Ending Rare Disease.

Rare disease has been the structural exception in drug development for four decades. Conventional pharmaceutical economics require each program to recover its cost against an addressable market large enough to justify the labor it consumes, and rare disease patient populations are by definition too small to clear that threshold. Approximately 300 million people live with one of more than 10,000 rare diseases worldwide, a population comparable in size to the United States. Fewer than one in ten of those conditions has an approved therapy. The unmet need is the largest in medicine, and the rare disease community has pursued cures for decades against economics that have not allowed the field to deliver them.

Those economics have changed. Agentic artificial intelligence has collapsed the per-hypothesis cost of drug discovery toward zero, and the Food and Drug Administration has opened a regulatory pathway for individualized therapies that do not fit the conventional randomized controlled trial model. The Plausible Mechanism Framework draft guidance issued February 23, 2026 accepts mechanism-anchored evidence as substantial evidence of effectiveness for genetic conditions with known biological cause, on the condition that high-quality natural history data support the mechanism. Taken together these two shifts make rare disease addressable in a way it has not been for forty years.

One constraint remains. Agentic research and the Plausible Mechanism Framework both require a data layer that is longitudinal, consented, reusable across programs, and provably compliant. The existing deidentified health data economy, dominated by aggregators including IQVIA, Datavant, Truveta, Komodo Health, and Tempus AI, does not supply such a layer for rare disease. The commercial aggregators serve common diseases where scale is achievable through back-end tokenization of records that patients never consented to monetize. Rare disease requires the opposite posture: direct, consented, patient-level longitudinal relationships that survive the exit of any individual sponsor and that compound in value as successive programs reuse the same evidence base. The consent failures exposed by the 23andMe bankruptcy and by the Lacks family's settlement with Thermo Fisher have made the inadequacy of the existing model visible to regulators, state attorneys general, and the public in ways that will not reverse.

Cureledger is the data layer the new regime requires. Patients contribute their health data into a Life Data Trust, a legal trust whose trustee administers the data under terms the patient sets and under a fiduciary duty the patient can enforce. The data is de-identified by default and subject to a conditional re-identification path that the patient controls and the trustee operates, used only to inform the patient of developments relevant to them. Revenue from authorized reuse flows back to contributors pro rata. Smart contracts and cryptographic proofs render the trustee's operations verifiable to contributors, regulators, and data purchasers without requiring any party to trust Cureledger itself. The architecture is purpose-built for rare disease and generalizes without modification to the broader drug development data economy, a market whose commercial ceiling is materially larger than the rare disease vertical alone.

Cureledger's thesis is that operational agentic AI, the Plausible Mechanism Framework, and a fiduciary-grade data layer are the three components of the next regime of drug development. Two are arriving under their own momentum. The third requires purpose-built infrastructure, and that infrastructure is what Cureledger delivers. Rare disease is the first proof. The larger market follows.