Yield

Yield

Salesforce screens cheap on reported free cash flow — about an 8.8% yield after a 59% drawdown. On the framework's adjusted basis, reported FCF minus stock-based compensation minus the five-year average of acquisition spend, FY2026 free cash flow of $14.4 billion falls to roughly $5.4 billion, a 3.3% yield — some 470 to 570 basis points below even the 8–9% fortress bar. The most charitable version, stripping only stock comp, still leaves 6.7%. Consensus reversion above the bar within one to three years is low-probability on the adjusted basis.

A note on the feature file: the deterministic derivation (fit_features.adjusted_fcf) returns not_computable because the structured cash-flow feed carries no stock-comp or acquisitions line — so it could not subtract SBC and defaulted acquisitions to zero. Every adjusted figure below is derived directly from the filed 10-K cash-flow statements (FY2021–FY2026) and is captioned as such. The gap is recorded in the claims ledger.

The adjustment, line by line

Salesforce's reported free cash flow has compounded from $4.1 billion (FY2021) to $14.4 billion (FY2026) [1]. Two lines sit between that number and cash a shareholder can claim. The first is stock-based compensation, a genuine non-cash expense that the cash-flow statement adds back but that dilutes the owner: $3.5 billion in FY2026 [2]. The second is acquisition spend. Salesforce is a serial acquirer — MuleSoft, Tableau, Slack ($14.9 billion of cash in FY2022) [3], and Informatica ($8.1 billion, net, inside FY2026's $9.3 billion) [4]. The framework charges the five-year average of that spend against FCF, because for this kind of company acquisitions are how growth is bought and that cash never reaches the owner.

Source: reported FCF and SBC derived from filed cash-flow statements — FY2026 10-K [5], FY2024 10-K [6], FY2023 10-K [7]. Adjusted FCF = FCF − SBC − trailing five-year average acquisition spend; shown only for FY2025–FY2026, the years whose full five-year acquisition window falls inside the corpus.

The five-year average acquisition charge is $5,480 million for FY2026 (the mean of FY2022–FY2026 business-combination outflows: $14,876M, $439M, $82M, $2,734M, $9,268M) and $3,882 million for FY2025 (FY2021–FY2025). That gives adjusted FCF of $5,413 million for FY2026 ($14,402M − $3,509M − $5,480M) and $5,369 million for FY2025 ($12,434M − $3,183M − $3,882M). The earlier years' adjusted column is left blank because their five-year acquisition window reaches before FY2021, the oldest year the indexed filings cover.

Source: derived from FY2026 10-K Consolidated Statements of Cash Flows [8]; five-year average acquisition spend from filed cash-flow statements FY2022–FY2026 [9].

The reported figure overstates the owner's yield because roughly $9 billion of FY2026 free cash flow was consumed by stock issued to employees and cash paid for a company — neither of which is distributable.

The yield, three ways

At the current price of $170.77 (July 17, 2026) and 956 million shares, the market capitalization is $163.3 billion. Three readings of the yield follow, and they diverge sharply depending on how much of the reported number one is willing to keep.

Source: adjusted and after-SBC free cash flow derived from FY2026 10-K cash-flow statement [10]; market cap from the price feed (956M shares at $170.77, July 17, 2026).

Current adjusted yield: 3.3%. Adjusted FCF of $5,413 million over the $163.3 billion market cap. Strip only stock comp and leave acquisitions untouched, and the yield is 6.7% ($10,893 million over $163.3 billion). The headline reported yield is 8.8%.

Three-year average: about 2%. Applying the normalized $5,480 million five-year M and A charge to each year's FCF-after-SBC gives adjusted FCF of $1,231 million (FY2024), $3,771 million (FY2025), and $5,413 million (FY2026) — a three-year mean of $3,472 million, or 2.1% on the current market cap. On each year's own trailing five-year window (which lifts FY2025 to $5,369 million), the three-year average is closer to 2.5%. Either way it sits below the current single-year figure, because FY2026 free cash flow is the high point of the series.

The baseline, and whether this is a jump. The framework's own five-year adjusted-yield baseline (fit_features.yield_baseline) is not_computable — the historical adjusted-FCF series was never built, for the same reason the current one had to be derived by hand. What can be shown is the reported free-cash-flow yield at each fiscal year-end, which does carry the fortress signature Ruchir looks for: a stable 2–4% through FY2021–FY2025 that jumped to 8.8% after the drawdown.

Source: reported FCF from filed cash-flow statements [11]; year-end and current market caps from the price feed. Adjusted yield shown only for FY2025–FY2026, where the acquisition window is sourceable.

The reported line reproduces the pattern the framework hunts — a name that yielded ~2–4% now yielding ~8.8% on a forced-selling drawdown, structurally like Microsoft's 4%-to-9% move. The adjusted line tells the other half: the jump is real but small in absolute terms, from roughly 1.6% to 3.3%, and it clears no bar. The reported optics and the adjusted arithmetic point in opposite directions, and the framework runs on the adjusted number.

Which bar applies

The balance-sheet class selects the reference line, and fit_features.balance_sheet_class returns unknown only because FY2026 EBITDA was missing from the feed. It computes cleanly from the filings. Net debt is total debt of $14,439 million ($4,000M current plus $10,439M noncurrent) less $7,327 million of cash — $7,112 million; counting the $2,238 million marketable-securities book alongside cash ($9,565 million of total liquidity), net debt is $4,874 million [12] [13]. EBITDA is operating income of $8,331 million [14] plus $3,631 million of depreciation and amortization [15], or $11,962 million.

Source: balance sheet [16], operating income [17], and D and A [18], FY2026 10-K.

Net debt runs 0.4 to 0.6 times EBITDA, straddling the fortress line (≤0.5x). A year earlier, before the $6 billion debt raise for Informatica, cash and securities of $14.0 billion exceeded $8.5 billion of debt principal — a net-cash, unambiguous fortress. Salesforce is investment grade, generates $15 billion of operating cash a year, and its leverage is a recent, modest, deal-financed step. The fortress bar of 8–9% is the right reference line, with the 10% moderate bar a secondary check.

Against the 8% fortress bar, the current adjusted yield of 3.3% sits roughly 470 basis points short; against 9%, about 570 basis points short. Even the most charitable reading — 6.7%, charging stock comp but nothing for acquisitions — is 130 to 230 basis points short of the fortress line, and 330 short of the 10% bar. On no accounting does the adjusted yield reach Ruchir's entry line.

Normalization

Salesforce is not a cyclical. Revenue has risen every year for a decade, decelerating from 25% growth to about 9%, with no down year to normalize toward a mid-cycle. So there is no depressed-margin add-back to make.

The judgment call here is acquisition lumpiness, not the cycle. Two mega-deals — Slack in FY2022 and Informatica in FY2026 — account for $24.1 billion of the $27.4 billion of acquisition cash in the FY2022–FY2026 window; the intervening years ran $82 million to $2.7 billion. A skeptic can argue the five-year average of $5.5 billion overstates a normal run-rate, and that ex-mega-deals the M and A drag is closer to $1–1.5 billion. That is why the after-SBC yield of 6.7% is the fair upper bound: it charges nothing for acquisitions at all. The counter is that Salesforce has closed a multi-billion-dollar acquisition roughly every two to three years for a decade, so a five-year average is the honest run-rate rather than an outlier. At a $1.5 billion normalized M and A charge, FY2026 adjusted FCF is about $9.4 billion and the yield 5.8% — still under the fortress bar, which is why the reversal case does not change the read.

The consensus check

Consensus forward free cash flow, from CapIQ estimates, rises steadily and does clear the bar on a reported basis — 7.4% on FY2025, 8.7% FY2026, 9.1% FY2027, 10.0% FY2028, 10.9% FY2029, all on today's market cap (fit_features.consensus_forward_yield, vintage as of the run). But the vendor metric is reported free cash flow — the FY2025 consensus of $12.1 billion sits right on the reported figure of $12.4 billion — so it carries no stock-comp or acquisition adjustment. On the framework's basis, the picture is different.

Source: consensus reported FCF from CapIQ estimates via fit_features.consensus_forward_yield (data/sp/estimates.json); adjusted estimate deducts about $4B of stock comp and the $5.5B five-year M and A charge from consensus FCF — derived, illustrative.

Take FY2029 consensus FCF of $17.8 billion, subtract roughly $4 billion of stock comp (SBC has grown every year and would keep pace) and the $5.5 billion M and A charge, and adjusted FCF is about $8.3 billion — a 5.1% yield on today's price. Even stripping only stock comp and charging nothing for acquisitions, FY2029 lands at 8.4%, just touching the fortress line, and only under the assumption that Salesforce stops doing mega-deals.

So the mean-reversion underwrite is explicit: for the adjusted yield to reach the 8% fortress bar on the current $163 billion market cap, Salesforce needs about $13 billion of adjusted FCF. The path that gets there within one to three years requires two things at once — reported FCF continuing to compound toward $17–18 billion (which consensus grants by FY2029, four years out) and the market accepting a stock-comp-only adjustment, i.e. betting that the serial-acquisition run-rate falls to near zero. Given the decade-long deal cadence, I put the probability of the adjusted yield clearing 8% within three years at roughly 20–25%, and it rests almost entirely on the charitable no-more-M and A reading; on the full-framework basis (with the M and A charge intact), reversion above 8% is not on the consensus path inside three years. The reported-yield optics say fear; the adjusted arithmetic says the fundamentals do not clear the bar. What would move this: a durable collapse in stock comp as a share of revenue, or an end to multi-billion-dollar acquisitions — either would be visible in the Self-Help share-count and capital-allocation record.

FCF-to-revenue trend

Cash conversion is improving, which cuts in the company's favor and against the reflex that a serial acquirer bleeds cash. Reported free cash flow as a share of revenue climbed from 19% in FY2021 to 35% in FY2026, as operating margin expanded and capital intensity stayed near 1.4% of revenue.

Source: FCF from filed cash-flow statements [19]; revenue from reported financials, FY2021–FY2026.

Even net of stock comp, conversion improved from about 9% to 26% of revenue over the same span. A rising conversion trend is what the framework wants to see — it does not undercut the case; the shortfall is in the level of adjusted yield against price, not in the direction of cash generation. On consistency, the framework's fit_features.fcf_stability is not_computable (it needs the adjusted series it never built), but the reported record is unambiguous: free cash flow rose in every one of the last ten fiscal years, from $1.7 billion in FY2017 to $14.4 billion in FY2026, with no negative episode. Predictability, the property Ruchir prizes, is present; it is the adjusted yield on today's price that is not.