Cash Quality

Cash Quality

Salesforce turns reported profit into cash at roughly twice the rate of net income, and free cash flow has risen for ten straight years. The quality is real but not pristine: about $3.5 billion of that cash-over-profit gap is stock-based compensation, a genuine cost paid in dilution. And in March 2026 the company borrowed $25 billion to buy back stock, moving a near-net-cash balance sheet to about two times free cash flow of net debt. Customer prepayments, not borrowings, still fund the operation.

Earnings that become cash

Salesforce's cash-flow record is consistent. Operating cash flow and free cash flow have each risen every year for a decade, through the FY2020 profit trough, the FY2022–FY2023 growth slowdown, and the FY2024 restructuring. Free cash flow went from $1.7 billion (FY2017) to $14.4 billion (FY2026) without a single down year [1]. Because capital expenditure runs near 1.4% of revenue, operating cash flow and free cash flow track each other closely — this is a business where heavy capex does not depress the free-cash-flow line, so the consistency the reader looks for is not being flattered by an under-investment choice that will reverse.

Loading...

Source: Consolidated statements of cash flows and operations, FY2017–FY2026 10-Ks; FY2026 figures per the FY2026 10-K [2].

In FY2026, operating cash flow of $15.0 billion was 2.0 times net income of $7.5 billion, and free cash flow of $14.4 billion was 1.9 times [3]. A conversion rate that high is not, by itself, a mark of quality — it has to be explained. Three items do most of the work: $3.6 billion of depreciation and amortization (much of it amortization of intangibles from past acquisitions), $3.5 billion of stock-based compensation, and a $2.9 billion increase in unearned revenue [4]. The first two are non-cash charges the cash-flow statement adds back; the third is customers paying ahead. None of them is an aggressive accrual, but two of them mean the cash figure sits above true economic earnings.

The stock-compensation wedge

Stock-based compensation is the item that most separates Salesforce's cash from its economics. The $3.5 billion charged in FY2026 is 8.5% of revenue [5]. Cash flow adds it back because no cash leaves the company when equity vests — but shareholders still pay for it, in dilution. Subtract it from free cash flow and the FY2026 figure falls from $14.4 billion to about $10.9 billion, and the roughly 9% headline free-cash-flow yield established in The Cash Machine drops to close to 7% — below the 8% threshold a cash-focused buyer typically wants.

Reported FCF ($B)

$14.4

Stock-Based Comp ($B)

$3.5

FCF less SBC ($B)

$10.9

SBC-Adj. FCF Yield

6.9%

Source: FY2026 10-K, statement of cash flows [6]; yield derived against the ~$158 billion market value in The Cash Machine.

The counter to that adjustment matters as much as the adjustment itself. Salesforce is not letting the share count drift upward: repurchases have exceeded the dilution, so the stock-compensation cost is being paid for in cash through buybacks rather than borne silently by holders. That is a better outcome than most software companies deliver, and it connects directly to the balance-sheet change below. But the accounting is unambiguous — the true owner-cash yield on this business is nearer 7% than 9%, and a reader who takes the reported free-cash-flow figure at face value is crediting Salesforce with about $3.5 billion of cash that is, in substance, a wage bill settled in shares. Seen whole, the compensation wedge and the balance-sheet swing are two sides of one trade: The market prices Salesforce's free cash flow to grow near zero forever, but that cash flow is overstated by ~$3.5B of add-back stock compensation — leaving a ~7-8% true owner-yield — and management met the low price by borrowing $25B to repurchase stock at $198.34, above today's ~$171, levering a near-net-cash balance sheet to ~2.1x free cash flow.

Deferred revenue: float, not debt

The item that makes Salesforce's cash consistent rather than lumpy is unearned revenue — customer prepayments for subscriptions not yet delivered. At January 31, 2026 the balance was $24.3 billion, equal to about 214 days of revenue and far above the 30-day level at which a deferred balance starts to read as a genuine forward-demand signal [7]. It is also growing: billings ran to $45.1 billion in FY2026, up 13.8% year on year and ahead of the 9.6% reported revenue growth [8]. Roughly half of all revenue recognized in FY2026 came out of the unearned balance carried into the year [9]. Billings running ahead of revenue is a modest positive: the book of paid-for, not-yet-delivered service is expanding faster than the income statement.

No Results

Source: FY2026 10-K, unearned revenue rollforward [10] and consolidated balance sheet [11].

A specific reader question — whether the net-debt calculation should treat deferred revenue as debt — resolves cleanly: the answer is no, because unearned revenue is discharged by delivering software the company already runs at near-zero incremental cost, not by paying cash. It is interest-free customer float. The right way to see it is that Salesforce runs on negative working capital: customers fund the business in advance. If one wrongly added the $24.3 billion of unearned revenue to borrowings, reported net debt would balloon past $30 billion; that would misread a funding advantage as a liability.

The balance sheet just changed

For most of the past decade Salesforce carried little net debt; at January 31, 2025 it held slightly more cash and equivalents than borrowings [12]. Two events changed that. In November 2025 the company drew $6 billion of bank debt to help fund the roughly $9.3 billion cash acquisition of Informatica [13]. Then in March 2026 it issued $25.0 billion of new senior notes and used the net proceeds to fund a $25 billion accelerated share repurchase, while refinancing the Informatica bridge with a $6 billion term loan [14]. Total debt principal went from $14.5 billion at January 31, 2026 to $39.5 billion by April 30, 2026 [15].

Loading...

Source: FY2026 10-K balance sheet [16] and Q1 FY2027 10-Q balance sheet and debt note [17]. Net debt is total debt less cash and equivalents; including marketable securities trims it by roughly $3 billion.

The result, as of April 30, 2026, is net debt of about $30 billion against cash and equivalents of $8.9 billion — roughly 2.1 times FY2026 free cash flow, or nearer 1.9 times counting marketable securities [18]. That is a real change from the fortress balance sheet described at the year-end snapshot, but it sits inside the leverage a high-free-cash-flow business can carry: interest on the enlarged debt load runs to roughly $1.8 billion a year against $8.3 billion of operating income and $14.4 billion of free cash flow, so coverage falls from about 26 times to roughly 5 times operating income — lower, not stretched [19].

The buyback that drove this is aggressive on price as well as size. The accelerated repurchase delivered about 103 million shares at an average $198.34, cutting shares outstanding from 929 million to 819 million in a single quarter [20] and lifting quarterly diluted earnings per share to $2.42 from $1.59 a year earlier [21]. The $198.34 average was struck well below the December-2024 peak but still above the ~$171 the shares fetched by mid-2026 — the same pattern seen in the FY2026 open-market buybacks: management is shrinking the count and exploiting the selloff, but has not been buying at the lows.

The shape of the debt

The duration of that debt is barbelled, and the two ends were financed in very different worlds. The $8.5 billion of notes issued in 2021 carry coupons of 1.50% to 3.05% and run out to 2061 — cheap, long money locked in at the bottom of the rate cycle, worth only about $6.7 billion at year-end fair value [22]. The March-2026 tranche, by contrast, was priced at 4.50% to 6.70% across maturities from 2028 to 2066 [23]. There is very little near-term repayment pressure: only $6 billion matures before 2029, comfortably inside a single year's free cash flow.

Loading...

Source: Q1 FY2027 10-Q, debt note; 2021 vintage carries 1.50%–3.05% coupons, the 2026 vintage 4.26%–6.70% [24].

Days, receivables and current liabilities

Two of the reader's forensic checks — a rise in days-sales-outstanding, and a build-up of current liabilities — flag on the year-end numbers and then dissolve on inspection. Receivables of $14.3 billion at January 31, 2026 were up 20% on a year when revenue grew 9.6%, lifting year-end days-sales-outstanding to about 126 days from 115 [25]. Taken alone that looks like slowing collections. It is not: Salesforce bills disproportionately in its January-ending fourth quarter, so the balance-sheet date catches receivables at their seasonal peak, and the first quarter is its largest collection quarter [26]. The proof is in the next quarter: by April 30, 2026 receivables had fallen to $5.1 billion, a $9.4 billion collection in three months [27].

Loading...

Source: FY2026 10-K balance sheet [28] and Q1 FY2027 10-Q balance sheet [29].

The current-liability build tells the same seasonal-and-structural story. Total current liabilities rose from $28.0 billion to $37.1 billion over FY2026, but the increase is not a stretch of supplier payables: roughly $4 billion was the Informatica bridge loan reclassified as current, about $3.6 billion was the growth in unearned revenue already discussed, and the rest was accrued compensation and acquired liabilities [30]. By April 30 the figure had fallen back to $27.5 billion as the bridge was refinanced into long-term debt [31]. There is no days-sales-of-inventory to track — a software company holds no inventory — and no sign of payables being stretched to manufacture cash flow.

What the book value does and doesn't say

One balance-sheet lens the reader asks for — tangible book value — returns almost nothing here, and the reason is itself the finding. Of $112.3 billion in total assets at January 31, 2026, $57.9 billion is goodwill and a further $6.8 billion is acquired intangibles: together 58% of the balance sheet is the accounting residue of acquisitions — Slack, MuleSoft, Tableau, and now Informatica [32]. Strip those soft assets out and tangible common equity is negative — roughly minus $5.6 billion — so a tangible price-to-book multiple is not a meaningful gauge for this company. The intangibles amortize over a weighted 5.4-year life and flow through the depreciation-and-amortization add-back that inflates the cash-conversion ratio [33]. Salesforce has not recorded a goodwill impairment; it tests the balance annually in its fourth quarter [34].

Book value understates rather than overstates this business: the assets that matter — a recurring, prepaid contract base and the software behind it — are worth far more than their carrying value, while the goodwill that dominates the balance sheet is worth whatever the acquired revenue keeps earning. For a subscription compounder, price-to-free-cash-flow is the lens that works; tangible book is not. What the reader should carry forward from this chapter is narrower and firmer: the cash is real and unusually consistent, it flatters economic earnings by about $3.5 billion of stock compensation a year, the deferred-revenue float is a funding asset rather than a debt, and the balance sheet has just traded a decade of net-cash conservatism for roughly two times free cash flow of net debt in the service of buying back its own shares.