Financebotresearch desk研究台

V › analyze

V · Analyze from before

TRIM Fintech

This updates: Output/Stocks/Fintech/V/analyze-2026-03-21.md (2026-03-21, BUY, conviction 8.5, price then ~$301.62) Today: 2026-08-27, price $379.66 (+25.9% since baseline, 1.5% off the 52-week/all-time high of $385.57)

1. What this updates

The March file called Visa a wide-moat, fortress-balance-sheet compounder trading at an uncommon discount to its own history (24.5x forward FCF vs a 30-33x band), and set a Trim $380+ dollar level almost as an afterthought. On 2026-08-27 the price crossed that level, and the current watchlist flags a live question: is this a sell signal, or should the position simply rotate into Mastercard (MA), which is flat over the last 52 weeks at a similar multiple?

Events since the baseline: - Q3 FY2026 results (2026-07-28): net revenue $11.6B +14%, non-GAAP EPS $3.32 +11%, payments volume crossed $4 trillion for the first time, VAS revenue +34%. Guidance raised for Q4/FY26 (low end of low-teens revenue growth, mid-teens EPS growth). - Nine sell-side price-target raises between 2026-07-08 and 2026-08-25 (targets moved from the $390-435 range to $405-460); still "Strong Buy" consensus, zero sells. - CCCA reintroduced Jan-2026 gained three new Senate co-sponsors in August but has not reached a floor vote — unchanged from baseline. - DOJ debit-monopolization suit: fact discovery now scheduled to close 2026-10-16, expert discovery through April 2027 — a firmer date than the baseline's vague "trial 2027." - Interchange-fee settlement: second distribution of funds court-approved 2026-06-15 — moving forward on schedule, no escalation. - No dividend increase yet (still $0.67/quarter, set November 2025) — the next raise is expected with Q4 FY26 results, per the company's usual cadence.

2. The delta ledger

Lead items — reversed, superseded, or retracted:

# Claim (baseline) Type Status What changed
P2 "Trading at 24.5x forward FCF, a discount to the 30-33x historical band" — the core BUY reasoning Price ❌ RETRACTED Reversed. TTM FCF/share is $9.80 (see T3); at $379.66 that is 38.75x TTM P/FCF — above the top of the band the baseline itself cited, not below it. The true current-fiscal-year forward P/E (see below) is 28.7x, also above the baseline's 20.7x. The valuation argument that carried the BUY call no longer holds at this price.
T4 "Gross margin rock-steady ~80%" Trend 🔄 SUPERSEDED Visa's GAAP income statement has no true cost-of-revenue/gross-profit line; Yahoo synthesizes one, and the synthesis is unstable — the same field now reads 97.7% TTM. Operating margin is the reliable line (a real reported figure): 65.6% TTM vs 66.4% FY2025 — still elite, still flat, just measured correctly. Replacing this claim rather than editing it, per the KB rule.
J2 Conviction 8.5/10, BUY Judgment 🔄 SUPERSEDED Moved to 7.0, TRIM — see §4.
P4 Entry <$280, strong buy <$260 Price 🔄 SUPERSEDED Zone sits 26-31% below spot — a textbook stale entry zone per pitfall-stale-entry-zone-suppresses-a-name. Re-derived to $300-330 / strong buy <$290.
P5 Trim $380+ (fixed dollar) Price 🔄 SUPERSEDED This was never re-set as a multiple after March — exactly the failure mode the KB warns about. Replaced with 30x ttm ≈ $352.50, computed off the verified-clean TTM EPS field (see §3 on why ttm, not fwd, is the safe basis here).
T3 FCF 3yr CAGR 6.5%, FCF/share CAGR 8.8% Trend 📉 DRIFTED TTM FCF is $21.01B, essentially flat vs FY2025's $21.58B (-2.6%). FCF/share TTM $9.80 vs FY2025's $9.83 — the multi-year uptrend has plateaued on the trailing-12-month view. Breaks the multi-year trend outright if Q4 FY26 (reports ~late Oct) prints full-year FCF below FY2025's $21.6B.

Carried (re-tested, held):

# Claim Check
S1 Asset-light toll-booth network, two-sided network effects, Visa is infrastructure not a fintech competitor No structural change found. Stripe/Adyen/Square/PayPal remain Visa's customers, not rivals.
S2 Widest moat in fintech, Evergreen 9/10 No attack vector materialized past the baseline's own stress-test.
S3 VAS-led revenue diversification strategy Confirmed and accelerating — VAS +34% in Q3 FY26 vs +24-28% at baseline (folded into T1 below, kept structural framing here).
S4 Complementary, not competitive, with FISV/PYPL/BR in the portfolio Portfolio unchanged — PYPL, FISV, BR all still held; no overlap logic to revisit.
T5 ROIC rising, 29.2%→35.4% (FY22-25) Directionally continuing (TTM estimate ~37%, single-source approximation — see §7).
T6 Dividend 5yr CAGR ~15.4%, FCF payout ~21.5% Recomputed from the full dividend-actions history: 15.9% 5yr CAGR ($1.28→$2.68). Payout now 22.1% (TTM EPS basis). Both essentially unchanged.
ST1 Net debt/EBITDA 0.31x, Debt/Assets ~25%, fortress balance sheet Recomputed at ~0.32x — unchanged.
ST2 SBC/FCF ~4.2%, well-controlled Recomputed TTM at 4.37% — unchanged.
ST3 Insider activity: routine 10b5-1 option-exercise-and-sell, zero purchases Confirmed through Aug 2026 — same four executives, same pattern, zero opportunistic buys.
J1 Moat width Wide Carried — see §3.

Refreshed (same claim, new number):

# Claim Old → New
T1 Revenue 3yr CAGR 10.9% → TTM growth 14.4% yoy — accelerating, not decelerating.
T2 EPS CAGR 13.4% → GAAP TTM growth ~10-11% yoy; FY26 non-GAAP guided "mid-teens" — a mild deceleration on GAAP (litigation-charge base effects), still double-digit.
T7 Share count -2.2%/yr, buybacks ~85% of FCF → Buybacks accelerated: TTM $21.4B vs FY2025's $18.3B, partly debt-funded (a $3.0B net long-term debt issuance landed in the softest FCF quarter of the year). Total capital return is now 125% of TTM FCF, up from 106% in FY2025.
ST4 Analyst consensus 92% buy/strong buy, tightening → Still "Strong Buy" (1.45 mean), and nine firms raised price targets between the Q3 print and today — sentiment has gotten more bullish even as the stock nears the top of its own historical multiple band. Named as a tension in §3.
ST5 Regulatory: CCCA uncertain, DOJ suit ongoing, interchange settlement pending → CCCA still stalled (assessment unchanged); DOJ fact discovery now has a firm close date (Oct 16, 2026); interchange settlement progressing on schedule. No escalation.
P3 DYT: yield 0.89% near the high end of the 5yr band = "cheap" → Yield now 0.70% (dividend flat, price up 26%). See §3 — this is a legitimate re-rating signal, not the DYT-inversion failure mode.

3. How the close calls were decided

Was the price move a re-rating of an intact business, or did something break? The Structural and Trend rows say the business is intact — revenue and VAS growth actually accelerated, the balance sheet is unchanged, dividend growth is unchanged, insider behavior is unchanged. Nothing broke. This makes it a pure multiple story, and the multiple simply ran further than the March discount thesis anticipated — and then past it.

The vendor forward P/E is on the wrong fiscal year, and correcting it makes the stock look more expensive, not less (pitfall-vendor-forward-eps-is-the-wrong-fiscal-year). Yahoo's headline forwardPE 25.30 prices epsForward $15.00, which is Visa's FY2027 consensus — a fiscal year that has not started. Visa's fiscal year ends September 30, and FY2026 is already three-quarters reported. The correct current-year read is priceEpsCurrentYear/epsCurrentYear: $13.23 non-GAAP → true FY2026 forward P/E = 28.7x, consistent with the company's own guided "mid-teens EPS growth" off a reconstructed FY2025 non-GAAP base. This is 13% above the vendor's naive 25.3x — the error runs in the same direction the pitfall predicts (manufactures a false discount). The TTM P/E (32.31x, off the verified trailingEps 11.75, which ties to the summed quarterly EPS $11.76 to the cent) needed no correction and is the cleanest number in the file.

FCF told the same story a different way. Summing the four most recent quarterly cash-flow statements (not reading a single column — pitfall-single-quarter-fcf-read-as-ttm): OCF $22.58B − capex $1.57B = FCF $21.01B TTM, which independently ties to Yahoo's snapshot OCF figure exactly. Dividing by the four-quarter average diluted share count (2,144.8M) gives FCF/share $9.80, which reconciles closely to fin.py's FY2025 annual figure of $9.83 — a second independent tie-out. At $379.66 that is 38.75x TTM FCF — clear above the 30-33x band the March report itself established as Visa's historical range. The one quarter behind this (Q2 FY26, Dec-Mar) shows a working-capital and tax-timing anomaly (OCF fell to $3.0B against a $2.69B tax payment and an $8.3B working-capital swing) rather than a demand problem — flagged as a watch item for the Q4 FY26 print, not treated as a trend break today.

DYT flipped, and the flip is the correct read, not the pitfall's failure mode. pitfall-dyt-inverts-when-price-caused-the-yield warns that a falling price can manufacture a false "cheap" signal. Here the mechanism ran the other direction and is not the trap: the dividend did not fall — it held flat at $0.67/quarter while the price rose 26%, so the yield compressed from 0.89% to 0.70%, now sitting below its own 5yr average (0.75%) rather than near the top of the band. A yield falling because the price rose on an intact, growing dividend is DYT doing its job correctly — it now reads fair-to-slightly-rich, which agrees with the P/FCF and current-year P/E findings above rather than contradicting them.

The MA rotation question, resolved on a same-basis comparison. The watchlist's framing — "MA is flat over 52 weeks while V hit its trim" — reads as an unexploited discount. It is not, once measured on comparable fields for both names:

V MA
Trailing P/E (verified TTM EPS, no fiscal-year ambiguity) 32.31x 32.53x
PEG (same vendor field, same methodology, both names) 1.74 1.83-1.84
Recent quarterly EPS growth (adjusted/non-GAAP) +11% (Q3 FY26) +22% (Q2 CY26)
52-week total return +9.7% +1.4%
Distance from own 52-week high -1.5% -1.6%

Trailing multiples are within a rounding error of each other, and MA's PEG is higher, not lower, meaning the vendor's own growth-adjusted read does not call MA cheap. The "flat 52 weeks" is a round-trip, not a discount — MA fell to a 52-week low of $464.52 mid-year and has since recovered to within 1.6% of its own high, the same distance V sits from its high. Both names are, right now, near the top of their own respective ranges. MA's faster recent earnings growth (22% vs 11%) is real and worth naming, but the market has not mispriced it relative to V on any of the fields checked here — it shows up in the (nearly identical) PEG, not as a free discount. Resolution: this is not a clean rotation trade on valuation grounds. The correct action is to trim the name already held and priced above its own historical FCF and earnings band — not to chase MA on the theory that its flat one-year chart means it is cheap. If the user's goal is single-name concentration reduction rather than a valuation call, that is a legitimate but separate reason to add MA — it should not be justified as "the cheaper way to own the same economics," because on today's numbers it measurably is not.

A live instance of the dual-class share-count trap, contained. Yahoo's sharesOutstanding (1.704B) does not reconcile against marketCap ÷ price (1.867B implied) — a 9.6% gap, the signature of pitfall-yahoo-share-count-dual-class-fpi. Visa's 10-Q cover page (Mar-31-2026) confirms: Class A 1,659.7M + Class B-1 4.8M + Class B-2 120.3M + Class C 8.9M = 1,793.8M raw shares, below even Yahoo's flawed 1.704B for Class A alone plus a partial B/C count — Visa's Class B/C convert to Class A at a ratio above 1:1 tied to a litigation-escrow schedule, which is why the implied fully-diluted figure (1.867B) sits higher still. This would corrupt any hand-built per-share figure, but it does not corrupt the numbers in this report, because every EPS and FCF/share figure above was taken from Visa's own reported diluted weighted-average share count (which already applies the correct as-converted methodology) or reconciled against independently-summed quarterly statements — never from the raw sharesOutstanding field. Filed as a new confirmed case in the existing pitfall note (§6).

4. Thesis persistence and conviction delta

Structural + Trend claims: 11 total. Survived as CARRIED or REFRESHED: 9 (S1-S4, T1, T2, T5, T6, T7). Drifted: 1 (T3, FCF plateau). Superseded: 1 (T4, gross-margin field). Thesis persistence: 82%.

This is a high-persistence result, but it points the opposite direction from what high persistence usually means. In the classic pattern (pattern-shortlist-runs-when-unexecuted and similar), a business holding intact while the price falls is the setup — the multiple re-rates up to catch a discount. Here the business held intact while the price rose past fair value: every Price-layer claim (P1-P5) reversed or superseded, and the one claim that most directly carried the March BUY case (P2, "cheap vs its own history") is the one retracted outright. The company did not change. The price ran ahead of it, and by more than the multiple ceiling the March report itself set.

Conviction: 8.5 → 7.0. Driven specifically by: - P2 retracted — the discount thesis inverted into a premium. - T3 drifted — FCF/share growth stalled on the trailing year, the metric the framework weights most heavily for this business type. - The MA rotation question resolving as "no discount found" removes the one plausible argument for treating the breach as noise.

Not moved lower than 7.0, because every Structural claim carried clean, the balance sheet and dividend trajectory are untouched, and the regulatory overhang — the risk factor the March report worried about most — did not escalate. This is "excellent business, priced past fair value, trim into strength," not "thesis broken."

5. What is genuinely new

  • N1 — MA rotation, resolved. See §3. Not previously addressed by any V or MA report; the question originated in the 2026-08-27 watchlist scan.
  • N2 — Dual-class share-count trap confirmed live on V. New ticker for the existing pitfall-yahoo-share-count-dual-class-fpi note (§6), contained without corrupting this report's figures.
  • N3 — FCF/share plateau, TTM basis. A genuinely new finding not visible in March (the baseline only had FY2025 annual data). Tied to one soft quarter (Q2 FY26) with a plausible tax/working-capital explanation; needs the Q4 FY26 print to confirm or break.
  • Sell-side sentiment diverging from valuation discipline. Nine price-target raises land the mean target ($417.90) still above spot, even as this analysis's own fair-value range tops out at $370. Named as a tension, not resolved — the Street is chasing the print, this report is pricing the multiple.

6. Updated verdict

Visa remains one of the widest moats available to buy — nothing in this pass found a crack in the network-effect, switching-cost, or scale advantages the March report described. The change is entirely in what the market is now paying for it.

Valuation, rebuilt from scratch (Price rows never carry forward)

Model Fair value Weight Basis
Graham IV $70.67 0% (inapplicable) Buyback-hollowed book, same as baseline — disregarded.
Bogle Expected Return ~9-16%/yr (center ~13%) 20% 0.70% yield + 10-14% earnings growth ± P/E change; narrower and lower than March because most of the room for further re-rating is gone.
DYT $282-357 (center ~$320) 15% Reset to the 0.75-0.95% band on a flat $2.68 dividend — reads fair-to-rich at spot, reversed from March's "cheap."
DDM (2-stage, 15%×10yr then 5-6% terminal, 9.5-10% discount) $340-445 20% Inputs re-verified (dividend CAGR, payout ratio) — largely unchanged from baseline.
P/FCF (primary) $294-352 (30-36x TTM FCF/share $9.80) 45% Historical band re-confirmed at 30-33x; spot is at 38.75x, above the top of the band.

Composite fair value: $320-370 (center ~$345). Current price $379.66 sits ~10% above the top of this range.

Entry / Trim

  • Entry $300-330, strong buy <$290 — re-derived off the fair-value range above, not carried from the stale March zone (which sat 26-31% below spot and had gone silent per pitfall-stale-entry-zone-suppresses-a-name).
  • Trim 30x ttm (≈ $352.50 at TTM EPS $11.75). Written on the ttm basis deliberately — the fwd basis would have the site compute the dollar level off Yahoo's forwardPE, which is contaminated by the wrong-fiscal-year EPS field described in §3 (pitfall-multiple-trim-inherits-the-broken-vendor-field); the ttm field ties to reported quarterly EPS to the cent and carries no such defect. 30x is a deliberately generous multiple for a wide-moat compounder — a stricter cap matched to the 30-33x P/FCF band would trim lower still, but this keeps the judgment anchored to earnings rather than double-counting the FCF discipline already expressed in the fair-value range.
  • Spot ($379.66) is already above this trim level — the breach the watchlist flagged is a real signal on this pass's numbers, not a stale-zone false alarm.

Recommendation

TRIM. Not an exit — the business case is unchanged and this remains a top-tier holding. Take some profit into strength above the $352-370 zone; do not add at current levels; do not rotate into MA as a "cheaper" substitute (§3). Revisit sizing after the Q4 FY26 print (~late October), which will also settle whether the FCF/share plateau (T3) was one quarter's timing or the start of something durable.

Break triggers (any one moves this toward a harder sell): - Full-year FY2026 FCF (reporting ~late Oct) prints below FY2025's $21.58B — confirms the TTM plateau as a real trend break, not a timing artifact. - CCCA advances to a floor vote with credible passage odds. - DOJ debit suit produces a document disclosure or ruling materially worse than the baseline's "manageable" read, around the Oct 16 discovery-close date. - A dividend raise below the ~13-16% cadence of the last five years (watch the Nov 2026 raise).

Upgrade conditions (back toward ACCUMULATE): - Price retraces into the $320-350 fair-value band without a fundamental deterioration. - Q4 FY26 FCF resumes its multi-year growth pace, confirming T3 as a one-quarter anomaly.

7. What this pass did NOT test

  • ROIC was re-estimated at ~37% TTM using operating income growth and an approximated invested-capital base, not rebuilt from a full MRQ balance sheet. Single-source, directionally consistent with the baseline's rising trend, but not independently corroborated this pass — treat as CARRIED-with-caveat, not verified.
  • Debt-to-Assets was not recomputed from the MRQ balance sheet in full; net debt/EBITDA (~0.32x) was cross-checked and is unchanged, and is a reasonable proxy, but the literal debt/assets figure is carried from FY2025 annual (25.3%) rather than refreshed to Q3 FY26.
  • Cross-border revenue growth and segment-level VAS detail (the granular breakdown the March report gave by sub-segment) was not rebuilt quarter-by-quarter this pass — only the headline VAS +34% and cross-border +12-13% figures from the Q3 print were checked.
  • The FCF/share plateau (T3/N3) is a single-source finding from this pass's own quarter-by-quarter reconstruction. It reconciles to two independent totals (Yahoo's snapshot OCF and fin.py's FY2025 annual FCF/share), which is real corroboration, but the explanation (tax/working-capital timing in Q2 FY26) is an inference, not confirmed against the 10-Q's own MD&A language. This is the single most important open question for the next pass — the Q4 FY26 print will resolve it either way.
  • MA was analyzed only on the fields needed to answer the rotation question (P/E, PEG, 52-week range, recent EPS growth) — this is not a full /analyze MA and should not be cited as one. If MA is added to the watchlist as its own name, it needs its own pass.

Sources