Even though AI tools are changing medicine, they are also being used against it. To fight back, one startup is raising a lot of money.
You can't ignore the numbers. Healthcare fraud cost the U.S. federal estimates that the system will be worth $100 billion in 2025, and more and more of it is being built with artificial intelligence. Synthetic patient identities, AI‑generated clinical documentation, and automated billing manipulation have outpaced the detection tools that hospitals and insurers built in a pre‑AI era.
Into that gap steps MedShield, a Boston‑based healthcare security analytics firm that announced a $95 million Series C round this week, led by Andreessen Horowitz's bio fund, with participation from UnitedHealth Group Ventures, Intermountain Health, and the Mayo Clinic Platform. MedShield is now worth $580 million, which is a lot of money for a compliance software company five years ago, but it seems almost too small when you think about how big the problem it's trying to solve is.
What MedShield Does
Founded in 2016, the company spent its first decade building a proprietary dataset of claims patterns, provider behavior, and patient record anomalies across dozens of health systems. That dataset now feeds a machine learning platform that screens claims in real time, flagging suspicious activity before the system processes reimbursement rather than after.
The distinction matters enormously. Audits happen months after payments are made, and recovery rates on false claims rarely go above 20%. This means that fraud in healthcare is usually found after the fact. MedShield's platform intercepts at the point of adjudication, allowing payers to pause and investigate before money leaves the system. The company says it currently screens over 800 million claims annually for more than 500 hospital networks, health plans, and government programs across 28 states.
Why Institutional Investors Are Paying Attention
It's not a coincidence that UnitedHealth Group Ventures and Intermountain Health are involved; it's a sign. Large payers and provider systems don't typically back compliance vendors unless they intend to deploy the technology at scale. Both organizations have indicated they plan to deepen their integration with MedShield's platform as part of broader AI governance initiatives.
The timing aligns with mounting regulatory pressure. In early 2026, the Centers for Medicare and Medicaid Services released new guidance that said health plans over a certain size had to show they could detect fraud in real time by 2028. This effectively set a deadline for the whole industry to comply.
What the Money Is For
MedShield plans to deploy the new capital across three priorities: expanding its AI model to cover pharmacy benefit fraud, a fast‑growing vector; building integrations with electronic health record platforms; and hiring in its federal government division, where contracts with the Department of Defense and Veterans Affairs are already in advanced discussions.
The broader message is straightforward. As AI lowers the cost of generating fraudulent medical claims, the institutions responsible for catching them need AI infrastructure that scales at the same pace. MedShield thinks the race will never end and that someone will win it.
