Bunkerhill Raises $55M to Scale Carebricks Across Health Systems

Bunkerhill Health announced the close of a $55 million Series B to expand Carebricks, an agentic AI platform that lets hospitals run AI agents on live clinical data. Investors in the round include Sequoia Capital, Felicis, Optum Ventures, Y Combinator and Khosla Ventures.

Carebricks is designed to let health systems build and operate agents that act on patient charts and clinical databases. The platform aims to move models out of research sandboxes and into production systems that access live clinical records at institutional scale.

Nishith Khandwala, co-founder and CEO of Bunkerhill Health, stated, “Every leading health system has more opportunities to improve patient outcomes than its workforce has capacity to address.” He said the company will use the new capital to broaden clinical and operational use cases and to add governance, monitoring and safety features.

Bunkerhill says Carebricks supports agents for tasks such as cardiology imaging review, prior authorizations and registry maintenance. The company reports deployments at Cleveland Clinic, the University of Texas Medical Branch (UTMB) and Intermountain Health.

UTMB provides the clearest example of the platform in production. Dr. Peter McCaffrey, UTMB’s Chief AI Officer, reported more than 20 agents live on Carebricks. In its first month a coronary calcium detection agent built on an FDA-cleared algorithm flagged a patient judged at imminent risk of heart attack; cardiology confirmed the finding and the patient received a triple bypass. McCaffrey credited the agent with saving the patient’s life and noted the case is a single report, not a controlled trial.

UTMB and Bunkerhill published additional operational results from live use. A nephrology triage agent that prioritizes patients and routes some cases to telemedicine reportedly reduced average specialist wait times by more than 50 percent. A lung nodule agent that tracks incidental CT findings prompted an 80 percent faster response on urgent cases, doubled guideline-concordant follow-up and reduced manual coordinator work.

Those performance figures come from UTMB and the vendor and have not been independently audited. They reflect UTMB’s specific data and staffing conditions and may not be replicated at other hospitals without similar setup and oversight.

Bunkerhill intends to invest the new funding in expanding the range of agents and in governance, monitoring and safeguards. Health systems evaluating deployments must address liability assignment, the cadence for monitoring agent performance, and procedures for resolving disagreements between an agent’s recommendation and a clinician’s judgment before scaling.

Investors cited the platform’s ability to run AI in production and the presence of live deployments as reasons for backing the company. UTMB’s multi-agent footprint provides a reference case for other providers considering agentic AI.