Pomegra
Baba, a New York startup pairing AI agents with human care teams for Medicare and Medicaid patients, raised a $40M Series A led by FirstMark in October 2026.
Key Takeaways
- Baba raised a $40M Series A led by FirstMark, with CVS Health Ventures and Valtruis joining.
- The round follows a $6.5M seed from February 2026, eight months earlier.
- Baba serves 10,000+ patients, has a 100,000+ family waitlist, and did not disclose a valuation.
Lead
Baba, a New York company that combines human care coordinators with AI agents to manage care for Medicare and Medicaid patients, announced a $40 million Series A on October 20, 2026. FirstMark Capital led the round. CVS Health Ventures and value-based care specialist Valtruis participated, alongside returning investors General Catalyst, Genius Ventures and Ground Up Ventures. Baba did not disclose a valuation.
The company says it supports more than 10,000 patients and has a waitlist above 100,000 families. It plans to spend the money on clinical care teams, AI research and partnerships with health plans.
What Does Baba Actually Do?
Baba assigns each patient a dedicated care team of care coordinators, licensed clinical social workers and therapists, then uses proprietary AI agents to handle the administrative work around them. That work includes syncing specialist referrals, coordinating hospital-to-home discharges, arranging in-home aide support, fighting billing denials and locating community resources such as housing and transportation help.
The model competes with the same tangle it targets: fragmented care navigation services, health plan case management programs and the family caregivers who now do much of this work unpaid. Baba sells to national commercial and Medicare Advantage plans, according to the company's announcement, which makes the plan, not the patient, the paying customer.
Founder and CEO Connor Sweeney framed the economics around cost. "By dramatically lowering care operation costs, we enable economically viable delivery for historically expensive-to-serve populations," he said.
Why Did FirstMark and CVS Back It?
FirstMark partner Rick Heitzmann described Baba as the firm's "latest bet on the consumerization of healthcare," aimed at Americans who are "rural, older and have more complex needs." That thesis treats high-cost, high-need members as an underserved consumer segment rather than a utilization problem.
For CVS Health Ventures, the draw is the human-in-the-loop design. Carter Prince of the CVS venture arm said that "training on specific long-horizon activities and keeping humans deeply in the loop" is what could help more patients get care. CVS owns Aetna, a large Medicare Advantage insurer, so a corporate investor with plan exposure has a plain interest in tools that cut the cost of coordinating complex members.
What Does the Round Imply About the Last One?
The Series A arrives eight months after Baba emerged from stealth in February 2026 with a seed round of more than $6.5 million led by General Catalyst. A roughly sixfold step up in that time suggests the seed milestones were met quickly, or that investors are paying for a position in a crowded category before margins are proven.
The announcement offers volume, not unit economics. Patient counts and a waitlist are not revenue, retention or medical-cost savings per member. Health plans typically want outcome data before they commit contracts at scale, and Baba has not published any.
What Is the CMS Connection?
In June 2026, the Centers for Medicare & Medicaid Services selected Baba for its 10-year ACCESS Model, which the CMS Innovation Center runs. Baba was chosen across all four clinical tracks: early cardio-kidney-metabolic conditions, advanced cardio-kidney-metabolic conditions, musculoskeletal health and behavioral health.
Participation gives Baba a federal pathway to reimbursement-linked work, and a long runway. It also ties the company's revenue to a government program whose design and payment rules can change over a decade.
How Is Baba Measuring Its AI?
Baba released BabaBench, a benchmark for evaluating AI agents on long-horizon patient advocacy tasks such as insurance calls, billing appeals and prior authorizations. The paper is on arXiv and was accepted to a NeurIPS workshop, with privacy safeguards for protected health information built into the tests.
A self-authored benchmark favors the company that wrote it. Still, publishing task definitions gives outside researchers something to test, which is more than most healthcare AI vendors offer.
The company also added Alicia Tennenbaum, formerly of Mount Sinai, as executive director of clinical operations, and designer Alan Li, formerly of Ramp.
Outlook
Baba now has $40 million, a federal program slot and a corporate investor from the insurer-and-pharmacy world. The open questions are commercial: whether health plans will pay for the service at scale, and whether AI-driven administration can keep care-team costs low enough to serve patients that plans consider expensive. Plan contracts, outcome data and the first full year of ACCESS Model participation will show whether the waitlist converts into a durable business.
