S02.7 · Healthcare & Life Sciences
US hospital and physician care worth roughly $2.4T combined, with AI attacking a 25-30% administrative overhead.
Healthcare providers and physician services — hospital systems and physician practice groups — represent roughly $1.5T in US hospital care and roughly $900B in physician and clinical services (2023 actuals, CMS National Health Expenditure, the standard reimbursement-linked benchmark), with growth projected at 5-7% forward per CMS Office of the Actuary projections on aging demographics and utilization. The economics split starkly — thin at the hospital level, materially better in scaled physician groups — and the largest controllable cost pool in the segment, administrative overhead, is precisely what AI is now attacking directly.
CMS NHE growth ran 4-6% historically; the Office of the Actuary projects 5-7% forward. The US structure — a fee-for-service/value-based mix layered on an employer/Medicare/Medicaid payer split — has no real analogue in the single-payer systems of the EU and Canada, where provider consolidation is far less M&A-driven. Fragmentation is the defining fact: thousands of health systems and physician groups operate independently. PE-backed practice roll-ups in dermatology, orthopedics and gastroenterology remain a small share of the market, but that share is growing quickly.
Device, pharma and diagnostics suppliers sit upstream; payers and patients downstream. Hospital operating margins run a thin 2-4%; scaled physician groups run 15-25% EBITDA. Both live under Medicare fee schedules, commercial payer contracts and site-of-service reimbursement differentials — where a procedure is performed can matter as much to the economics as what the procedure is.
Where agentic AI lands first is clinical documentation, coding and billing, and prior-authorization workflows — all of it inside an administrative overhead estimated at roughly 25-30% of total US healthcare spending. That is one of the largest addressable cost pools in the entire sector, so even partial automation moves real money. It also erodes a specific moat: large health systems have absorbed reimbursement complexity through scaled administrative staff that smaller practices could not afford, and automation shrinks that advantage. Documentation and coding automation is already underway; AI-driven clinical decision support at scale runs a longer horizon, 5-10 years, because anything touching clinical judgment directly carries higher stakes and heavier regulatory scrutiny.
Two strategic moves dominate, with opposite economics. The site-of-care shift toward ambulatory surgical centers captures margin as procedures move to lower-cost, higher-margin outpatient settings — new economics, not defended ones. Payer-provider vertical integration — providers taking on risk-bearing functions historically held by insurers — is about controlling risk and reimbursement rather than growing volume. Behavioral health integration is the further adjacency, connecting this segment to specialty and post-acute care.