Financial Shenanigans

Financial Shenanigans — Adobe Inc.

Adobe's reported numbers look like a faithful representation of economic reality. Cash flow runs ahead of net income every year, receivables grow in line with revenue, the auditor has issued clean opinions, and there is no restatement, material weakness, or regulator action of record. The forensic concerns here are not accounting manipulation; they are governance concentration, an unusually long auditor relationship, and an EPS engine that increasingly leans on share count reduction rather than operating leverage.

Forensic verdict

The two yellow flags that matter most: (1) the non-GAAP EPS premium reached 25.4% in FY2025 ($20.94 vs $16.70 GAAP), driven mainly by excluded stock-based compensation of $1.94B (8.2% of revenue) — recurring "compensation expense" treated as non-recurring; and (2) share count is shrinking faster than operating income is growing, with $11.28B of FY2025 buybacks (158% of net income) acting as the dominant EPS lever rather than operating leverage. The cleanest offsetting evidence is the cash flow statement: FY2025 OCF of $10.03B exceeds GAAP net income by 41%, FCF after acquisitions of $9.85B fully funded the buyback after netting debt issuance, and trade receivables grew slower than ARR or RPO. One data point that would change the grade up to "Elevated": a quarterly slowdown in deferred-revenue/RPO growth coupled with receivables jumping faster than revenue — the classic SaaS pull-forward signal. We do not see that today.

The 13-category shenanigans scorecard

This is the standardized accounting-quality map. Every category is named; passes are stated, not omitted.

Eleven of thirteen categories pass cleanly. The four yellow items cluster around the same theme: management presentation choices (non-GAAP exclusions, new ARR cuts, buyback-driven EPS) that flatter the optics without distorting the underlying statements. None of the four warrant a position-sizing change.

Cash conversion: the strongest piece of evidence

CFO has exceeded GAAP net income in every year of the past decade by a comfortable margin. This is the single most informative chart on the page.

CFO ran below net income only once in the last decade — FY2020 — and even then it stayed above 1.0x. The FY2020 dip was a non-cash tax benefit of $1.08B that lifted GAAP NI to $5.26B without a matching cash flow. The mechanism behind the persistent CFO premium is identifiable: subscription customers prepay, so deferred revenue (a current liability) grows every year and generates an ongoing operating cash inflow even before any working-capital efficiency is captured. That is structurally repeatable, not a one-time working-capital lifeline. RPO grew 13% in FY2025 — a leading indicator that the same mechanism continues.

Stock-based compensation has grown from 7.5% of revenue to 8.2% over five years. Roughly $1.9B of SBC each year flows through CFO as a non-cash add-back — which means CFO mechanically includes a non-cash item that the income statement has charged. Adobe excludes this from non-GAAP EPS but books it inside CFO; investors should net SBC against headline cash generation if they are using CFO as a shareholder-economics proxy. SBC-adjusted CFO for FY2025 is roughly $8.1B versus the headline $10.0B.

Revenue quality: receivables and contract balances behave correctly

The classic forensic test for a SaaS business is whether trade receivables and contract assets grow faster than revenue. They do not.

DSO peaked at 40.7 days in FY2022 (a period of enterprise mix-shift and FX disruption) and has since improved to 33.9 days — collections are better than they were three years ago, the opposite of the "stretching receivables to book revenue" pattern. FY2025 receivables grew 13.1% versus revenue of 10.5%, a 2.6-point gap that is well within normal variance and explained by Q4 timing of billings disclosed in the MD&A. RPO (remaining performance obligation) — the cleaner forward signal — grew 13% versus revenue 11%, indicating bookings are outpacing revenue, not lagging it.

The Figma fee and FY25 EPS optics

GAAP EPS rose 35% year-over-year in FY2025. That is not operating leverage; it is a 2024 base distortion plus a buyback tailwind.

Net income's "28% growth" collapses to roughly 9% once the non-deductible Figma termination fee is normalized out — about in line with revenue. The remaining EPS uplift comes from the 6%+ reduction in average shares outstanding (470.9M FY22 → 426M FY25). This is not earnings manipulation; it is one-off comparability that any model should adjust. The point is that the forensic test (does operating leverage explain reported EPS growth?) returns "partially" — investors should not treat 35% GAAP EPS growth as a recurring growth rate.

Non-GAAP wedge and what is being excluded

The non-GAAP premium narrowed in FY2025 because the Figma fee fell out of the GAAP base; on a clean basis, the premium has been a consistent 25–30% for years and is dominated by stock-based compensation. Adobe's FY2026 guide is non-GAAP operating margin of ~45.0% — the GAAP equivalent at current SBC-to-revenue intensity would land closer to 36–37%. There is no metric-definition change to flag this year, but the introduction of "AI-influenced ARR" (now over one-third of the book of business) is an unaudited, internally defined construct without a clean reconciliation to reported revenue. It is a narrative metric, not a forensic one — track it, but do not underwrite to it.

Buybacks: the EPS engine that is doing more work each year

Buybacks ran at 158% of net income in FY2025. They are economically funded — FCF after acquisitions of $9.85B covers most of it, with $1.5B of net debt issued in January 2025 closing the gap — but the EPS optics depend on continued repurchase intensity. If buybacks normalized to 70% of FCF (a level still aggressive for most peers), the FY2025 share-count reduction of ~6% would be closer to 3.5–4%, and EPS growth would step down by the same magnitude. This is not a forensic flag — Adobe discloses the buyback transparently and the cash is real — but it is the single largest source of EPS growth and the part of the model most exposed to a CFO succession or capital-allocation reset.

Soft-asset and impairment risk: low

Goodwill and intangibles have not been driving the balance sheet. Adobe is not a serial acquirer.

Goodwill has been flat at ~$12.8B since the FY2021 close, after the Marketo/Magento cycle of FY2018–FY2019. Intangibles are amortizing toward zero. The Figma transaction would have added ~$20B of goodwill had it closed; its termination preserved a much cleaner balance sheet. The pending Semrush deal ($1.9B announced November 2025, expected close H1 FY2026) is a real but small addition. The risk of a goodwill impairment under a more bearish AI-disruption scenario is low to moderate in dollar terms — Adobe is not over-capitalized on prior deals.

Governance and breeding-ground read: the only material yellow

This is where the page tilts away from "Clean" toward "Watch." Three structural items deserve a name:

KPMG has audited Adobe since 1983 — 42 fiscal years. That is among the longest auditor relationships in the S&P 500 software complex. SEC and PCAOB rules require lead engagement partner rotation every five years, which mitigates the worst of the long-tenure risk, but it is still a structural yellow that the audit committee has not chosen to retest with a competitive review. The CFO transition matters more in real time: Dan Durn's planned departure (announced June 2026) overlaps a period when management has lowered ARR growth guidance and shifted to freemium acquisition — exactly the macro setup where guidance discipline and metric definitions can drift. We do not see evidence today; we are naming the risk vector.

Breeding-ground checklist

Net read: the breeding ground dampens rather than amplifies the few accounting yellow flags. There is no founder concentration, no related-party plumbing, no reverse-merger history, and the board is genuinely independent. The risk that remains is incentive design — a comp plan keyed to non-GAAP EPS and an internally-defined ARR metric does create a soft pull toward presentation choices that favor those numbers. That is a watch-item, not a verdict-changer.

SaaS-specific tests

The forensic playbook for software is: billings vs revenue, RPO/cRPO growth, deferred revenue trend, contract assets, SBC intensity, and NRR. Adobe discloses some but not all of these.

The two SaaS disclosures Adobe does not provide — net revenue retention and explicit billings — are the two most useful for a forensic check on whether the subscription base is genuinely growing or being defended by price increases on a shrinking installed base. The omission is consistent with most large legacy-to-SaaS converts and does not in itself flag manipulation, but it is the single highest-value disclosure that would upgrade or downgrade the forensic grade.

What to underwrite next

Five specific items to track over the next 4–6 quarters. These are the disclosures and signals that would actually move the forensic grade, not generic diligence boilerplate.

The single signal that would upgrade the grade to "Clean (≤20)": Adobe disclosing net revenue retention and explicit billings in the FY2026 10-K, eliminating the two largest SaaS-disclosure gaps. The single signal that would downgrade the grade to "Elevated (41–60)": a quarter in which receivables grow materially faster than revenue while RPO growth decelerates below revenue growth — meaning revenue is being pulled forward into a deteriorating bookings environment.

Closing read

The forensic risk here is a footnote, not a thesis-breaker. Adobe converts revenue to cash at 1.4x net income, runs the cleanest possible subscription P&L, and has no restatement or auditor history that argues otherwise. The realistic decision-useful caveats are non-forensic: the non-GAAP EPS run rate excludes $1.9B of recurring SBC, the GAAP EPS print is being amplified by aggressive buybacks, and management is in the middle of a CFO succession that will set the metric-definition tone for the next several years. Position size and valuation can be set on operating fundamentals; no haircut for accounting risk is warranted today, but the FY2026 Q1 reconciliation tables and the 10-K MD&A under the new CFO are worth reading line by line.