Ortet is now public, and it starts with $500M. The money comes from Thoreau, an outfit that describes its own work as building compute, governance rules and research infrastructure for the health sector.
Running the New York company is Kyunghyun Cho, its co-founder and chief executive. His name sits on two ideas that became load-bearing for modern AI: the attention mechanism and the gated recurrent unit, the machinery underneath today’s transformer models.
Cho’s career has run through molecular design. He spent five years on an AI platform that paired wet-lab experiments with models, and a company called Prescient Design grew out of that work. Genentech bought it and kept him on to set up its frontier research group. He also holds a chair in health statistics at New York University.
The thesis is that narrow tools fail. A model trained on a single task with a thin slice of data cannot learn how a patient’s biology, treatment, outcomes and system interactions connect. Ortet wants a shared foundation instead, one that keeps improving and carries knowledge from scientific discovery into clinical care, operations and billing.
Thoreau says the aim is combining compute and data infrastructure with frontier research and deployment, all shaped around the individual patient. The founding team blends AI research, health, life sciences and large-scale infrastructure work.
Why it matters: health systems have spent years buying point solutions. One model that reasons across a full patient record, from labs and imaging to claims and outcomes, would be a different kind of product and a far harder one to validate. Ortet has not named clinical partners or published benchmarks.
