Chapter 4

Spending 1.8x Free Cash Flow

The margin chapter left the reader with a company that turns reported earnings into cash at an unusually high rate — $45.70 billion of free cash flow over FY2021 through FY2025, about 1.32 times net income, under a clean audit. This chapter follows that cash out the door. For five years it went to the same three destinations in almost the same proportions, on a formula so steady it reads like a rule. In FY2026 the rule is being broken on both sides at once: the company plans to spend roughly 1.8 times its free cash flow, take on debt for the first time in its public life, and shift the character of what it buys from people to software.

The formula that ran for five years

Over FY2021 through FY2025, Accenture generated $45.70 billion of free cash flow and put $21.29 billion of it into share purchases, $14.46 billion into dividends and $2.94 billion into property and equipment, while spending a further $18.20 billion on acquisitions of businesses and investments [1] [2]. Shareholder returns — buybacks plus dividends — came to $35.75 billion, or 78.2% of free cash flow, and the annual figure barely moved: 70.7%, 74.5%, 79.6%, 90.1% and 76.5% across the five years. Returns plus acquisitions ran to $53.96 billion, about 118% of free cash flow; the gap above internal cash was covered by $6.69 billion of proceeds from employee share issuance and, latterly, by debt.

Source: derived from the Consolidated Cash Flows Statements, FY2021–FY2025 annual filings [3] [4].

The steadiness is the point. Accenture publishes no target payout ratio, no return-on-invested-capital disclosure and no acquisition hurdle rate; what it publishes each September is a dollar floor for shareholder returns, and that floor has been met or beaten every year on record — at least $6.3 billion promised and $6.6 billion returned in FY2022, rising to at least $8.3 billion promised and $8.3 billion returned in FY2025, with the quarterly dividend lifted roughly 10% a year to $1.63 [5]. The floor is honoured mechanically, which is what makes the two components underneath it worth separating: one of them did far less than its headline suggests, and the other is where the money and the risk actually sit.

The buyback that barely moved the count

The $21.29 billion of share purchases over five years reduced the diluted share count from 645.91 million to 632.44 million — a cumulative 2.1% [6]. The reason the largest discretionary use of cash produced the smallest change in the share base is that a second stream of shares ran the other way. Share-based compensation expense over the same five years was $8.97 billion, and roughly 41.2 million Class A shares were issued to employees, bringing in $6.69 billion of cash equal to nearly a third of the gross buyback. And of the $4.61 billion Accenture spent on its own shares in FY2025, $776 million was share withholding for employees' payroll taxes on vesting equity — purchases that, by the company's own disclosure, do not consume repurchase authorization at all [7].

The buyback, in other words, has functioned mostly as an offset to dilution from employee equity rather than as a shrinking of the base. That distinction sharpens against the price paid. The 57.3 million shares Accenture bought in the open market over FY2021 through FY2025 cost $17.42 billion, an average of $303.83 per share; the pace then continued into a falling market, at $245.32 in the first quarter of FY2026, $246.09 in the second and $198.84 in the third [8] [9] [10] [11]. The close on 31 July 2026 was $165.92 — below every one of those averages.

Source: Note 14 share-purchase tables, FY2021–FY2025 Forms 10-K, and quarterly earnings calls; the 31 July 2026 close of $165.92 (dashed reference in text) sits below all eight bars [12] [13].

Source: derived from Note 14 share-purchase tables and data/prices/daily.json [14].

As arithmetic, the 57.3 million shares bought for $17.42 billion are worth $9.51 billion at the 31 July close — a $7.9 billion decline in the value of what was purchased, before any judgment about the decisions. That the pace held into the decline is a choice management made explicitly: on 23 June 2026, five days after cutting the year, the board added $2 billion to the FY2026 repurchase program, taking it to $7.5 billion and a 62% increase over FY2025, with the CEO stating that "we do not believe our current share price reflects that position or the strength of our business fundamentals" [15].

That top-up also underlines how the authorization now works. The board has authorized an aggregate $54.1 billion for repurchases since 2001, but the available balance is refilled in roughly annual tranches: $2,851 million at 31 August 2025, topped up by $5,000 million that September, drawn to about $3.2 billion by 31 May 2026 and expected to leave roughly $1 billion after Q4 [16] [17]. At the current pace a $5 billion authorization lasts under four quarters, so future buybacks depend on board top-ups rather than standing headroom.

Acquisitions no one priced

The $18.20 billion spent on acquisitions over five years is disclosed in a single sentence per year. Note 6 of each 10-K reports that Accenture "completed a number of individually immaterial acquisitions," gives one line each for total consideration, goodwill and intangibles, and names no target, no purchase price, no acquired revenue and no acquired margin [18]. Between 82% and 90% of consideration lands in goodwill every year, and goodwill reached $22.54 billion at 31 August 2025 — 69.9% of the company's $32.24 billion of total shareholders' equity — carried without a single impairment in FY2024 or FY2025 [19]. The only disposition of the period was the FY2022 exit from Russia, a $96.3 million non-operating loss.

Because no deal is priced, the only way to judge the program is against management's own measure of what it bought: the inorganic contribution to revenue growth. Lined up by year, that measure exposes the awkward pairing at the centre of the record — the biggest spending year was the weakest growth year. FY2024 saw $6.58 billion deployed, the largest acquisition outlay in the company's history, in the year revenue grew 1.2%, the slowest of the five; management described it as "$6.6 billion in strategic acquisitions" delivering "2% growth in local currency" [20].

Source: Consolidated Cash Flows Statements and quarterly guidance; revenue growth from reported financials [21] [22] [23].

A note on the numbers behind that table: the structured financials feed carries acquisition cash as zero for every year, contradicting the cash-flow statements, so every acquisition figure here is read from the filings themselves rather than the data series. Management's inorganic contribution has ranged from about 5% in FY2022 down to about 1.5% guided for FY2026 — the only published gauge of whether $18.20 billion of deals earned their keep, and a self-reported one.

The break

For five years the framework flexed acquisitions inside the residual left after the dividend and the buyback, and self-funded the whole of it. FY2026 raises both legs at the same time. Acquisition guidance, given at about $3 billion in September 2025, moved to about $9 billion by 18 June 2026 after a run of deals; shareholder-return guidance went from at least $9.3 billion to at least $9.5 billion, and the 23 June buyback increase pushed planned repurchases to $7.5 billion, implying roughly $11.5 billion of total returns [24] [25] [26]. Set against guided free cash flow of $10.8 billion to $11.5 billion, the combined roughly $20.5 billion of planned uses is about 1.8 times the cash the business is expected to generate [27].

Source: derived from the FY2021–FY2025 cash flow statements and FY2026 guidance; "returns and M&A" is buybacks plus dividends plus acquisitions [28] [29].

The gap is being funded two ways. The cash balance is being drawn down — from $11.5 billion at 31 August 2025 to $10.2 billion nine months later — and the balance sheet, run at net cash for two decades, has become a funding source [30]. Accenture carried no long-term debt through FY2023; on 4 October 2024 a finance subsidiary issued $5.0 billion of senior unsecured notes in four tranches maturing 2027 to 2034, with no financial covenants, and on 18 June 2026 the CFO said the company expects to access the long-term debt market again for its elevated acquisition outlook, while keeping "a strong investment grade credit rating with a low net leverage ratio" [31] [32]. No target leverage ratio has been published; that qualitative phrase is the only quantitative-adjacent commitment on record.

What changed is not only the size of the acquisition budget but the kind of thing it buys. The FY2026 deals move away from services tuck-ins toward software and product platforms: a majority stake in Dragos plus all of runZero and NetRise in operational-technology security, and Ookla, a 430-person business with about $231 million of 2025 revenue earned from subscriptions and licensing [33]. Management frames the intent as moving revenue toward models that are not billed by the hour. That is the first allocation move in five years capable of changing the labour arithmetic laid out in The 800,000-Person Payroll rather than adding to it — and because none of these deals carries a disclosed price, whether it reshapes the model or simply buys revenue at unknown multiples is a question this chapter can pose but not settle. It belongs to What $165.92 Is Asking.

Who is making these calls

The people directing this cash are a fully home-grown bench: every executive officer named in the FY2025 10-K is an internal promotion of twelve to thirty-nine years' standing [34]. They answer to a conventional-to-strong board with no control block — ten directors, nine of them independent, Julie Sweet as combined chair and CEO with Arun Sarin the independent Lead Director since February 2025, all committees fully independent and a retirement age of 75 [35]. The only holders above 5% are index managers — Vanguard at 10.4% and BlackRock at 7.8% — with Accenture itself holding 6.8% of the Class A shares as non-voting treasury [36].

What that board and bench do not have is much of their own money in the stock. As of 1 December 2025, all 21 directors and executive officers together beneficially owned 144,910 Class A shares — roughly 0.02% of the base against which Vanguard's stake is 10.4% [37] [38]. Sweet held 6,906 shares, of which 2,339 are units deliverable within sixty days, leaving about 4,567 owned outright — worth roughly $0.8 million at the July close, against the $19.9 million she realized on vesting equity in fiscal 2025 alone [39] [40]. The ownership requirement meant to supply alignment is set at six times base salary but is satisfiable with "unvested equity," so it clears on the forward grant stream before any purchased share is counted; the independent directors' own guideline, by contrast, counts only shares held and vested awards — the company draws the distinction for its directors that it waives for its officers [41].

Two design choices in the pay machinery shape how that grant stream is earned, and both were resolved in management's favour in a year of poor shareholder returns. The annual bonus is explicitly non-formulaic: the scorecard is weighted 60% financial and 40% strategic, but the proxy states the company applies no formula and no predetermined weighting and that no single objective is material, so the "exceeds" rating that produced Sweet's raised $4.5 million fiscal-2025 bonus cannot be reconstructed from disclosed inputs [42]. That "exceeds" was awarded in the same year the proxy discloses a three-year total shareholder return at the 3rd percentile of the compensation peer group, with $100 invested in Accenture on 31 August 2020 worth $117 five years later against $251 for the S&P 500 Information Technology index [43] [44].

The second choice is the relative-TSR hurdle inside the largest equity award — the 25% of it not governed by the adjusted operating-income metric examined in Ten to Thirty Basis Points. After the return components of two consecutive grants paid out at 0%, the committee, for awards granted in January 2026, lowered the entry threshold from the 40th to the 30th percentile and replaced a sixteen-company technology-and-consulting comparison group with the whole of the S&P 500 [45]. Both changes reduce the return needed to earn the same equity, and both took effect after the returns that would have failed the old bar.

The counterweight is that the machinery has, on the reported side, bitten. Compensation actually paid to Sweet fell from $37.4 million in FY2021 to $16.6 million in FY2025 — down 56%, and 44% below her $29.6 million reported package that year — as the two zero relative-TSR payouts flowed through [46]. Realized pay tracked the stock down even as headline grant values held up, so the same disclosures that show generous targets also show the targets going largely unmet. What a reader is left with is not a governance defect — there is no promoter, no dual-class control, no structural flaw — but a set of discretionary levers, each adjusted at the moment the stock made the prior settings expensive.

The allocation formula, the buyback that offsets dilution at prices above today's, the acquisitions that carry no price tag, the FY2026 break to 1.8 times cash flow, and the bench and board steering all of it now sit on the table. The final act prices them: what the $165.92 share embeds, what six months of the drawdown did and did not change in the estimates, and the dated events that will test the questions this record leaves open.