Financebotresearch desk研究台

FICO › analyze

FICO · Analyze

WATCH Technology

2026-09-06 · $932.26 · Technology / Software–Application (credit-score analytics) · United States

One line: A near-perfect franchise — 85% gross margin, 54% operating margin, ~$870M FCF, a decades-old credit-score standard — whose single growth engine of the last two years (mortgage-score pricing power, ~43% of revenue and ~100% of recent growth) had its regulatory mandate revoked two days ago. The stock is down 53% from its $1,998 high to a fresh 52-week low, which prices some of this. But FY2026 earnings are inflated by both a mortgage origination up-cycle and a monopoly price spike, both now at risk at once — so the ~$35 TTM EPS may be a peak, and "27x trailing" is cheaper than it looks only if that base holds. WATCH, conviction 5.5 — a great business at an honestly two-sided price. The reason to wait is a live, unquantified threat, not a doubt about quality.

A portfolio-specific passage was removed from the public build.

1. The event that defines the analysis

On 2026-09-04, FHFA Director Bill Pulte ordered Fannie Mae and Freddie Mac to approve VantageScore 4.0 for all lenders, effective immediately — ending the four-month, 50-lender pilot and, with it, FICO's ~decades-long monopoly as the only credit score accepted for conforming mortgages. Key facts:

Fact Detail Read
VantageScore GSE share already >9% of Fannie/Freddie securitizations by early Sept (from the pilot) adoption is real, not hypothetical
Pulte's framing accused FICO of raising per-score cost 1,800% since 2020 politically charged; price is the target
FHFA next step "seriously considering bi-merge scoring" (2 bureaus instead of 3) would cut score pulls ~⅓ on top of competition
Cost gap cited ~$2,000 to FICO per 100 mortgage apps vs ~$99 VantageScore the price umbrella the bear case targets
Independent study Milliman found FICO 10T more predictive than VantageScore 4.0 every vintage the bull's quality defense
FICO's own response cut FICO Score 10T to $0.99/score + $65 funding fee via direct licensing (−90% wholesale) FICO will defend on price — and cut out reseller markups

The single most important interpretation: the threat is overwhelmingly to price, not (yet) to volume. CEO Will Lansing, post-ruling: "not seeing volume loss," with potential share loss "in the 20s" (percent). Lenders pull both scores; the mortgage stack, GSE guidelines, and model validation are deeply embedded and slow to switch. But FICO's revenue explosion came from the price umbrella — and that umbrella is what just lost its legal support. FICO cutting its own 10T to $0.99 is the tell: the days of raising mortgage-score prices 40%+ a year are over, whoever keeps the volume.

2. Fundamentals — a superb business, but on a base inflated by the very thing now at risk

FICO's fiscal year ends Sept 30. TTM below = Q4 FY25 + Q1–Q3 FY26 (rebuilt from quarterly_income_stmt, not a single annual column).

Quarter (fiscal) Revenue Op. Income Net Income Diluted EPS
Q4 FY25 (Sep-25) $515.8M $248.1M $155.0M $6.42
Q1 FY26 (Dec-25) $512.0M $234.0M $158.4M $6.61
Q2 FY26 (Mar-26) $691.7M $402.5M $264.5M $11.14
Q3 FY26 (Jun-26) $674.2M $362.6M $237.2M $10.45
TTM (sum) $2,393.6M $1,247.2M $815.1M ~$34.6

FY2026 guide raised to $2.53B (+20% YoY) — driven, in management's own words, by "mortgage market resilience" and "higher mortgage origination score unit pricing."

Segment split (Q3 FY26) — this is the whole story:

Segment Q3 rev YoY Note
Scores $458.9M +41% ~88% operating margin; mortgage = 63% of Scores, ~43% of total company revenue
Software $215.0M +2% Platform ARR $413M (+62%) now exceeds legacy for the first time; Software ARR $816M (+10%)

Scores is ~68% of revenue and essentially all the growth; mortgage is ~63% of Scores. The Software leg — the part immune to the FHFA ruling — grew just 2% in reported revenue. Its Platform ARR is genuinely inflecting (+62%), but at ~$215M/qtr and mid-single-digit reported growth it is nowhere near large or fast enough to offset a mortgage-Scores hit in the near term.

Health scorecard

Metric Value Read
Gross margin 85% ✅ elite — Scores is ~88%
Operating margin 54% ✅ among the best in software
FCF (FY25 → FY26E) $739M → ~$870M ✅ but boosted by the mortgage boom
FCF 3yr CAGR 13.7% understates the recent mortgage-driven spike
Shares outstanding 21.6M, −2.3%/yr ✅ aggressive, consistent buybacks
Net debt ~$2.94B ($3.07B debt − $134M cash) ⚠️ real leverage, funded the buybacks
Book equity −$1.75B 🚩 negative from buybacks — Graham/P/B are VOID, see §5
Interest coverage OCF $779M vs interest $134M ✅ ~5.8x, no distress

Capital allocation: FICO is a buyback machine — $1.41B repurchased FY25, $821M FY24, ~$1.1B FY22 — funded partly by debt, driving equity negative. This is deliberate financial engineering, not distress: retire ~2–3% of shares a year on 85%-gross-margin cash flow. It works beautifully while earnings grow; it is riskier if earnings roll over, because the debt stays and the EPS tailwind from buybacks shrinks.

3. Moat — a two-tier moat where the more valuable tier just got breached

The Scores moat is a regulatory + standard-setting intangible, historically the strongest kind: FICO was written into the GSE selling guides, so every conforming mortgage had to buy a FICO score. That is the tier the FHFA just opened. The non-mortgage Scores moat is intact — credit-card, auto, personal-loan, and prescreen uses run on FICO by decades of convention, lender model validation, and consumer familiarity (myFICO). VantageScore has been "approved" in various channels for years and gained little outside where a regulator forced the door.

Adversarial stress test. A rival cannot rebuild FICO's brand, its validated model history, or its embedding in thousands of lender workflows. What a rival can do — and just did — is get a regulator to mandate optionality, which converts a price-maker into a price-taker in the one vertical (mortgage) where FICO had pushed price hardest. The Milliman "10T is more predictive" finding is a genuine defense if lenders choose on quality; the $2,000-vs-$99 gap is why they may not.

Revenue-stream map: (1) Mortgage Scores — highest-priced, now contested; (2) non-mortgage Scores — huge, durable, growing low-double-digits; (3) Software/Platform — decisioning + FICO Platform, ARR inflecting to platform, structurally lower margin, the long-term second leg.

Disruption forecast. Beyond VantageScore: the FHFA bi-merge idea would cut score pulls ~⅓ regardless of vendor; and long-horizon, AI-native underwriting could de-emphasize a single three-digit score. None of that kills FICO, but the era of "the score is a mandated toll you raise at will" is ending. Evergreen: the franchise, yes; the mortgage-pricing growth model, no.

Moat rating: 6.5/10 — would be 8+ without the mortgage breach; the non-mortgage franchise and the Platform inflection hold it up.

4. Type-specific overlay

Not a dividend grower (no dividend), REIT, or BDC. Closest to a high-multiple compounder with a regulatory-cycle overlay — §1–§3 and the FCF valuation below carry the weight. The relevant discipline from the framework is the peak-earnings guard: FY26 sits on a mortgage origination up-cycle and a monopoly price spike, so trailing EPS should be treated as potentially peak, not run-rate.

5. Valuation — cheap vs the peak, only fair-to-full on normalized earnings

DYT/DDM N/A (no dividend). Graham IV is VOID — book equity is negative (−$189.71 BVPS from buybacks), so √(22.5·EPS·BVPS) cannot be computed and the vendor's P/B of −4.91 is meaningless (same shape as [[pitfall-yahoo-book-value-lags-buyback-equity-collapse]]). Forward P/E of 17.57 is unreliable twice over — it likely maps to the wrong fiscal year ([[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]]) and it is built on FY27 estimates that predate the Sep 4 ruling, so it is stale in the exact direction that matters. Treat it as void.

Primary frame: EV/FCF on normalized cash flow. EV = $25.48B (mkt cap $20.13B + ~$2.94B net debt + a little). Current-year multiples:

Frame Multiple
EV / TTM revenue ($2.39B) 10.7x
EV / FY26E revenue ($2.53B) 10.1x
EV / FY26E FCF (~$870M) ~29x
P/E (TTM, ~$34.6 EPS) 27x — vs ~55x at the $1,998 peak

The de-rate from ~55x to 27x has stripped most of the monopoly premium. The question is what FCF this de-rated multiple sits on. Because ~43% of revenue is mortgage scores and that line grew +127% (Q2) / +41% (Q3) on price, a realistic normalized FCF haircuts the mortgage price spike:

  • Bear — VantageScore takes low-20s% mortgage share and FICO holds volume only by cutting price; Scores growth goes negative; a weaker origination cycle compounds it. Normalized FCF ~$650–700M, multiple de-rates to ~20x → EV ~$13–14B → ~$500–560/share.
  • Base — FICO keeps most volume (embedded, more predictive, lenders pull both) and its direct $0.99 licensing recaptures reseller markup, so mortgage revenue compresses only modestly; Platform keeps growing 60%+ ARR; normalized FCF ~$780–820M at ~24–27x → ~$800–970/share.
  • Bull — "permitted everywhere, paid for nowhere": VantageScore adoption stalls beyond ~10%, FICO's direct model grows wallet share, mortgage market recovers; FCF ~$900M+ re-rates to 30x → ~$1,150–1,250/share.

Fair value range: $640–970, central ~$800. At $932 the stock sits in the upper half of that range — clearly cheap against the $1,998 peak, but only fair-to-full against normalized earnings that assume the mortgage engine is partly impaired. You are not yet being paid much of a discount to take a live, unquantified regulatory risk.

6. Sentiment & positioning

The sell side is repricing in real time, and downward. JPMorgan $1,825 → $1,325; Mizuho → $1,344; Jefferies → $1,675; Wolfe Research downgraded to Peer Perform (early Aug, ahead of the ruling, citing real VantageScore competition). Mean target $1,464 is stale and falling — every revision post-Sep-4 has cut. Consensus is still "Buy" and every target is above spot, which is unfinished business, not support: the same setup where a de-rating name looks "cheap vs targets" precisely because the targets haven't caught down yet.

Management tone: confident and specific — Lansing's "not seeing volume loss / share loss in the 20s" is a falsifiable claim the Q4 print will test. The $0.99 direct-licensing move is a genuine strategic answer (recapture the reseller markup, compete on total cost), not a panic cut.

Insider / ownership: insider ownership ~0.03% (a serial-buyback, low-insider-holding structure); institutions ~100% of float; short interest ~10% of float — elevated, reflecting the contested thesis.

7. Synthesis — weighted verdict

Two questions (framework §0):

  1. Good business? Yes, unambiguously — 85% gross / 54% operating margin, ~$870M FCF, a genuine standard-setting brand, and a Platform second leg finally inflecting (+62% ARR). On quality alone this is a 7.5–8 name.
  2. Priced in? Partly, and that's the whole problem. The 53% drawdown has removed the monopoly premium, but FY26 earnings are inflated by a mortgage up-cycle and a monopoly price spike that both just came under threat simultaneously. On normalized FCF, fair value ($640–970) brackets the $932 price in its upper half — this is a de-rated great company at a fair price, not yet a bargain, with the single most important variable (how much mortgage revenue survives) unresolved and first-measurable at the November print.

Verdict: WATCH — conviction 5.5/10

The bull case. The threat has a ceiling: the mortgage score is a rounding error in the cost of a mortgage, FICO's model is independently more predictive, lenders aren't dropping FICO (they're pulling both), switching the origination stack is slow, and FICO's direct-licensing pivot could recapture reseller margin so revenue holds even as list price falls. Non-mortgage Scores and Platform are untouched and compounding. If FICO defends volume and its direct model defends revenue, ~$35 EPS holds and $932 is cheap.

The bear case. ~43% of revenue and ~100% of recent growth came from a mortgage-score pricing umbrella that just lost its legal mandate; FICO's own $0.99 cut proves the pricing era is over; VantageScore is already >9% of GSE volume and now open to everyone; the FHFA is dangling bi-merge on top; and the whole earnings base was flattered by a strong origination cycle that will turn. If Scores growth goes negative, TTM EPS is a peak, and 27x on peak earnings is not cheap — it's the [[pitfall-multiple-trim-inverts-on-peak-cycle-cyclicals]] trap in equity form.

Why 5.5, not higher or lower. The business quality argues for more; the timing argues for patience. This is the textbook "great company, freshly wounded moat, honestly two-sided price" — the 5.0–6.0 band ("interesting with a named flaw; not an add here"). The flaw is named, dated, and first-measurable in ~60 days. Buying today pays a fair-value price to take a binary regulatory risk blind; waiting for the Q4 print costs nothing and resolves the crux.

Key risks: (1) mortgage revenue compression (price and/or share) larger than the "20s%" management guides; (2) FY26 EPS is peak — origination cycle + price spike unwinding together; (3) FHFA bi-merge cuts pull volume ~⅓ regardless of vendor; (4) net debt ~$2.9B with negative equity means the buyback EPS tailwind shrinks if cash flow softens; (5) targets are still far above spot and falling — the de-rating may not be finished.

Disposition — NOT a holding; recommend adding to Watchlist as 👀 On Deck

FICO is not owned and not on the watchlist. It is worth tracking, not buying here. It fits the 🔧 Re-Rating Plays sleeve (a quality franchise dislocated by a discrete, datable event). Suggested watchlist entry: [5.5], On Deck, entry $650–760 (where the bear case is priced and a real margin of safety opens), break/avoid if FY27 is guided to Scores revenue down double-digits or VantageScore share crosses ~20% of GSE volume. Catalyst: Q4 + FY2026 print, early Nov 2026 — the first Scores read after the ruling and the first FY2027 guide, which is when "no volume loss" and the price defense get their first hard test. Shall I add it?

8. Data traps noted

  • Graham IV / P/B void — negative book equity from buybacks ([[pitfall-yahoo-book-value-lags-buyback-equity-collapse]]).
  • Forward P/E 17.57 unusable — wrong-fiscal-year risk ([[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]]) and built on pre-ruling FY27 estimates; stale in the direction that matters.
  • Peak-earnings guard applied — FY26 sits on a mortgage up-cycle + monopoly price spike; TTM EPS treated as potentially peak, not run-rate ([[pitfall-multiple-trim-inverts-on-peak-cycle-cyclicals]]).

Sources