Universal Health Services, Inc. UHS

Key data

The business

UHS runs two different businesses under one roof. About 56% of revenue comes from acute-care hospitals, the general medical centers where people go for surgery, childbirth and emergencies, concentrated in Nevada, Texas, California and Florida. The other 44% comes from behavioral health: psychiatric hospitals and addiction and residential treatment centers, several hundred facilities across the US and the UK. The acute side is bigger and more cyclical. The behavioral side is smaller, higher-margin and hard to replicate, because you cannot open a psychiatric hospital quickly and the beds stay full once you do. A large slice of both is paid by government: Medicare, Medicaid, and the supplemental "directed payments" states layer on top of base Medicaid rates. That last bucket is the swing factor, and Washington just put a clock on it.

The numbers

Growth is real and steady, not spectacular. Revenue has compounded about 8% a year since 2021, from $12.6B to $17.4B, while earnings per share compounded roughly 18%, from under $12 to $23. Two levers stacked on top of revenue: operating margin recovered from the 2022 trough, and the share count shrank about 23% as the company bought back stock.

Q2 2026 was the largest quarter the company has ever printed: revenue $4.64B, up 8.3% on the year, diluted EPS $5.98, up 10.1% and a few cents ahead of the $5.94 consensus estimate. The number that jars is net income attributable to UHS, up only 1.5% year on year against 8.3% revenue growth, because other operating expenses climbed 16%. The profit growth shareholders saw came from the buyback, not from the operations. That gap is the thing to watch.

The buyback is the engine, and at this price it is a good one.

Operating cash flow ran $1.86B in 2025 against $1.02B of capital spending, leaving roughly $850M of free cash, and the company spent $968M repurchasing stock, about 9% of today's market value in a single year. Buying your own earnings back under 7 times is worth far more than buying them at 15, and this team has done it consistently. Behavioral segment margin is the tell: consolidated operating income held, $517M in Q2 against $503M the quarter before, and EPS kept climbing while the count shrank, so the better half has not cracked, though this filing does not break out the segment to confirm it line by line. What this memo believes that the tape does not: the directed-payment cut is real but slow, phased over years from 2028, and the price treats it as if it lands all at once. The print that settles it is 2027 revenue guidance.

Management

Insiders are net sellers, but trivially so: five sales totaling $1.2M over the last year, the largest an $820,000 sale by director Warren Nimetz in November, no open-market buys, plan status not disclosed on the Form 4s, so read them as routine rather than as signal. The real capital story is the buyback, funded from cash flow, not leverage. Long-term debt sits at $4.9B, about 1.8 times EBITDA, interest covered 13 times, and both Moody's and S&P rate the company investment grade. In August it issued new 2031 and 2036 notes, ordinary refinancing. The all-cash Talkspace deal at $5.25 a share tucks a telehealth therapy platform into the behavioral franchise. Pay is tied to earnings and the guidance record has held. This is a team compounding at low prices, which is exactly what you want and rarely see.

How it fails or surprises you

The directed-payment reprice. The 2025 budget law caps state directed payments at 100% of Medicare in expansion states and 110% in non-expansion states, and from state fiscal 2028 grandfathered plans shrink by up to 10% a year until they reach the Medicare rate. That supplemental money flows straight to margin. First sign: 2027 revenue and margin guidance, and any state rate notice before then.

Costs outrunning revenue. Net income attributable grew 1.5% in Q2 while revenue grew 8.3%, because other operating expenses jumped 16%. Operating margin was 11.1% this quarter against 11.7% a year ago. Two more quarters of that drift and the buyback is papering over a real problem, and the cheap multiple is cheap for a reason.

The discount closes (right tail). At 6.9 times earnings UHS is cheaper than at any point in the twelve years the record covers, which bottomed near 9, against peers near 22. If the payment cut proves as gradual as the law reads and EPS holds in the $24 to $26 range, a re-rate to just 10 times is roughly +45%, and the buyback shrinks the count into it. The market is paying for the fear, not the arithmetic.

Closing thoughts

The 2027 guidance and the quarterly margin prints settle this. The downside is legislated and gradual, directed payments phase down from 2028 at 10% a year, and the company throws off enough cash to keep buying back stock while it plays out. What's genuinely at risk if cost drift and the payment reprice compound is a few points of acute-care margin, call it a tenth of earnings, not the franchise. Against that sit the behavioral business and a 7-times multiple. Judgment, not a probability: the odds favor the buyer more than the price implies.

The bet is still that UHS keeps its acute and behavioral beds full and keeps getting paid for them, and that the extra Medicaid money holds up longer and cuts shallower than a 7-times multiple assumes. What breaks it is a faster, deeper directed-payment reprice than the law now schedules. The one pair of numbers that tells you first: operating margin against revenue growth, quarter over quarter. If margin stops falling while revenue keeps compounding, the discount was a gift. If margin keeps sliding as those payments phase down, the market was right to worry.