UnitedHealth Group

Some analyses are about buying blood in the streets. This one is about the morning after — because by the time you read this, the blood has largely dried. UnitedHealth Group is the largest healthcare company in the United States and one of the largest companies in the world by revenue: $447 billion in 2025, roughly one dollar in every eight the country spends on health. It is two businesses fused into one bloodstream: UnitedHealthcare, the nation's biggest health insurer (~48 million members), and Optum, which owns the doctors, the data, and the pharmacy-benefit machine that sit on the other side of the insurer's own claims. For a decade it was one of the great compounders in America, turning $6-7 a share into $630 by November 2024.

Then it broke — spectacularly, and largely by its own hand. Medical costs blew through the prices it had set; a botched expansion at Optum compounded the damage; and in the space of a year operating margins were cut in half (from 8.7% to 4.2%) and earnings collapsed. The stock fell ~63% to $235 by mid-2025. Along the way its insurance-unit CEO was shot dead on a Manhattan sidewalk, a federal criminal investigation opened into its Medicare billing, and it became, by most polls, the most hated company in America. The founding architect of Optum, Stephen Hemsley, was pulled out of the chairman's seat to run the company again — and, five days into the job, bought $25 million of stock with his own money. This week's second-quarter results, reported just before we published, confirmed the repair is real: adjusted profit nearly a third above expectations, the medical-cost ratio falling hard, full-year guidance raised. The operational crisis is ending. The investment question is entirely different: at $423, up ~80% from the lows, is there anything left on the table?

"Be fearful when others are greedy, and greedy when others are fearful." — Warren Buffett

Berkshire Hathaway lived that maxim on this exact stock — and its trade is the most instructive fact in this report. It bought UNH in mid-2025, in the depths of the fear, a ~$1.6 billion position revealed that August. Then, in the first quarterly filing of the Greg Abel era this spring, it sold every share. Berkshire bought the blood, banked the bounce, and walked — before the criminal probe was resolved, before the recovery was complete. When the most patient capital in the world treats a name as a trade rather than a forever-holding, an owner should ask why. The answer runs through this entire analysis: the greatness is real, but so is the reason even Buffett's firm wouldn't marry it.

The history is essential here, because it tells you the collapse was a cost-and-conduct crisis, not a demand crisis — and that this company has turned scandal into its greatest era before.

Two lessons. First — the reassuring one — UNH has done this before: the 2006 backdating scandal preceded its greatest decade, and the man who led that recovery, Hemsley, is the one leading this one, with his own capital on the line. Second — the sobering one — the 2025 collapse exposed how much of this company's economics depend on things other than serving customers well: aggressive Medicare risk-coding (now under criminal investigation), an insurer that owns its own doctors and grades its own claims, and a denial-heavy reputation so toxic that a CEO's murder drew cheers instead of only grief. The business is being repaired. The questions the collapse raised — about how durable those economics are once the regulator and the public are done with them — are not.

The machine, in five steps:

This is the honest heart of the matter, and it is why even a repaired UNH cannot be valued with the confidence of a Coca-Cola. A majority of this company's economics ride on administered rates — Medicare Advantage and Medicaid, set by the government. In January 2026 a single CMS 'advance notice' proposing a near-zero rate update cut $80 billion of market value in one session; in April, the final notice came in far more generously and handed it back. The business model is simple to describe; the earnings power is a function of Washington's decisions, of a criminal probe's outcome, and of a social licence the 2025 crisis showed to be dangerously thin. A Buffett-style buyer can underwrite the insurance and the pharmacy toll. What no one can underwrite is the regulator's next document or the prosecutor's next move — and those, not underwriting, are what move this stock now.

Two halves, one bloodstream (FY2025 shape):

The flywheel — and why critics call it a conflict machine. A premium dollar enters UnitedHealthcare, which pays Optum's own physicians (often a fixed fee per patient) and its own pharmacy manager for the drug benefit — so profit is captured twice or three times inside the same corporate wall. To bulls, this is aligned incentives and a data feedback loop no competitor can replicate: the insurer that owns the doctors can manage cost better than one that doesn't. To critics — and, more consequentially, to the Department of Justice — it is an insurer that decides its own claims, sets its own drug prices, and grades its own homework, and the criminal probe reportedly covers precisely how UnitedHealth reimburses its own physicians and codes its Medicare diagnoses. The genius and the legal jeopardy are the same structure. And note what the 2025 collapse did not touch: revenue kept growing all the way through ($372B → $448B). Americans did not stop needing healthcare. The company simply mispriced the cost of providing it — a fixable error, now being fixed.

UNH's moat is, in pure business terms, one of the widest we have examined — and, uniquely on our shelf, it is being attacked not by competitors but by its own government and its own reputation. Three walls:

1 · The scale trifecta. No company in world healthcare holds more than one of these crowns; UNH holds all: the largest US insurer (~48M members), the largest employer of physicians (~10% of the entire US doctor workforce, via Optum Health), and a top-three pharmacy-benefit manager (~23% of all US prescriptions) — plus Optum Insight's claims-processing layer, which even competitors run on. Trend: intact.

2 · The data flywheel. Claims, clinical, pharmacy and physician-workflow data at population scale, feeding pricing and care management — the "compounding information asset" a value investor recognises as a right-of-way. Trend: widening.

3 · Regulatory complexity as a barrier. Operating compliantly across Medicare, Medicaid in dozens of states and 50 insurance departments requires fixed-cost scale that locks out entrants. Trend: stable — but double-edged, because the same government is the biggest customer and the biggest threat.

The crucial distinction. Note what did not happen in 2025: no competitor took UNH's share. Humana, CVS/Aetna, Elevance and Centene all bled on the identical Medicare-cost cycle; the whole oligopoly retreated together. That is the proof the competitive moat held and the crisis was cyclical, not structural. But the moat is under a different kind of siege: from the regulator (CMS rates, a DOJ criminal probe), from the legislature (live bills to break up the PBM-insurer structure), and from the public (a majority of Americans now tell pollsters they would support ending private health insurance). The width of the moat is not in question. What is in question is the toll rate the political system will permit it to charge. I score the moat a 7: exceptionally wide, and under political review.

The operational question — will margins recover? — has essentially been answered: yes, and fast. The medical-cost ratio has fallen from 88.9% to 86.7%, this week's guidance was raised toward ~$19.50–$20.00 of adjusted earnings, and CMS handed the industry a favourable 2027 rate in April. Health insurance is a repricing business: costs surged faster than the annual pricing cycle could adjust, and now the cycle has caught up. The bull case is playing out in the numbers. So the investment question is not "will it recover" — it's "how much of the recovery is already in the $423 price, and what protects you against the tail?"

Our read is the one the whole report has been building toward, and it is unsentimental: this is a wonderful business, genuinely repaired — bought at the wrong time. The fat pitch was $235 in May 2025, or $283 in the January panic, when the fear was maximal and the margin of safety enormous. At $423, up ~80%, the recovery is largely priced in, the DCF reads overvalued, and the one thing that could still hand you a second fat pitch — a bad turn in the criminal docket — is the one thing no model can price. Buffett's own firm demonstrated the correct playbook here: buy the terror, sell into the relief, and don't stay for the trial. We arrive too late to buy the terror. So the honest posture is to admire the franchise, respect the repair, and watch the docket for the next dislocation rather than pay up for a recovery someone braver already banked.

The competitive table tells a clear story: UNH's rivals are no threat to its moat — its government is. The single most important fact of the 2025 collapse is that everyone bled together: Humana targeting a mere ~3% Medicare margin, CVS/Aetna closing ~90 plans across 34 states, Elevance and Centene all repricing and retreating on the identical cost cycle. No competitor gained a yard of UNH's share. That is the definitive evidence that the crisis was a cyclical cost shock plus self-inflicted Optum missteps, not a competitive moat breach. The threats that matter sit in the bottom row — CMS setting the rates, the DOJ running a criminal probe, Congress drafting bills to unwind the PBM-insurer structure, and an electorate that has made attacking health insurers a bipartisan applause line into the midterms. You are not underwriting a competitive contest. You are underwriting American health politics — which is precisely why the multiple should be, and now is, lower than the compounder once commanded.

The management story is a genuine strength; the ownership story is a genuine warning.

Capital allocation is exemplary — and honestly complicated. Even in its worst year UNH generated $16 billion of free cash flow, never lost money, never cut the dividend (a 17th consecutive raise in 2026), and returned over $13 billion to shareholders. The balance sheet carries real weight — $78 billion of debt, $131 billion of goodwill from decades of Optum acquisitions, net debt ~2.2× EBITDA — but the cash easily services it. The genuinely hard part for a values-minded owner is the ethical dimension, which is also a financial one: UnitedHealthcare's prior-authorisation denial rates were repeatedly found among the industry's highest, and part of the historical margin was extracted through friction imposed on sick people. The 2026 trust campaign — eliminating prior authorisation on 30% of services, pledging 100% pharmacy-rebate pass-through, rebating ACA profits, a $1 billion foundation — is both genuine reform and margin the company is voluntarily handing back to buy back its legitimacy. An honest owner must model permanently thinner "friction margins" from here. I score management a 6: a superb, aligned repair job, on a business whose past profitability and present legal jeopardy are uncomfortably intertwined.

Read these numbers as a story of a wound and a scab. The wound is the margin line: operating margins run razor-thin by the nature of insurance (a "good" year is single digits), and in 2025 they halved as costs outran prices — dragging GAAP EPS from $23.86 to $13.23. The scab is the estimate line: analysts see earnings climbing back to $18.49 next year and $25 by 2028, essentially recovering the old peak, as the repricing flows through. The constant, through it all, was cash: even in the worst year the company generated $16 billion of free cash flow and never touched the dividend. The honest caveat is the balance sheet — $131 billion of goodwill from a decade of Optum deal-making, against which a razor-thin-margin business carries $78 billion of debt — which means the reported returns flatter the underlying capital intensity, and a structural squeeze on those "friction margins" would bite a leveraged model harder than a screen suggests. The recovery is real. It is also, at today's price, no longer a secret.

Here is what makes UnitedHealth the odd one out among this summer's fallen angels. For Novo Nordisk the DCF said +73%; for IBM, +85%; for Nike, +14%. For UNH, the automated DCF reads $275 — thirty-five percent below the price. It is the only stock on our entire board where the model says the market is paying too much, not too little. The analysts agree from the other direction: a mean target of $451 is barely 7% above the quote, and the most bearish published target, $373, sits below it. This is not a mystery — it is the arithmetic of a recovery that has already happened. At $235 last May, the stock traded at perhaps 15 times a trough $16 of adjusted earnings, with the fear at its peak; a genuine fat pitch. At $423, up ~80%, it trades at ~21 times the recovering forward number, with the operational crisis largely resolved and only the un-modellable risks — the criminal probe, the political toll — left to hang over it. The margin of safety that once justified the risk has been spent. On the earnings the company will actually produce by 2028 (~$25), today's price is fair, not cheap — roughly 17 times, for a wide-moat franchise with a fat legal tail. Fair value on a wonderful business is a perfectly respectable place to be. It is simply not, for a value investor, a reason to act — especially when the same business was on sale, screaming, fourteen months ago.