The Legendary Debate

Warren Buffett: sets down his reading glasses "Healthcare is the one industry where your biggest customer is also your regulator. The U.S. government sets the price for roughly eighty percent of UNH's premium revenue — Medicare rates, Medicaid rates, ACA rules — and then simultaneously investigates whether UNH is too powerful. That's the tension at the heart of this business. So let me pose it plainly: if we owned the whole thing privately and could never sell it, would we sleep well at night for the next twenty years? That's where I want to start."

Charlie Munger: "I'll invert your question, Warren. Instead of asking whether we'd sleep well owning it, let me ask: how does this business die? I see three scenarios, and only one scares me. Single-payer healthcare would eliminate the entire private insurance model — but that's about as politically likely as Congress voluntarily cutting its own pay, so I'll set it aside. The second is that medical cost inflation permanently outpaces the company's ability to reprice premiums. That's happened for three consecutive years now, and it's concerning, but there's a natural corrective mechanism — UNH sets premiums annually and can shed unprofitable members, which is exactly what they're doing with the four-to-five million members they're deliberately dropping in 2026. That tells me the repricing mechanism still functions, even if it works with a painful lag. The third — and this is the one that makes me genuinely uncomfortable — is the DOJ forcing structural separation of UnitedHealthcare and Optum. That would be like telling Berkshire it can't own both GEICO and the railroad — it would destroy the very architecture that makes the sum greater than the parts. I can't probability-weight that outcome with any confidence, and that uncertainty is why my conviction is only a five out of ten."

Dev Kantesaria: "Charlie, I want to challenge your premise that the repricing mechanism is a moat. It's not — it's a survival mechanism, and there's a world of difference. Let me explain what I mean. When Moody's raises its rating fees three to five percent annually, issuers pay without blinking because the alternative — issuing bonds without a Moody's rating — carries a thirty-to-fifty basis point borrowing penalty on billions of dollars of debt. Even if someone offered ratings for free, issuers would still pay Moody's. That is a toll booth. The toll is set by market forces, and the position is structurally permanent. Now consider UNH. CMS sets Medicare Advantage rates through a political process that has delivered below-trend increases for three consecutive years running. Management's own language on the earnings call was almost confrontational — Tim Noel said the 2027 advance notice 'simply doesn't reflect the reality of medical utilization and cost trends.' When your CEO is publicly begging the government to set a more realistic toll rate, you do not have pricing power. You have a government contract. And four-to-five million members leaving in 2026 proves the position isn't permanent either — unlike Moody's, where no issuer can credibly walk away, UNH's customers switch plans every annual enrollment period."

David Tepper: "Dev, I hear you, but you're describing a structural constant and calling it a fatal flaw. CMS has set Medicare rates through a political process for the entire history of Medicare Advantage — and during that entire history, UNH compounded revenue at eleven percent annually for thirteen consecutive years without a single decline. The government-set pricing is the water this fish has always swum in, not some new development. What's actually changed is that medical costs accelerated from five-to-six percent to seven-and-a-half to ten percent, creating a temporary lag against repricing. I've seen this exact pattern in distressed credit: a high-quality issuer hits a liquidity squeeze, the market prices in default, and then the business reprices and recovers while bondholders earn thirty percent in twelve months. That's the setup here. Revenue grew twelve percent in 2025 — twelve percent! — even as operating income dropped forty-one percent. The top-line engine is fully intact. What broke is the margin structure, and margins in managed care are repriced annually. This isn't a structural decline — it's a repricing lag being mistaken for permanent impairment."

Mohnish Pabrai: "David, I don't disagree with your cyclical-versus-structural argument — you may well be right that margins recover. My objection is mathematical, not philosophical. This company has a two-hundred-fifty-billion-dollar market cap. For me to achieve a three-to-one asymmetric return — which is the minimum I require before committing capital — UNH would need to reach seven hundred fifty billion. Name me a scenario where a managed care company with three-to-five percent net margins reaches seven-hundred-fifty-billion-dollar valuation. Even your most optimistic recovery — twenty-five dollars in EPS at twenty times, roughly four hundred fifty billion — gives less than two-to-one from here. I rejected Penske Automotive at eighteen times cash flow when peers traded at eight times, and Penske is arguably the finest auto dealer on Earth. Quality doesn't excuse the math. At this size, the risk-reward simply doesn't work for my framework. I'd rather deploy the same capital into a two-billion-dollar cyclical at trough where tripling is plausible."

Robert Vinall: "Mohnish, your math is valid for your framework, but let me explain why the opportunity still interests me despite my reservations on price. What drew me in is the management situation. Hemsley is the architect who built Optum from nothing — he took UNH from a hundred-billion-dollar insurer to a two-hundred-fifty-billion-dollar integrated platform during his first tenure, growing EPS from roughly five dollars to eleven dollars while expanding ROIC from twelve percent to fifteen percent. His return fits a pattern I've seen work before: Schultz coming back to Starbucks, Iger returning to Disney. The common thread is that these are founders or founder-equivalents who understand the institutional DNA deeply enough to fix what professional managers broke. And what broke at Optum Health is specific and fixable — eighteen different electronic medical record systems, structurally unprofitable contracts that required a six-hundred-million-dollar reserve, a provider network that had expanded twenty percent beyond what could be efficiently managed. Conway is consolidating to three EMR systems, narrowing the network by twenty percent, streamlining risk membership by fifteen percent. These are the kind of blocking-and-tackling operational improvements that build moats through execution, which is what my framework prizes above all — moat as the output of execution, not the input."

Pulak Prasad: "Robert, I share your respect for Hemsley, but I'm troubled by something the financial data reveals that the qualitative narrative glosses over. The operating income trajectory — thirty-two billion in 2023, thirty-two billion in 2024, nineteen billion in 2025 — shows a forty-one percent collapse in the organism's energy output while it continued consuming more resources. Four hundred forty-eight billion in revenue, up from three hundred seventy-two billion. When a dominant organism takes in more resources but produces dramatically less energy, evolutionary biology has a term for that: metabolic mismatch. And the quarterly EPS trajectory through 2025 — from six-ninety-one in Q1 to three-seventy-six in Q2 to two-fifty-nine in Q3 — shows accelerating deterioration, not stabilization. Add eight C-suite changes in ten months, and what I see is an organization under severe adaptive stress with its institutional knowledge being deliberately disrupted at exactly the moment it needs stability. I've studied organisms that survived existential threats — Coca-Cola, Nestlé India, HDFC Bank — and the common denominator is that leadership remained stable through the crisis while making targeted surgical changes. UNH is doing the opposite: wholesale leadership replacement during a multi-front assault. That pattern precedes organizational failure more often than recovery."

Warren Buffett: "Pulak raises a genuinely uncomfortable point about the leadership instability, and I want to give it proper weight. Eight C-suite changes in ten months is not normal. But I'd argue the analogy isn't the organism losing its immune system — it's more like a hospital bringing in a new surgical team because the old one was botching operations. Hemsley didn't just change people randomly; he brought in Wayne DeVeydt from Elevance as CFO, someone who ran the finances at UNH's most direct competitor for a decade. That's not panic — that's deliberate recruitment of people who understand the specific patient on the operating table. Now, let me address what the financial data tells us, because the numbers both support and complicate the qualitative story."

PHASE 2: FINANCIAL EVIDENCE

Warren Buffett: "The single most important financial fact about UNH is this: revenue compounded at eleven-point-one percent for thirteen consecutive years — from a hundred-two billion in 2011 to four hundred forty-eight billion in 2025 — without a single year of decline. Through the financial crisis tail, through COVID, through the cyberattack, through this earnings collapse. I cannot think of another business processing half a trillion dollars in annual revenue that has achieved that consistency. That is the financial signature of a genuine franchise. But I want to be honest about what else the numbers show. Operating income fell from thirty-two billion to nineteen billion in a single year. Verified free cash flow has been genuinely volatile — thirteen-and-a-half billion in 2023, three-point-seven billion in 2024, eleven billion in 2025, and negative two-point-three billion in 2022. You cannot call that predictable by any definition I use. The more stable anchor is operating cash flow — nearly twenty billion in 2025 at one-and-a-half times net income — which for an insurer of this scale better represents the franchise's cash generation than the FCF line, which is distorted by insurance investment portfolio activity and acquisition spending."

Charlie Munger: "Let me raise the elephant in the room that everyone's dancing around: the balance sheet. Total debt went from forty-six billion in 2021 to seventy-eight billion in 2025 — a seventy-percent increase in four years. During that same period, net income fell from seventeen-point-seven billion to twelve-point-eight billion. In 2024 specifically, the company issued seventeen-point-eight billion in new debt while simultaneously spending nine billion on gross share buybacks and seven-and-a-half billion on dividends. Let's call that what it is: borrowing money to return capital to shareholders during an earnings downturn. That's not disciplined stewardship — that's the prior management team juicing per-share metrics with leverage while the underlying business deteriorated. The paused buyback in Q4 2025 wasn't prudent foresight — it was forced necessity. At seventy-eight billion in debt with depressed earnings, they simply ran out of room. I want to own this business, but I want to own it with a management team that treats the balance sheet as a fortress, not a piggy bank."

Dev Kantesaria: "Charlie's capital allocation analysis reinforces exactly why this fails my framework. The seventy-five-point-seven billion in cumulative acquisitions from 2016 to 2024 produced what? Six-hundred-twenty-five million in lost contract reserves for structurally unprofitable relationships. Eight hundred million in cyberattack write-offs from the Change Healthcare acquisition. Eighteen different EMR systems that had to be consolidated to three. The ROIC expansion from eleven to seventeen percent that Warren celebrates was fueled by aggressive acquisition-driven growth on borrowed capital — and now we're seeing the hangover. My holdings — Moody's, FICO, Visa — generate thirty-to-fifty-percent ROIC with minimal debt and zero acquisition dependency. They compound organically because the toll booth does the work. UNH's ROIC required seventy-five billion in acquisitions and thirty-two billion in new debt to achieve. That's a fundamentally different — and more fragile — compounding model."

David Tepper: "Dev, you're looking at the same ROIC trajectory I am and drawing the opposite conclusion. ROIC expanded from ten-point-seven percent to seventeen percent while the capital base tripled. Most acquisitive companies see ROIC dilute as they deploy capital — UNH achieved the opposite. That's actually the strongest evidence the integrated flywheel works. Yes, some acquisitions were messy — Change Healthcare's cybersecurity was inadequate, and Optum Health expanded too fast. But the Catamaran acquisition built Optum Rx into a top-two PBM processing one-point-four billion prescriptions annually. That's a genuine competitive asset. The question isn't whether the prior team made mistakes — they clearly did — it's whether the franchise value underneath those mistakes is intact. And the answer is yes: eight hundred new PBM relationships won during the worst earnings year in company history. Clients are voting with their contracts even as the stock drops."

Robert Vinall: "I want to address the profitability question head-on because it determines the entire valuation debate. Operating margins went from eight-point-seven percent in 2023 to four-point-two percent in 2025. Some of that — roughly four billion in charges — is demonstrably one-time: the cyberattack reserves, the restructuring, the lost contract reserve. Stripping those out, underlying operating margin is probably closer to five-and-a-half to six percent — still well below the historical eight-to-nine percent band, but not as catastrophic as the headline suggests. Management is guiding for eighty-eight-point-eight percent MCR in 2026, which represents only thirty basis points of improvement from the 2025 figure of eighty-nine-point-one percent. That tells me the repricing is working incrementally, not dramatically. Full margin recovery to eight percent probably extends into 2027 or 2028. The honest answer is that normalized EPS power is probably twenty to twenty-two dollars, not the twenty-four to twenty-six the most bullish estimates assume, because some of the medical cost inflation may prove persistent."

PHASE 3: VALUATION AND VERDICTS

Warren Buffett: "So let's put a price on it. At $275.59 today, you're paying roughly twenty times trough GAAP earnings of fourteen dollars. But that's misleading — you're also paying about thirteen-and-a-half times the twenty-dollar-and-change operating cash flow per share, and roughly sixteen times what I'd estimate as normalized owner earnings of sixteen-to-seventeen-dollars per share once you strip SBC from the FCF figure. For a franchise that has generated fourteen-to-seventeen percent ROIC for a decade and still produces twenty billion in operating cash flow during its worst year, those are fair-to-attractive multiples. I'd begin accumulating at current levels — small, two percent of portfolio — and add more aggressively below two-forty where the margin of safety genuinely widens. My conviction is a six because the DOJ risk is real and unquantifiable, and the balance sheet isn't the fortress I prefer."

David Tepper: "I'm more aggressive than Warren because the asymmetry math works at current prices. Bear case: operating margins stay permanently compressed at five-to-six percent, giving you sixteen-to-eighteen-dollar normalized EPS at thirteen times — roughly two-twenty to two-thirty-four. That's fifteen percent downside. Base case: margins recover to seven-to-eight percent by 2028, producing twenty to twenty-two-dollar EPS at sixteen times — three-twenty to three-fifty. That's twenty to twenty-seven percent upside. Bull case: full recovery to eight-plus percent margins with DOJ resolution, twenty-four-dollar EPS at eighteen times — four-thirty-plus. The risk-reward skews heavily toward upside, and the April 2026 Q1 results provide a near-term catalyst to confirm or deny the margin recovery thesis. I'd size this at four percent immediately."

Mohnish Pabrai: "David, your base case of twenty-to-twenty-seven percent upside over two-to-three years is a fine return for most investors. But that's roughly ten percent annualized, which is what the S&P 500 delivers without any of these specific risks — DOJ, medical cost inflation, leadership instability. I need situations where the risk-reward is dramatically asymmetric: risk one dollar to make three to ten dollars. At two-hundred-fifty billion market cap, the upside is capped by the sheer size of the company. I'd need UNH at a hundred-ten dollars or below — a hundred-billion-dollar market cap — before the asymmetry works for me. Until then, I'll admire this from the sidelines and deploy capital elsewhere."

Dev Kantesaria: "I'll reiterate my categorical position. Despite my medical training — perhaps because of it — I understand that healthcare economics are fundamentally unpredictable over the ten-year horizon I require. The CMS rate-setting process means you're investing in a business where the single largest pricing input is determined by political negotiation, not market forces. My portfolio — Moody's, Visa, FICO, MSCI — consists of businesses where the toll is mandatory, permanent, and market-determined. UNH's toll is government-set, periodically inadequate, and customers demonstrably leave when prices shift. No price makes this investable for me."

Pulak Prasad: "I'll acknowledge this organism has extraordinary survival fitness — thirteen consecutive years of revenue growth through multiple crises is genuinely rare. But at $275.59, my conservative fair value of roughly three hundred dollars provides only nine percent upside — nowhere near the margin of safety I need for a healthcare franchise with an active DOJ investigation, seventy-eight billion in debt, accelerating quarterly earnings deterioration, and the most disruptive leadership transition in the company's history. I want to see this organism demonstrate adaptation — two consecutive quarters of operating margin improvement — before I commit capital. At two-twenty to two-thirty, with evidence of stabilization, I'd be genuinely interested."

Robert Vinall: "I'm in, but barely. Hemsley's return is the specific catalyst that tips me from 'wait' to 'small position.' At $275.59, my fair value of three-ten to three-thirty gives me roughly twelve to twenty percent upside — tighter than I'd prefer for my fifteen percent annual hurdle rate. I'll take a one-and-a-half percent position because the management quality justifies participation, but I'll add meaningfully only below two-fifty where the margin of safety becomes adequate, or upon DOJ resolution, which would remove the single largest source of uncertainty in the thesis."

Charlie Munger: "I'll buy a small position at current prices — one-and-a-half percent — because the inversion test says the probability of permanent capital loss is genuinely low. But I want to be clear about what concerns me: the debt. When you owe seventy-eight billion dollars and your earnings just fell forty-three percent, you're not in a fortress — you're in a building with a cracked foundation. The paused buyback tells me management knows it. I'd double my position below two-thirty, where even the most pessimistic scenarios — eighteen-dollar normalized EPS at thirteen times — provide downside protection at two-thirty-four."

PHASE 4: SYNTHESIS

Warren Buffett: surveys the room "Let me try to pull this together, because this was one of the more genuinely divided discussions we've had. On the qualitative side, we have broad agreement on one critical point: the integrated payer-provider-pharmacy-technology model that UNH built is unique in American healthcare — no competitor has replicated it despite a decade of well-funded attempts, and the demographic tailwind of ten thousand daily Medicare entrants makes the demand side of this business as close to inevitable as anything in investing. Amazon tried to enter with Haven and failed. That tells us something real about the durability of the competitive position, even if the current financial trajectory is painful.

Where we divide is on three specific questions. First, Dev and Mohnish raise legitimate structural objections — Dev argues this isn't a toll booth because CMS sets the pricing, and Mohnish argues the math doesn't work at a quarter-trillion-dollar market cap. I respect both positions, though I disagree: the repricing mechanism has worked through every prior cycle, and twelve-to-fifteen percent annualized returns are attractive for this quality level even if they don't meet Pabrai's three-to-one threshold. Second, the balance sheet genuinely worries all of us — seventy-eight billion in debt accumulated through aggressive acquisitions during a period of declining profitability is exactly the kind of leveraged overreach that Munger and I have warned about for decades. Third — and this is what makes the situation genuinely interesting — the operating cash flow of nearly twenty billion dollars in the worst year proves the franchise cash generation engine is intact underneath the charge-depressed earnings.

The four of us who would buy — myself, Charlie, David, and Robert — do so with modest sizing and genuine humility about the DOJ uncertainty. At $275.59, we're paying a fair price for a franchise business at cyclical trough, with the April 2026 Q1 results serving as the critical confirmation point. If operating margins show recovery toward six-to-seven percent, the market will begin repricing this toward three-twenty-plus. If they don't — if the margin compression proves structural and the DOJ imposes meaningful restrictions — then Dev and Pulak and Mohnish will have been right to wait, and we'll have a small, manageable loss cushioned by the three-point-two percent dividend yield. The honest answer is that reasonable, intelligent people disagree on this one, and the data to resolve the disagreement arrives in April. Position accordingly — small, patient, with clear exit triggers — and let the business prove which camp is right."