Financebotresearch desk研究台

INTU › analyze

INTU · Analyze

ACCUMULATE Technology

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

1. What this updates

The baseline was written 21 days before the single largest scheduled event in Intuit's year and said so explicitly: "That single event delivers FY2026 results, first FY2027 guidance, and the annual dividend raise." That event has now happened — 2026-08-25 — and it was the whole point of the recheck date. This pass exists to collect the answer.

Event list since 2026-08-04

Date Event Ledger rows touched
2026-08-25 Q4 + full-year FY2026 results; first FY2027 guidance; annual dividend raise St1, T1–T8, S3–S7, all Price rows
2026-08-25/26 Stock fell on the FY27 guide despite the beat; some analysts downgraded on AI-disruption concerns Price, J1
— Price $319.97 → $347.22 (+8.5%); analyst mean target $454.65 → $428.61 Price rows
2026-09-17 Investor Day — has not happened; still ahead —

One event dominates. Per the adjudication rule, the print is one force that moves many rows, not fifteen independent confirmations — and that is how it has been weighed below.

2. The delta ledger

25 baseline claims. 2 retracted, 1 superseded, 8 refreshed, 12 carried, 2 untested.

❌ RETRACTED — wrong when written

# Baseline claim What was actually true Cause
St1 "TTM FCF has fallen to ~$5.23B from the FY2025 $6.08B… this is not yet resolvable and must not be waved away." FCF ROSE. FY2026 OCF $8.838B, capex $221M, FCF ≈ $8.617B — up 41.7% from $6.08B. True TTM at the baseline's own date was $7.715B, not $5.23B. Reading error, not a vendor trap. $5.236B is the Q3 FY26 single-quarter FCF — Intuit's tax-season quarter, which alone carries ~two-thirds of annual cash generation. It was read off one column and labelled TTM. New note: [[pitfall-single-quarter-fcf-read-as-ttm]].
St5-proxy "Intuit's non-GAAP flatters less than half as much as Adobe's" — INTU Q3 GAAP $11.09 vs non-GAAP $12.80 = +15.4%, vs ADBE's +35.9% On the full year: GAAP EPS $16.46 vs non-GAAP $24.27 = +47.4% — wider than the ADBE reference, not less than half. Same class of error: a single seasonal quarter generalised to the business. The favourable comparison inverts on annual figures.

Both retractions were caused by generalising one quarter to a year, and both trace to the same fiscal-calendar seasonality. That is one methodological failure, not two.

⚠️ The direction of the first retraction matters and cuts in Intuit's favour. The baseline reached ACCUMULATE [7.0] while carrying a false cash-flow decline and a conservative dividend payout ratio (25% on the wrong denominator; the real figure was ~17%). The verdict was correct despite a headwind that did not exist. The second retraction cuts the other way.

🔄 SUPERSEDED — facts changed

# Baseline claim Replaced by Cause
T5 "Share count flat, −0.1%/yr — per-share compounding has come from the business, not from financial engineering." Share count −2% in FY2026. Buybacks $5.5B, +96% vs FY25's $2.77B. New $8B authorization. Diluted shares 283M → 273.5M. Deliberate capital-allocation change. The baseline predicted this — "the FY26 buyback acceleration now adds the other lever on top" — so the claim is superseded by a development it anticipated, not falsified.

🔁 REFRESHED — same claim, new number

# Claim Baseline Now (FY2026 actual)
T1 Revenue growth +14.0% 3yr CAGR $21.4B, +14% — held exactly
T2 Operating margin 26.2%, +610bps/3yr 27.5% GAAP, +130bps further; non-GAAP 41.5%
T3 Net income / EPS growth +23.3% / +23.4% 3yr GAAP EPS $16.46, +20%; non-GAAP $24.27, +20%
T4 FCF CAGR +18.5% 3yr, FY25 $6.08B FY26 $8.617B, +41.7%
T7 Dividend, 15 straight raises ~15% Last raise +15.4% 16th raise: +15% to $1.38/qtr
T8 CapEx 0.66% of revenue $124M $221M = 1.03% of revenue — still asset-light
St2 FCF payout ratio ~25% (on the wrong denominator) ~17% — survives an ~83% FCF cut
St3 ROIC ~15% on $13.98B goodwill ~15% ~17-18% inferred from op income $5.9B ⚠️ single-source, FY26 balance sheet not yet filed

T7 deserves separate note. The baseline wrote an explicit falsification test: "~15% (to ~$1.38/qtr) says the FCF story is intact; anything under 8% would be the first management-signaled crack." Management delivered +15% to $1.38 — the exact number, to the cent. A pre-registered prediction that lands is materially stronger evidence than the same fact observed after the fact.

✅ CARRIED — re-tested, still true

# Claim The check
S1 Tax compliance is a legally mandated, annually recurring workflow with an accountable-party requirement; Intuit's risk is price, not existence Nothing in the print touches this. Management's FY27 response is to cut entry-level price — a price response, precisely as the claim predicts.
S2 Regulatory intangible — IRS e-file authorization, 41 state engines, EFIN/ETIN, 45-day suitability check, annual re-certification; AI reduces none of it No regulatory change. Untouched by the print.
S3 Accountant channel — 200k+ ProAdvisors, ~80% US QBO share Strengthened: QBO Accounting +23%, GBS $12.9B +16%, Online Ecosystem $9.9B +19%. The best-defended segment grew fastest.
S4 Switching costs stable at complex, ~zero at simple; sub-$50k DIY already lost, structurally unrecoverable at a premium price Directly confirmed: TurboTax units −2% while TurboTax revenue +7% ($5.3B) and TurboTax Live +37%. Units leaving at the bottom, revenue held by mix at the top — exactly the claim's shape. Management acknowledged "a loss of quality DIY tax customers to lower-cost providers."
S5 Brand stable but no longer supports a price premium Strengthened to the point of management action: FY27 TurboTax guided +2-3% because the company is lowering entry-level pricing. The baseline inferred it; management is now executing on it.
S6 Mailchimp is the worst asset (3.0) Confirmed and worsened — guided flat-to-negative for FY27. Online Services excluding Mailchimp grew +24%.
S7 GBS/QuickBooks is the largest segment and the best-defended GBS $12.9B of $21.4B = 60% of revenue, growing +16%. Both halves hold.
T6 Gross margin stable ~79.6%; no margin-compression signature Gross margin 81%; operating margin expanded. The moat-erosion signature remains absent.
St4 Debt/assets 18%, positive tangible book Debt $6.64B on assets $36.96B = 18.0% ⚠️ FY25 balance sheet; FY26 not yet filed
J2 Moat 7.0/10, EVERGREEN Every component re-tested; see §3. Score held.
P-DYT DYT void — the yield rose because the price fell Still void. Yield 1.39% vs 5yr avg 0.67% = 2.07x, but the price is −56% from the high. [[pitfall-dyt-inverts-when-price-caused-the-yield]] — carried from the baseline, which applied it correctly.
P-Graham Graham IV weight 0.5, reality check only Graham IV $166.90 vs $347.22. BVPS $75.22 is ~$51/sh goodwill. Still weight 0.5.

⏳ UNTESTED

# Claim Why
St5 SBC not retrieved — baseline refused to estimate it Still not pulled. Second consecutive pass. The GAAP/non-GAAP gap is now known ($3.0B of add-backs on non-GAAP op income $8.9B vs GAAP $5.9B = 14% of revenue) but the SBC line itself needs the 10-K, not yet filed. This is the system deferring the same question twice — flagged per protocol.
St6/St7 Insider activity — founder sold $276M pre-collapse; one director bought $541k fin.py --holders returned no insider section this run. Not verified. Does not carry forward as true.

🆕 NEW — no baseline counterpart

Finding Detail
FY2027 guidance exists Revenue $23.279–23.512B (+9–10%) · GAAP op income $7.408–7.490B (+26–27%) · GAAP EPS $20.12–20.36 (+22–24%) · non-GAAP EPS $22.88–23.12 (+23–24%)
Online paying customer growth is slowing +3% to 8.9M. Management acknowledged the slowdown. Not a metric the baseline tracked; it should be tracked from here.
"Big bets" disclosure Assisted Tax + Money + Mid-Market grew 34%, now 30% of full-year revenue
Credit Karma re-accelerated +20% to $2.6B — against a baseline moat rating of 4.0 ("lead-gen, not a moat"). Growth is not moat; the rating stands, but the asset is performing.
New $8B buyback authorization On a $94.9B market cap = 8.4% of shares
Vendor tripwire fired Yahoo's snapshot shows OpMargin 18% below NetMargin 21% — impossible. See §3.

3. How the close calls were decided

The FCF retraction — three independent confirmations before it was called

A retraction on a single source would be a guess. Three: 1. Arithmetic tell — "$5.23B TTM" matches the Q3 FY26 quarterly cell ($5.236B) to three significant figures. A TTM reproducing one quarterly cell that precisely was copied, not summed. 2. Recomputation — summing the four quarterly cells (356 + 599 + 1,524 + 5,236) gives $7.715B, not $5.23B. 3. Primary source — Intuit's own release: OCF $8.838B less capex $221M = $8.617B for FY26.

The vendor tripwire — a pattern note written yesterday caught a live error

fin.py's snapshot reports OpMargin 18%, NetMargin 21%. Net margin cannot exceed operating margin absent non-operating gains — [[pattern-net-margin-above-operating-margin-is-a-tripwire]], written 2026-08-26. Recomputed from the quarterly statements, TTM operating margin is 27.5% and net margin 21.9%. Yahoo's OpMargin field is simply wrong, and it agrees with the company's own reported 27.5% GAAP operating margin. No claim was built on the broken field, because the tripwire fired before it was used. Recorded, not acted on.

Consumer weakness — one force, not four

TurboTax units −2% · FY27 TurboTax guided +2-3% · entry-level price cuts · online paying customers +3%. These read as four negatives. They are one: Intuit has conceded the low-end DIY filer and is now pricing to defend the boundary. The baseline already named this (S4, S5) and rated that segment 3.0/10 — "already lost, structurally unrecoverable at a premium price."

So the question is not whether Consumer is weak — the baseline said it was — but whether the weakness has spread beyond the segment already written off. It has not: GBS is 60% of revenue growing +16%, QBO Accounting +23%, TurboTax Live +37%, Credit Karma +20%. A Structural claim holding explains the State weakness rather than being broken by it. Consumer is behaving exactly as modelled. That is why S4/S5 are CARRIED rather than DRIFTED, and why the ledger does not count this four times.

What would change that reading: the price cut failing to stabilise units, or GBS decelerating below ~12%. Both are named in the break triggers.

The two retractions pull in opposite directions

The FCF retraction is favourable — the baseline underwrote a decline that was actually a 41.7% increase. The non-GAAP retraction is unfavourable — a claimed quality advantage over ADBE inverts. They do not cancel to zero. The FCF error touched the central open question of the baseline and resolved it decisively; the non-GAAP gap is a comparative talking point that changes no cash flow. Weighted accordingly: net favourable, but the conviction move is held to +0.5 rather than +1.0 partly because of it.

Why the FCF multiple was weighted down rather than used

FY26 FCF margin jumped to 40.3% from 32.3%. At 18–21x, FY26 FCF/share of $31.50 implies $567–662 — far above the earnings-based $368–460. Two models disagreeing by 50% means one is mismeasured, and that must be resolved before either is used. OCF grew +42% against net income +20%; the gap is working-capital and deferred-revenue timing that the 10-K has not yet explained. The earnings range is the defensible one. The FCF read is carried only as a statement that the name is not expensive, at reduced weight, with the divergence flagged for the next pass.

4. Thesis persistence and conviction delta

Thesis persistence: 14 of 15 Structural + Trend claims survived as CARRIED or REFRESHED = 93%.

  • Structural: 7/7 held. Not one business-model claim was overturned.
  • Trend: 7/8 — T5 superseded by an anticipated capital-allocation change.

93% persistence against a −56% drawdown from the high and a +8.5% price move is the re-rating signature: the business held and the multiple moved. The one claim that changed changed because management started buying back stock, which the baseline predicted.

Conviction: 7.0 → 7.5

Driven by, and only by, these rows:

Row Move
St1 RETRACTED The baseline's single largest open risk — a falling FCF trend — did not exist. FY26 FCF +41.7%. This alone is most of the upgrade.
T7 REFRESHED The baseline's own pre-registered falsification test passed to the cent (+15% to $1.38).
T5 SUPERSEDED Buyback doubled to $5.5B, share count −2%, new $8B authorization — the second per-share lever the baseline said was coming.
St2 REFRESHED Dividend coverage is ~17% of FCF, not 25%.

Held back from 8.0 by:

Row Drag
St5-proxy RETRACTED The non-GAAP quality comparison inverts on annual figures.
S5 CARRIED (strengthened) Management is now cutting entry-level price — the concession is real and dated.
NEW Online paying customers +3%; FY27 revenue guided +9–10%, a genuine deceleration from +14%.
St5 UNTESTED ×2 SBC still unpulled on a second consecutive pass.
Price +8.5% since the baseline — less discount for the same thesis.

5. Valuation — re-derived from scratch

⚠️ Yahoo's PE(fwd) 12.56 is a phantom and was not used. It implies forward EPS of $27.64, which matches no guided year — FY27 non-GAAP is $22.88–23.12, FY27 GAAP is $20.12–20.36. [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]], the same trap the baseline flagged at $11.7x and it is still live.

Basis EPS Multiple at $347.22
FY27 non-GAAP guide (mid) $23.00 15.1x
FY27 GAAP guide (mid) $20.24 17.2x
TTM GAAP (= FY26 actual) $16.46 21.1x ✓ matches Yahoo's trailing PE and the press release
Model Weight Range
Forward earnings — 16–20x on FY27 non-GAAP $23.00 45% $368–460
FCF multiple — 18–21x on normalized FY25/26 FCF $26.87/sh 25% ⬇️ (from 40%) $484–564 (divergent — see §3)
Bogle — 1.39% yield + durable 12–14% EPS growth 20% ~13.5–15.5%/yr, supports the range
Graham IV $166.90 10% (0.5 weight) Drags; goodwill-punished
DYT · DDM 0% Void — reported and discarded

Fair value: $385 – $475 (midpoint ~$430)

Barely moved from the baseline's $380–470. That is the honest result: FY27 guidance came in close to what the baseline modelled. Spot $347.22 → +11% to +37%, midpoint +24%.

Baseline Now
Fair value $380–470 $385–475
Entry $300–345 $305–350
Strong buy <$280 <$285
Trim 17x fwd 28x ttm ≈ $461
Conviction [7.0] [7.5]

⚠️ The trim is deliberately ttm, not fwd. The site computes a multiple-trim as mult ÷ vendor P/E × price, using Yahoo's forwardPE — which here is the phantom 12.56. A 20x fwd trim would render $553, ~20% too high. Yahoo's trailing PE (21.09) and EPS(ttm) ($16.46) are verified correct against the company's own FY26 GAAP EPS, so the ttm basis is sound. 28x ttm ≈ $461 ≈ 20x FY27 non-GAAP — the intended level, reached through a denominator that is not broken. [[pitfall-multiple-trim-inherits-the-broken-vendor-field]].

Zone note, stated plainly: spot $347.22 sits at the top edge of the re-derived $305–350 band — inside by 0.8%. That is a boundary, not a buy signal, and the band moved $5 only because fair value moved $5. Do not read "back in zone" as a green light on a 0.8% margin.

6. Verdict — ACCUMULATE [7.5]

FY2026 delivered +14% revenue, +20% EPS, 27.5% operating margin (+130bps), $8.6B of free cash flow (+41.7%), a 16th consecutive dividend raise of +15%, and a doubled buyback that took the share count down 2% for the first time. The one large uncertainty the previous report carried — a cash-flow trend it believed was deteriorating — was an artifact of reading one quarter as a year, and the true trend is strongly positive.

Against that: the Consumer segment is conceding the low end on price, by management's own decision, FY27 revenue is guided to decelerate to +9–10%, and online paying customer growth has slowed to +3%. None of this is new information about the business — the baseline rated the sub-$50k DIY filer 3.0/10 and called it structurally unrecoverable at a premium price. It is that prediction being executed.

The name is a 15.1x forward-earnings business guiding +23–24% EPS growth, 56% below its high, with 93% of its prior thesis intact.

Break triggers: GBS revenue growth below ~12% · QBO Accounting below ~15% · the FY27 entry-level price cut failing to stabilise TurboTax units (a second year of unit decline with lower price) · FY27 non-GAAP EPS guided below +15% at any quarterly update · operating margin failing to expand · a dividend raise under 8% · online paying customers flat or negative.

Upgrade to [8.0]: SBC pulled from the 10-K and confirming the non-GAAP gap is amortization-led rather than SBC-led · the FCF margin jump to 40.3% explained and shown durable · TurboTax units returning to growth on the new pricing · GBS holding ≥16% through FY27 H1.

Downgrade to [6.5]: GBS decelerating below 12% · a second consecutive year of TurboTax unit decline despite the price cut · any FY27 guidance cut.

7. What this pass did NOT test

  • SBC — second consecutive pass untested. The baseline refused to estimate it; this pass still could not pull it (FY26 10-K not yet filed). The system has now deferred the same question twice, which is exactly what the protocol says to call out. Pull it at the next touch.
  • Insider activity — fin.py --holders returned no insider section. The baseline's findings (founder's $276M sale, a director's $541k purchase) are not re-verified and do not carry forward as true.
  • FY2026 balance sheet — not yet filed. ROIC ~17-18% and debt/assets 18% are inferred from the FY25 balance sheet against FY26 income. Both are single-source figures.
  • The FCF margin jump to 40.3% — real and primary-sourced, but unexplained. Working capital and deferred revenue are the likely cause; the 10-K settles it. This is why the FCF model was weighted down rather than trusted.
  • Price vs mix in TurboTax's +7% revenue on −2% units. ARPU rose ~9%, but Intuit does not disclose an ASP series at the granularity that would separate price increases from Live mix — see [[pattern-disclosed-asp-series-separates-price-from-volume]] for what the disclosure would need to look like. Recorded as unresolved, not assumed.
  • Investor Day, 2026-09-17 — has not happened. It is the next scheduled information event before the Q1 FY27 print.

Sources