THE COMPANY-BUILDING FIELD NOTEBOOKRESEARCH EDITION / SEPTEMBER 2026
Startup
Research.
Search
POST-MORTEM LIBRARY / AI-era failures (2023–2026)

Forward Health

  • Company — Forward
  • Sector — Digital health / primary care
  • Founded / Died — 2016 / November 2024
  • Lifespan — ~8 years
  • Capital raised~$650M total, including a $100M Series E in November 2023 (Founders Fund, Khosla Ventures, SoftBank, General Catalyst); peak valuation reported ~$1B+
  • Peak scale — 19 physical clinics at peak; a reported ~$149/month membership base in the tens of thousands; ~200 employees. Revenue never disclosed.
  • What it built — First a chain of technology-heavy primary care clinics with staff doctors; then CarePods — unstaffed automated diagnostic booths placed in malls and offices, where a patient would self-administer tests guided by AI.
  • Stated cause of death — Founder Adrian Aoun said the company shut down after failing to scale CarePods and cited the difficulty of the model. No detailed public explanation was given.
  • Evidenced causeDivergence, and unusually instructive. Reporting established that CarePods had frequent hardware problems — including patients getting stuck inside — that the clinical scope was narrow, and that a pod could not perform the examinations primary care consists of. Forward spent the majority of $650M solving an automation problem in a domain where the binding constraint was clinical liability, reimbursement and trust, not throughput. The pivot from staffed clinics, which had real patients, to pods was a capital-intensive bet against the part of the business that worked.
  • Warning signs visible earlier — CarePods launched November 2023 to immediate technical criticism; the company shut down twelve months later. Forward was cash-pay only, never solving reimbursement, which caps the addressable market to affluent urban customers.
  • Money outcome — Investors lost most of ~$650M, including $100M invested twelve months before the end. Members lost access and, in some cases, continuity of care — the most serious customer harm in this library after the fraud cases. Employees laid off.
  • Post-mortem qualityPoor. Brief statements; no analysis.
  • Transferable lesson — Identify the binding constraint before you automate. Automating the non-binding constraint is the most expensive way to discover which one it was.
  • SourcesFierce Healthcare · Modern Healthcare · TechCrunch on the CarePod launch

Read the wider evidence

This entry is reproduced from the supplied research, with its inline source links retained. It has not been independently re-reported for this website conversion.

Read the complete chapter, source list, and methodological notes →
← Back to post-mortems