Phase Space AI

Financial Model Notes

Adobe [ADBE]

Adobe Inc. [ADBE] — Financial Model Notes

As of 2026-07-29. Provenance, construction, adjustments, and the weaknesses in my own numbers. Reproduction scripts are committed under work/.

1. Data provenance

Item Source Retrieval
Prices Alpaca v2/stocks/snapshots, v2/stocks/{sym}/bars work/fetch_prices.py, work/gap_check.py
Fundamentals SEC XBRL companyfacts CIK 0000796343 work/edgar_pull.py, work/extract.py
Balance sheet, income statement 10-Q Q2 FY26 adbe-20260529.htm, read directly work/fetch_docs.py, work/q.py
Disclosure language FY23/FY24/FY25 10-Ks + Q2 FY26 10-Q, full text work/censor.py, work/mentions.py
Earnings detail, ARR, cRPO, guidance 8-K Ex-99.1 adbeex991q226.htm, adbeex991q425.htm work/fetch_pr.py
Multiple history own P/S by fiscal quarter work/multiple_hist3.py
Implied path assets/reverse_dcf.py work/run_paths.py, work/run_exit.py
Peer anchor XBRL + balance sheets, 11 comparators work/anchor.py
Option chain Alpaca options contracts + snapshots work/liquidity.py

Recency asserted. Period 2026-05-29, filed 2026-06-15 — 44 days old at the memo date, the freshest of the five. The GOOGL (485-day-stale companyfacts) and Alcon (filing history ending 2010) failures do not apply.

2. Quarterly series construction

Adobe's fiscal Q4 is not separately tagged in XBRL; only the full year is. Q4 is derived as the annual value less the three filed quarters, with contiguity enforced (consecutive period ends 80–100 days apart) before any TTM is summed.

Quarter ended Revenue ($m) Derivation
2025-02-28 5,714 filed 10-Q
2025-05-30 5,873 filed 10-Q
2025-08-29 5,988 filed 10-Q
2025-11-28 6,194 derived: FY25 23,769 − (5,714 + 5,873 + 5,988)
2026-02-27 6,398 filed 10-Q
2026-05-29 6,618 filed 10-Q

TTM revenue = 5,988 + 6,194 + 6,398 + 6,618 = $25,198m. Matches the screen exactly. Last-FY revenue ($23,769m) would have understated the base by 5.7%.

Revenue tag note: Adobe reports under Revenues, not RevenueFromContractWithCustomerExcludingAssessedTax. A pipeline keyed only on the latter finds nothing for Adobe. My extractor falls through a tag priority list; this is recorded because it is the same class of error as §4 below.

3. Share count

Basis Count Use
dei cover page, 2026-06-11 397.5m Market cap. What the screen used — correct.
Balance sheet, 2026-05-29 399m 601m issued − 202m treasury. Cross-check ✓
Diluted weighted average, Q2 FY26 402.5m EPS cross-check only
Company guidance, FY26 diluted ~399m Company's own forward figure
12 months forward (my estimate) ~385m 397.5m less ~12.5m net

Cross-check, mandatory per the brief: net income ÷ diluted shares = 1,712 ÷ 402.5 = $4.253 against filed diluted EPS of $4.25. ✓ Scale confirmed. No dual-class complication — Adobe has a single class of common stock (900m authorised, 601m issued), so the dei tag survives SEC aggregation, unlike WDAY's.

Buyback evidence for the forward count, so it is not an assumption: treasury stock $48,847m → $53,418m in six months ($4,571m in Q2 alone), treasury shares 188m → 202m, FY25 retired 30.8m. My ~12.5m annual net reduction is conservative against a run-rate above 30m.

4. Net cash — where the screen went wrong, and the general lesson

The screen reported +$4,447m of net cash. Adobe has −$1,019m of net debt. Itemised from the balance sheet:

Component 2026-05-29 ($m) Screen Mine
Cash and cash equivalents 4,919 ✓ ✓
Short-term investments 707 ✓ ✓
Debt, current (1,843) used (850) from 2026-02-27 ✓
Debt, long-term (4,802) OMITTED ✓
Operating lease liabilities, current (91) omitted disclosed, excluded
Operating lease liabilities, noncurrent (329) ✓ as debt disclosed, excluded
Net cash ex-lease (1,019) +4,447 HEADLINE
Net cash incl. leases (1,439) — Alternative

Root cause. Adobe tags its long-term debt as LongTermDebt (4,802 @ 2026-05-29). Adobe's LongTermDebtNoncurrent tag was last filed on 2015-08-28 at $1,900m. A pipeline keyed on LongTermDebtNoncurrent therefore either finds nothing (and omits $4.8bn) or, worse, finds a well-formed eleven-year-old number.

This is calibration item D1 in a form the D1 sweep did not cover. D1 is "existence is not validity" for missing tags returning nan. This is the same disease for stale tags returning a plausible value: nothing errors, nothing is nan, and the output is a confident wrong number. Every balance-sheet tag must be date-filtered against the period end, not merely tested for existence. The same pattern is live in CRM (MarketableSecuritiesCurrent last filed 2014) and PANW (OperatingLeaseLiabilityCurrent last filed 2025-07-31).

Tags checked per the brief: LongTermDebt ✓ 4,802 @ 2026-05-29 · LongTermDebtCurrent ✓ but the live figure is the balance sheet's "Debt, current" of 1,843, and the XBRL tag's latest value is 850 @ 2026-02-27 — so the balance sheet is authoritative over the tag here · LongTermDebtNoncurrent ✗ stale to 2015 · ConvertibleDebtNoncurrent absent · MarketableSecurities* absent; Adobe uses ShortTermInvestments.

Consequence: EV $105,738m rather than $94,614m — 11.8% higher, EV/S 4.20x rather than 3.75x. The error flattered the cheapest name in the cluster. Correcting it still leaves ADBE the cheapest, which is why the conclusion survives the correction; but the conclusion was not derived from the screen's number.

5. Adjustments applied, with reasons

# Adjustment Effect Why
1 Add back the $68m goodwill impairment to Q2 operating income margin 33.8% → 34.8% Non-cash, legacy Publishing & Advertising unit, disclosed as $0.17/share. The run-rate is what matters. Note this adjustment makes the margin better and the trend still compresses.
2 Strip ~$40m Semrush from Q2 revenue growth 12.7% → 12.0% Company-disclosed.
3 Use constant currency as the demonstrated rate 12.0% → ~10.3% The company reports 13% and 11% CC in the same bullet. FX is not a business capability and it reverses. This is the figure used in the implied-path test, and using the reported 12.0% instead would move the margin from +9.3pp to +11.0pp — i.e. my choice is the conservative one.
4 Decompose guided FY26 ending-ARR growth of 10.2% for the ~1.8pp Semrush contribution organic ARR ~8.4% Arithmetic on two disclosed figures ($480m Semrush ARR within $27.10bn). This is the most important derived number in the memo and it is not disclosed as such by the company.
5 Step shares to ~385m for forward work +3.2% to forward per-share values Evidenced by an observed >30m/yr retirement rate; deliberately conservative.
6 Exclude operating leases from headline net cash EV −$420m vs incl-lease Consistency with the peer anchor set, computed on the same basis.

No SBC adjustment. SBC is real, it is 8.2% of revenue, and the 36.1% GAAP operating margin carries it. The non-GAAP ~45% margin is reported for reference and is never used as a terminal margin. Using it would move the required CAGR from 1.0% to deeply negative and would be exactly the error the framework exists to prevent.

6. What the model deliberately does not contain

7. Known weaknesses in my own numbers

Stated so a reviewer can attack them.

  1. The 36% terminal margin is ~1pp generous. The latest clean quarter is 34.8% and the trend is down. At 34% the required CAGR rises from 1.0% to ~2.6% and the margin falls from +9.3pp to +7.7pp — the verdict does not change, which is the point: unlike CRM, ADBE's conclusion is not hostage to this input. At 30% the required CAGR is ~5.6% and the margin is still +4.7pp.
  2. The ~8.4% organic ARR figure is my arithmetic on two disclosed numbers, not a disclosure. It assumes Semrush's $480m of ARR is entirely incremental to the FY25 base. If some of Semrush's ARR displaced existing Adobe spend, organic is higher than 8.4%; if the guide embeds further FX, it is lower. It is the load-bearing bear-side number and it carries this uncertainty.
  3. cRPO in dollars (~$14.9bn) is my multiplication of 67% by $22.27bn. Adobe discloses only the percentage. Rounding in the disclosed 67% implies ±$110m.
  4. NTM revenue of $28,070m stitches two guided quarters to two modelled ones at ~11% growth. If FY27 is guided to 9%, NTM is ~$27,700m and the 12-month target falls from $303 to ~$299 — immaterial, because the target's sensitivity is overwhelmingly to the multiple, not the revenue.
  5. The P/S history uses a 45-day post-period-end price proxy, not actual filing dates. Noise of a few percent per observation; it does not move the 10th-percentile conclusion, which holds against the whole range.
  6. The anchor set has six usable members after excluding WDAY and ADP for non-normalised margins. Six is thin for a median, so the full 11.6x–27.7x range is carried through every sensitivity rather than treating 17.5x as a point estimate. For ADBE specifically this matters least of the five, because the margin is positive at every point in the range.
  7. I have not verified that "AI-first ARR" excludes Semrush. The release states total ARR "including approximately $480 million from Semrush" but makes no such statement about the >$500m AI-first figure. If Semrush's SEO/marketing-analytics ARR is counted as "AI-first," the organic AI number is materially smaller than $500m. The definition is not disclosed and I have not resolved it.