Forward Health promised to deliver 3,200 AI-powered “CarePods” within a year. It deployed five. In November 2024, after raising $657 million in total funding, the company shut down every location and cancelled patients’ scheduled visits. It is one of the most expensive failures in tech-enabled healthcare — and one of the most instructive, because almost everything that killed it was visible in the business model from day one.

The diagnosis

Forward was founded in 2016 by Adrian Aoun, a former Google executive. The original product was a chain of membership primary-care clinics operating outside insurance. The pivot that defined — and ended — the company came in late 2023, when Forward raised a $100 million Series E from backers including Khosla Ventures and SoftBank to build the CarePod: a self-contained, AI-powered kiosk where patients could get blood drawn, throat swabbed, and blood pressure read for a $99 monthly membership, with no doctor or nurse present.

The vision was a doctor’s office you could mass-produce and drop into a mall. The unit economics were the opposite of mass-producible. Each CarePod reportedly cost over $1 million to build, and the machines routinely failed to deliver the promised experience — blood draws didn’t work, patients got stuck inside, and the self-serve diagnostics couldn’t handle the messy reality of frontline care. Forward had planned to roll out 3,200 CarePods across 2024; it deployed only a handful. Having raised $657 million and never found a sustainable model, it ran out of road. Aoun cited operational complexity and the inability to reach profitability.

Why it matters

Forward made three compounding bets, each of which raised the difficulty of the next. It bet on hardware (capital-intensive, slow to iterate, unforgiving of defects). It bet on removing the clinician entirely (turning every clinical edge case into a product failure rather than a human judgment call). And it bet on doing both at consumer-retail scale before proving the model worked once. A miss on any single bet is survivable. Stacking all three meant the company needed near-perfect execution on an unproven machine, in a regulated domain, with a cost base of seven figures per unit.

The seductive error was treating healthcare delivery like a software-deployment problem — build the platform once, then scale copies at near-zero marginal cost. But a CarePod’s marginal cost was over a million dollars and its marginal reliability was poor. The “AI replaces the practitioner” framing inverted where the risk actually sat: in a clinic, a nurse absorbs the thousand small exceptions that automated diagnostics can’t. Remove the nurse and every exception becomes a refund, a complaint, or a safety event.

The Charaka View

Forward belongs to a category Manthan Intelligence’s postmortem database tracks closely: AI ventures where the technology was real but the deployment economics were fatal. Our analysis of why a majority of healthcare AI agents fail in production found the recurring killer is not model accuracy — it’s the gap between a controlled demo and the uncontrolled, liability-laden reality of care. Forward is that gap rendered in steel and $1-million kiosks.

The transferable lesson is a sequencing rule. When a business stacks hardware risk, automation risk, and scale risk, it must retire them one at a time — prove the pod works and is safe and is economic at a handful of sites before promising thousands. Forward inverted the order: it raised and committed to scale first, then discovered the unit didn’t work. Capital can buy you time to find product-market fit. It cannot buy you a working product you haven’t built yet — and at over a million dollars per unit, it buys that time very, very fast.

This analysis draws on Fierce Healthcare’s reporting on Forward’s shutdown, TechCrunch’s CarePod launch coverage, Fortune on Forward’s Series E and founder background, Healthcare Huddle’s failure analysis, and eMarketer’s analysis of the failure. Human editorial oversight applied.

This analysis is informational and does not constitute investment advice, a research report, or a recommendation to buy, sell, or hold any security.

Charaka Notes by Manthan Intelligence. Subscribe