiveness.
“Physicians come to OpenEvidence for the best available evidence, but for thousands of diseases, that evidence doesn’t exist yet,” said Isabel Ostrer, OpenEvidence’s director of primary care initiatives.
Ostrer’s statement identifies the limit of reach: distribution cannot substitute for missing trial evidence. Under the partnership, OpenEvidence’s point-of-care role becomes useful only after Every Cure’s evidence work produces findings that clinicians can use.
Why does repurposing remain attractive?
Drug repurposing begins with an approved compound’s established safety data and existing manufacturing capacity, according to the announcement. New-drug development, by contrast, can require billions of dollars and more than 10 years.
For Every Cure, that inherited foundation makes approved medicines plausible candidates for diseases that would not support a conventional commercial launch. The nonprofit can pursue promising indications without relying on a conventional commercial return, according to its partnership rationale.
The existing foundation does not establish efficacy in a new indication. The announcement treats repurposing as an attractive starting point, not a guarantee that a candidate will work or that a clinical study will produce a positive result.
What does the founder’s experience add?
Every Cure says its approach grew from co-founder and CEO David Fajgenbaum’s experience with idiopathic multicentric Castleman disease. After being told that no options remained, he identified an overactive immune pathway in his blood and tested sirolimus, a drug approved to prevent kidney-transplant rejection that targets the same pathway.
Fajgenbaum has been in remission for more than 11 years and has advanced 14 additional repurposed drugs, Every Cure said. That history gives the nonprofit direct experience with the path from an individual diagnosis to a drug selected for a different use.
The announcement cites the 11-year remission and 14 additional drugs as operating experience. It does not name those 14 medicines or report their individual outcomes. Fajgenbaum’s case documents one personal repurposing path; the release cannot support a success rate for the wider portfolio.
What has to happen next?
Every Cure and OpenEvidence define the partnership’s central test: whether AI-assisted hypotheses, paired with specialist knowledge and physician reach, can become credible trial evidence and accessible treatments for patients who otherwise have few options.
For that test to produce evidence, the announced roles require:
- A specific repurposing program must move from systematic drug evaluation into a clinical study, with Every Cure responsible for safety and effectiveness evidence.
- OpenEvidence must match the program with physicians who understand the disease and support physician-led referrals when patients may qualify.
- OpenEvidence must then distribute trial results through its point-of-care network, while Every Cure addresses barriers to patient access.
The announcement provides no named candidate, target disease, trial start date, enrollment target, efficacy threshold, regulatory timetable, budget or funding amount. It also provides no model benchmark against which the partnership’s AI performance could be evaluated.
For the more than 300 million people whom Every Cure says have no FDA-approved treatment for their disease, the next meaningful milestone is a named clinical program that produces credible safety and efficacy evidence and reaches treating physicians. That is the test set by the partnership: credible evidence delivered to clinicians, rather than a larger list of AI-generated proposals.
Source: unite.ai
More from Marcus Bennett
Show full bio
Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
718 articles
Related articles
- Orakl Oncology raises $34 million to scale AI cancer predictions
- OpenAI President Greg Brockman: AI Could Make Death a Choice
- Isomorphic Labs Raises $5B as AI Drug Design Faces Trial Test
- Anthropic CEO Predicts AI Cures Within a Decade; Doctors Push Back
- China's Drugmakers Turn to AI Marketing as Anti-Graft Rules Bite