I have enough context from the prompt data. Let me write the debate now.

The Investment Council Debates FICO

Warren Buffett: adjusts his glasses and leans back "Charlie and I have a simple test — if you gave us ten billion dollars and said 'go compete with FICO,' could we do it? I've been thinking about that for weeks, and I keep arriving at the same answer: not a chance. Not in ten years. Maybe not in twenty. And the reason isn't the algorithm — algorithms can be replicated. The reason is that every mortgage underwriter in America, every secondary market investor, every risk model at every bank has been calibrated to FICO scores for three decades. You'd have to convince Fannie Mae, Freddie Mac, thousands of lenders, and their regulators to simultaneously adopt your alternative. That's not a commercial challenge — that's an infrastructure replacement problem on the scale of ripping out railroad gauges. So let's start there. Is this the widest moat any of us has seen in financial services, or am I missing something?"

Charlie Munger: "You're not missing anything — you're just describing it too politely. What FICO has isn't a moat in the traditional sense of brand loyalty or switching costs. It's something rarer and more durable: it's become the unit of measurement for creditworthiness in the United States. Think about that. When Moody's rates a bond, there are alternatives — S&P, Fitch. When FICO scores a consumer, there is functionally no alternative for conforming mortgage lending. VantageScore has existed for nearly two decades, backed by the three credit bureaus themselves — Experian, TransUnion, Equifax — companies with every incentive and every dataset to displace FICO. And VantageScore's share of mortgage scoring is still negligible. That tells you something profound about the depth of institutional embedding. But here's where I apply inversion: what could kill this business? Three things. First, a Congressional mandate forcing GSEs to accept competing scores on equal footing — possible but requires political will that hasn't materialized in twenty years. Second, a paradigm shift where AI-driven alternative data models make traditional credit scores obsolete — possible in theory, but lenders are the most conservative adopters of new technology I've ever seen. Third — and this is the one that keeps me up at night — antitrust enforcement. When a company dismisses lender concerns as 'misplaced, misguided,' as CEO Lansing did on the last earnings call, that's either supreme confidence or the kind of arrogance that attracts regulators."

Dev Kantesaria: "Charlie, I want to address the antitrust point directly because I think it's the most intellectually honest bear case, and also the one most likely to be wrong. I've studied toll booth businesses for eighteen years — Visa, Mastercard, Moody's, S&P Global, MSCI — and every single one of them has faced regulatory scrutiny at some point. Visa and Mastercard had the DOJ settlement. Moody's and S&P were dragged through Congressional hearings after 2008. MSCI faces constant pressure from index fund fee compression. None of them were broken. The reason is that regulators understand something that bears often don't: these businesses exist as infrastructure, and breaking infrastructure creates more harm than the monopoly pricing you're trying to prevent. FICO's case is even stronger than those precedents. The FICO score isn't just used for lending decisions — it's embedded in securitization structures, risk-based capital calculations, and fair lending compliance frameworks. A court-imposed pricing constraint might slow margin expansion, but it won't break the business model. And here's what I think gets underappreciated: the Direct Licensing Program isn't just a pricing mechanism, it's a deepening of the moat. Every lender that signs a direct contract with FICO is one more institution that has formally committed to the FICO ecosystem. They announced four new DLP resellers just this quarter. That's the moat widening in real time."

David Tepper: "I agree with Dev on the moat question, but I want to push on something everyone is glossing over — the mortgage concentration. The Q1 earnings call was very clear: mortgage origination Scores revenue was up 60% year-over-year, and mortgage now represents 42% of total Scores revenue and 51% of B2B revenue. That's not a footnote — that's a structural dependency on one of the most cyclical activities in the American economy. Now, I'm the guy who bought Bank of America preferreds in 2009 when everyone thought the world was ending, so I'm not easily spooked by cyclical exposure. But you have to price it correctly. When I hear Dev say 'buy now' at 37 times earnings, I need to understand what happens to those earnings if mortgage origination volumes drop 30% in a housing downturn. Because that's not a tail risk — that's happened twice in the last twenty years."

Mohnish Pabrai: "David just identified the key tension that makes this interesting. Let me frame it differently. The business itself is close to inevitable — I agree with Dev and Warren on that. But the financial structure layered on top of this business is anything but conservative. Management has tripled debt from $1.05 billion in 2021 to $3.46 billion today. Stockholders' equity is negative $1.75 billion. They bought back $1.415 billion of stock in fiscal 2025 against only $770 million in free cash flow — which means they borrowed $645 million to buy back stock. And here's the kicker: the Q1 buybacks were executed at an average price of $1,707 per share. The stock today is $995. That's management buying at a 72% premium to where we sit today. I'm not saying this destroys the business — the underlying economics are far too strong for that — but it does raise a question about capital allocation judgment that should give any value investor pause."

Robert Vinall: "Mohnish raises a fair point about the buyback pricing, but I want to complicate the narrative because I think the simple version — 'they overpaid for buybacks' — misses what's actually happening. The buyback program has reduced shares outstanding from 31 million to 24 million over the past decade. That's a 23% reduction that has been a primary driver of EPS compounding at 21% annually versus revenue growing at just 9.5%. The question isn't whether any individual quarter's buyback was executed at the 'right' price — it's whether the cumulative program has created per-share value. And the answer is unambiguously yes: EPS has gone from $3.54 in 2016 to $27.50 in 2025. FCF per share from $6.05 to $31.76. The compounding math dominates the timing noise. Now, I share Mohnish's concern about the debt funding of buybacks in excess of FCF — that converts a self-funding program into a levered bet. But the business generates $770 million in annual free cash flow at 47% operating margins with $9 million in capex. The debt service is $162 million at 5.2% weighted average cost. Interest coverage at 7x EBIT is adequate, not fortress-level, but adequate for a business with this level of revenue predictability."

Pulak Prasad: "I want to step back from the financial engineering debate and return to something more fundamental, because I think we're spending too much time on capital structure and not enough on evolutionary fitness. I study businesses through a Darwinian lens — which ones survive and thrive through adversity, and which ones break. FICO passed the ultimate stress test in 2008. The worst credit crisis in modern history — the very thing that should have destroyed a credit scoring company — actually reinforced FICO's dominance. Lenders didn't abandon FICO scores in the crisis; they relied on them more heavily. Regulators didn't say 'the scoring model is broken'; they said 'we need better data inputs into the FICO framework.' That's the sign of a keystone species — one that the ecosystem depends on so thoroughly that even a catastrophic shock strengthens its position. But here's my concern with the current incarnation of this business: management has optimized for growth-phase conditions. Negative equity, 3.5x debt-to-EBITDA, buybacks exceeding FCF — these are the choices of an organism that assumes the environment will remain favorable. In nature, organisms that over-optimize for one environment become fragile when conditions shift. I'd feel more comfortable if management carried a modest cash buffer and held debt below 2.5x EBITDA."

Warren Buffett: "Pulak, that's a beautiful analogy, and it connects to something I've been thinking about — Guy Spier's 'away from desk' test. Could I own FICO for twelve months without checking the price, reading an earnings call, or seeing a single headline, and sleep well? For the business, absolutely — the toll booth collects its royalty whether I'm watching or not. But for the balance sheet, I'm less certain. The negative equity and rising debt mean this is a business where the operating economics are magnificent but the financial structure requires monitoring. A pure toll bridge — I'd never think about it. A levered toll bridge — I'd want to check the debt covenants once a quarter. That distinction matters to me."

Charlie Munger: "Warren's right to draw that distinction, and it leads me to the AI disruption question, which I think we need to address honestly rather than dismissively. People ask whether AI could replace the FICO score. Here's my answer: the probability that AI fundamentally disrupts FICO's core mortgage scoring business in the next decade is somewhere between 5% and 15%. Not zero — alternative data models using bank transaction data, rent payments, and real-time income verification are genuinely more predictive for thin-file borrowers. The Plaid partnership Lansing announced this quarter is FICO's hedge — incorporating cash flow data into an enhanced UltraFICO Score. But here's why I put the probability low: credit scoring disruption isn't a technology problem, it's an adoption problem. Every new model needs to be validated against decades of default data, approved by regulators, accepted by the GSEs, integrated into securitization frameworks, and adopted by thousands of lenders simultaneously. The QWERTY keyboard isn't the best layout either — we've known that for sixty years. But the switching costs are so high that the installed base is permanent. FICO is the QWERTY keyboard of consumer credit."

Dev Kantesaria: "Charlie's QWERTY analogy is exactly right, and I want to build on it because it illustrates why I view FICO as the single most structurally inevitable business in my portfolio — more inevitable than Visa, more inevitable than MSCI. With Visa, there's at least a theoretical path to disruption through real-time payment rails or central bank digital currencies. With FICO, there isn't even a theoretical path that doesn't require rebuilding the entire American credit infrastructure. And the addressable market is expanding, not contracting. Score 10T incorporates trended credit data for better predictive accuracy. The UltraFICO Score with Plaid adds cash flow data. The Mortgage Simulator creates new revenue streams from the same institutional relationships. Management is doing exactly what a toll booth operator should do: raising the toll, widening the road, and adding lanes. To David's point about mortgage cyclicality — yes, 42% of Scores revenue is mortgage-related. But even in a severe downturn, mortgages still get originated. The 2008-2010 period saw origination volumes drop roughly 40% from peak, and FICO survived without breaking a sweat. At today's revenue base of $2 billion and 47% operating margins, a 15-20% Scores revenue decline would reduce operating income by perhaps $150 million. Painful? Yes. Existential? Not remotely."

David Tepper: "Dev, you're right that it's not existential, but you're wrong to dismiss it as merely 'painful.' Let's do the math. FICO has $162 million in annual interest expense. If operating income drops $150 million from a mortgage downturn — your own estimate — pre-tax income falls from roughly $760 million to $610 million. EPS drops from $27 to maybe $21-22. At the current 37x multiple, the stock would trade at $780-$815. And if the multiple compresses — because levered businesses with negative equity get punished harder in downturns — you're looking at 25-28x on depressed earnings: $525-$615. That's 35-47% downside from here. Now, I'd buy that hand all day — that's the kind of asymmetric setup I live for. But buying at $995 when that scenario has, let's say, a 25% probability over the next three years? The expected value math works, but it's not a screaming buy. It's a reasonable buy. There's a difference."

Warren Buffett: "Let's transition to the financial record, because I think the numbers tell an extraordinary story that either validates or contradicts everything we've just said about business quality. When I look at FICO's ten-year trajectory, one number jumps off the page: operating margins went from 19% in 2016 to 47% today. That's a 28-percentage-point expansion on a business that was already profitable. And it wasn't driven by cost-cutting — revenue nearly tripled from $881 million to $2 billion. That combination — tripling revenue while more than doubling margins — is the financial signature of a monopoly that discovered it could charge more."

Robert Vinall: "Warren, that margin trajectory is the single most important data point in this entire analysis, and I want to unpack why it happened because it tells you everything about the reinvestment runway ahead. Before 2018, FICO operated at 17-20% operating margins — respectable but unremarkable for a software company. Then something changed: management began systematically repricing the FICO score. B2B score prices increased year after year, and because the marginal cost of delivering a score is effectively zero — the algorithm runs on bureau infrastructure — every dollar of incremental pricing dropped straight to operating income. ROIC went from 12% in 2016 to 58.5% in 2025. That's not financial engineering; that's a business discovering and monetizing a monopoly. The question for the next decade is whether there's room for another 10-15 points of margin expansion — which would take operating margins toward 55-60%, the theoretical ceiling for an asset-light data business — or whether political and commercial resistance caps margins near current levels. My instinct says we're in the seventh or eighth inning of margin expansion, but the per-share compounding can continue through buybacks even if margins plateau."

Mohnish Pabrai: "Robert, I appreciate the compounding math — I really do — but I want to apply a cold shower of valuation discipline to this conversation because I think several people are falling in love with the business and forgetting to ask what it costs. FICO earns $27.50 in EPS. At $995, that's 36x earnings. Free cash flow is $31.76 per share. At $995, that's a 3.2% FCF yield — below the risk-free rate of roughly 4.25%. Stock-based compensation runs at $157 million, or about $6.50 per share. If I subtract that from FCF, owner earnings are closer to $25 per share. At $995, my owner-earnings yield is 2.5%. I need situations where I'm risking a dollar to make three. At these prices, even in my optimistic scenario — EPS grows to $40 in three years, stock re-rates to 33x — I get to $1,320. That's 33% upside against the downside scenario David just laid out of $525-$615. My heads-to-tails ratio is roughly 1:1. That's not my game. I need it cheaper."

Charlie Munger: "Mohnish is being disciplined, and I respect that, but I want to challenge one element of his math. Subtracting the full $157 million in SBC from free cash flow double-counts the dilution impact if management is simultaneously reducing shares through buybacks. The honest way to assess this is to look at net dilution. Shares went from 31 million in 2016 to 24 million in 2025 — a 23% reduction. That means buybacks are more than offsetting SBC issuance. Per-share FCF has grown from $6.05 to $31.76 — a 425% increase, or 18% annualized. If you're an owner, what matters is per-share value creation, and that number is unambiguous. Now, I still want a margin of safety because of the leverage and mortgage cyclicality — I agree with Mohnish on the direction of the adjustment, just not the magnitude."

Pulak Prasad: "Let me add a data point that I find remarkable. FICO spent $8.9 million on capital expenditure in fiscal 2025. Against $2 billion in revenue. That's a capex-to-revenue ratio of 0.45%. I have studied hundreds of businesses across decades, and I cannot find another company of this scale with capital intensity this low. Even Visa and Mastercard spend roughly 5-7% of revenue on capex. Moody's spends 2-3%. FICO's asset-light model is structurally different from every comparable. And it's because the scoring algorithm runs on infrastructure owned by the credit bureaus — FICO is, in effect, a royalty on someone else's capital expenditure. This is why ROIC has expanded from 10% to 58% while the capital base has barely grown. The implication for long-term compounding is extraordinary: virtually every dollar of incremental revenue converts to free cash flow available for buybacks or debt reduction."

Warren Buffett: "All right, let's get to brass tacks. The stock is at $995. Everyone in this room agrees the business is exceptional. The question is whether the price is right. Dev, you've been the most bullish — make your case for buying at today's price."

Dev Kantesaria: "My framework is simple: I value businesses on forward FCF yield relative to the risk-free rate, with an adjustment for the certainty of the cash flows. FICO generates $31.76 in FCF per share today. At 15% growth — below its five-year CAGR of 21.9% — FCF per share reaches $36.50 next year and $42 the year after. On a $995 cost basis, my Year 2 FCF yield is 4.2%, surpassing today's risk-free rate. By Year 5, I'm earning a 6.4% yield on my cost basis from a business with structural inevitability. The compounding math works at this price. Not spectacularly — I'd prefer $850 — but the opportunity cost of waiting is real. This business compounds EPS at 20% annually. Every year I sit in cash earning 4%, the intrinsic value moves $150-200 higher. Buffett himself says he'd rather buy a wonderful business at a fair price than a fair business at a wonderful price. FICO at $995 is a wonderful business at a fair price."

David Tepper: "I'm with Dev on buying now, but for different reasons. I see an asymmetric setup. The stock is down significantly from its highs — the market is pricing in deceleration, regulatory risk, and margin peak all at once. But look at Q1 FY2026: revenue up 16%, Scores up 29%, non-GAAP EPS up 27%. This business is accelerating, not decelerating. If Q2 guidance comes in above expectations — which the catalyst intelligence suggests is the top catalyst — the stock re-rates to $1,100-$1,300. That's 10-30% upside in a few months. My downside, even in the stress scenario, is roughly 25-35% from here, and I think that scenario has maybe a 20-25% probability over three years. Expected value is positive. I'm buying."

Robert Vinall: "I share Dev and David's enthusiasm for the business but not their urgency on price. I would start accumulating below $880. That's roughly a 20% discount to my fair value estimate of $1,100, which I derive from a ten-year DCF: $770 million in FCF growing at 13% annually, discounted at 9%, with a 25x terminal multiple. That gives me $26.5 billion in equity value, or roughly $1,100 per share after subtracting net debt. The 20% margin of safety at $880 compensates for the mortgage cyclicality risk and the leverage. I'd apply the away-from-desk test here: at $880 with a 3.6% FCF yield growing at 15%, I'd be comfortable owning this without looking at it for a year. At $995 with a 3.2% yield below the risk-free rate, I'd be checking quarterly. That's the difference."

Mohnish Pabrai: "My number is $830, and let me explain why with a specific scenario. At $830, the FCF yield is 3.8%, approaching the risk-free rate. More importantly, in my downside case — mortgage volumes drop 25%, Scores revenue falls 15%, EPS compresses to $21 — the stock at 28x trough earnings would be worth roughly $590. That's 29% downside. In my upside case — 18% EPS growth for three years to $45, at 30x — the stock is worth $1,350. That's 63% upside. My heads-to-tails ratio at $830 is roughly 2:1. Not the 3:1 I'd prefer, but adequate for a business of this quality. At $995, the ratio is closer to 1:1. I'm patient. I can wait."

Charlie Munger: "I'd start buying at $860. My reasoning is inverted — I start with what could go wrong and price accordingly. The two material risks are a housing downturn compressing Scores revenue by 15-20% and an antitrust settlement that constrains future pricing. If both happen simultaneously — maybe a 10-15% probability — EPS could trough at $18-20 and the stock could see $550-600. I need enough margin of safety that even in that scenario, I haven't permanently impaired capital. At $860, my downside to the worst case is roughly 30-32%. That's uncomfortable but survivable. And the business quality means the worst case is temporary — housing recovers, FICO reprices, and you're back to $1,000+ within three years. The asymmetry at $860 is acceptable. At $995, I'm buying the same business with 15% less cushion, and that 15% is the difference between a good investment and an exceptional one."

Pulak Prasad: "I'd begin accumulating at $870, which provides an 18-19% margin of safety below my fair value estimate of $1,080. I arrive at that fair value by weighting three approaches: normalized FCF times 33x for a business with a 30-year survival track record through multiple credit cycles, forward EPS of $33 times 33x, and an EV/EBIT framework at 28-30x. The average converges around $1,050-$1,080. My discount reflects the leveraged balance sheet — if management were carrying net cash instead of $2.9 billion in net debt, I'd buy ten percent higher. But I do want to say this: I would be comfortable owning FICO for twelve months without checking the price. The toll booth collects regardless. The business doesn't need me watching. It's the debt that introduces the variable I'd want to monitor, and even that is manageable at 7x interest coverage. This is a business I'd want to own through a cycle, not around it."

Warren Buffett: surveys the room "Let me try to pull together where this council has landed, because I think this has been one of the more illuminating discussions we've had. On the qualitative side, we reached something close to unanimity — and that's rare for this group. FICO isn't just a strong business; it's a structural monopoly embedded so deeply in the American credit system that displacing it would require the coordinated action of Congress, regulators, the GSEs, thousands of lenders, and the entire securitization market simultaneously. Charlie's QWERTY analogy captures it perfectly: the installed base is the moat, not the algorithm. VantageScore has had twenty years and the full backing of all three credit bureaus, and it still can't crack the mortgage market. That tells you the switching costs aren't just high — they're effectively infinite for the use case that matters most."

"Where we diverge — and this is where the honest disagreement lies — is on whether the current price adequately compensates for two specific risks. First, the mortgage concentration: 42% of Scores revenue tied to a deeply cyclical activity, which David quantified as potentially $150 million in operating income at risk in a downturn. Second, the leverage: $3.46 billion in debt, negative stockholders' equity, and a management team that has funded buybacks in excess of free cash flow by borrowing — including $1.4 billion in buybacks last year at an average price of $1,707, well above today's $995. Mohnish and Charlie rightly point out that this converts a fortress into a levered bet."

"Dev and David make the compelling counter-argument that waiting has a real cost when a business compounds EPS at 20% annually. Dev's forward FCF yield math — showing the yield on a $995 cost basis surpasses the risk-free rate within two years — is intellectually honest and directionally correct. But five of us, myself included, believe the margin of safety at $995 is insufficient given the cyclical exposure and financial leverage. We'd start buying in the $830-$880 range, which provides a 15-20% cushion and a more attractive FCF yield approaching the risk-free rate. At those prices, FICO becomes what I'd call a 'buy-and-forget' holding — a toll booth you own through cycles without losing sleep. At $995, it's a business I admire enormously but would want to own at a better price. Five to two in favor of buying lower — and reasonable people, as this discussion proves, can disagree."