The Common Health Coalition said on 13 August that it has received $100m from the OpenAI Foundation for an initiative aimed at a specific and unglamorous problem: people who could be cured, and are not.
The target
The Breakthroughs to Follow-Through programme starts with hepatitis C, aiming to at least double cure rates within two years across eight initial states and localities. The first awards go to collaborations in Alabama, Illinois, Louisiana and Massachusetts, with more jurisdictions to be named within six months. Later phases target HIV prevention and curable cancers.
Why hepatitis C
Direct-acting antivirals have cured hepatitis C at very high rates for more than a decade, and most infected Americans still never complete treatment. The failure is not pharmacological. It is finding people, linking them to care and keeping them in it — a data and follow-up problem, which is the kind of problem this money is aimed at.
What the grant actually buys
Not software licences. The funding covers technical assistance, implementation expertise and evaluation for public health departments — the capacity layer that determines whether any tool survives contact with a county health system. Dave Chokshi, a former New York City health commissioner, chairs the coalition.
The clause worth noting
OpenAI Foundation's Anna Makanju said recipients keep "the agency to decide which tools to use — or what not to use." A frontier lab funding AI deployment without requiring its own models is unusual, and it is the difference between philanthropy and distribution.
How to judge it
The programme has set a falsifiable target with a date on it. In two years there will either be a doubled cure rate in eight jurisdictions or there will not — a cleaner test than almost anything else in health AI.