Report contents
At a glance
- Business model: Salesforce sells per-seat subscriptions across sales, service, marketing and data software, and subscription and support were 94% of FY2026's $41.52 billion of revenue. Growth comes from adding clouds to about 202,600 existing customers, which deepens switching costs; the obstacle is Microsoft bundling comparable tools while Agentforce shifts the unit of sale toward consumption.
- Financial quality: FY2026 operating cash flow of $15.0 billion against $7.46 billion of net income funded $14.4 billion of free cash flow, 34.7% of revenue, on $0.6 billion of capex. Roughly half that gap is stock compensation added back. Notes issued in March 2026 lifted gross debt to $39.8 billion, and goodwill at 54% of assets leaves tangible equity negative.
Business model
Business model: seats still pay, agents are the test
Salesforce sells subscription access to a suite of cloud applications built around a company's customer records: Sales Cloud for pipeline and account management, Service Cloud for support cases, Marketing and Commerce Cloud for campaigns and storefronts, MuleSoft for connecting other systems, Tableau for analytics, Slack for internal collaboration, and Data Cloud as the customer-data layer underneath. Agentforce, its autonomous-agent product, runs on Data Cloud and is billed per conversation or action rather than per user.
This is platform software at maturity, not a growth startup. Subscription and support produced $39.03 billion of FY2026's $41.52 billion in revenue, or 94%, up 10% year over year, Salesforce reported in its 10-K. The paying customer is a mid-market or large enterprise, usually represented by a CIO or a line-of-business executive signing a one-to-three-year contract. The people who use the software daily, sales reps and service agents, have little influence over the purchase. Revenue is weighted to North America and Europe. Salesforce does not break out cloud-level revenue in dollars, so the relative size of Sales versus Service versus Data Cloud has to be inferred from commentary; no table gives it. Agentforce revenue is not separately disclosed at all, which is the largest gap in the disclosure for anyone judging the pricing transition.
The installed base is the growth engine
Salesforce counted about 202,600 customers in its FY2026 materials, and more than half of the top 100 deals in Q4 FY2026 included six or more Clouds. That is the business model in one line. New logos are not where the growth comes from; the economics come from selling an additional module to a customer who already runs one. Each added cloud raises switching costs, because the data model, permissions and integrations are shared, and it pushes the account toward enterprise pricing.
Procurement is the counterweight. Microsoft bundles Dynamics with Teams, Azure and Microsoft 365 at a combined price a Salesforce-only bid struggles to match, giving a buying committee a credible alternative even when it prefers Salesforce. Deals rarely flip outright; they get discounted. That shows up as flat revenue per customer more than as lost logos, which is why customer counts and cloud-attach rates tell you more than the top line.
Contracted backlog is expanding at roughly the rate of revenue: current remaining performance obligation (cRPO, contracted revenue billable within twelve months) rose 11% to $29.4 billion in Q3 FY2026, and total RPO rose 12%, per the company's earnings release. An 11% backlog against 10% revenue growth means the forward book is not accelerating ahead of current sales.
Consumption pricing cuts both ways
The seat model ties revenue to customer headcount. That worked while buyers added users; it stops working when a customer hands to AI agents the work ten service reps used to do. Per-conversation pricing for Agentforce is a real change in the unit of sale. A consumption model can grow faster than seats if adoption is deep, and shrink faster if customers cap agent usage or negotiate volume tiers. Salesforce has not disclosed attach rates, average consumption per customer, or whether agent capacity is being added on top of existing contracts or substituted for seats that would have been sold anyway.
Underneath, Salesforce rents compute from AWS and Google Cloud and depends on third-party foundation models for agent capability. It owns the workflow and the customer data, not the infrastructure or the model. That supports high software gross margins but caps how much of each AI dollar Salesforce keeps, because a rising share of the cost of answering an agent query is paid to someone else.
Margin came from cost cuts, and part of the cost is stock
The margin gain since FY2024 came mostly from headcount reductions and spending discipline after activist pressure in 2023, not from pricing. FY2025 GAAP operating margin was 19.0% against a 33.0% non-GAAP figure, a 14-point gap driven almost entirely by stock-based compensation. That compensation is a real cost and is how a meaningful part of the workforce is paid. The margin level holds only while revenue growth holds; a company that trims its way to a target and then stalls has bought a level, not a trend.
What would change this view
The judgment is that Salesforce remains a sticky, high-gross-margin subscription business whose growth depends on deepening existing accounts, and whose next leg rests on a pricing model the company has not yet shown it can scale. Two observations would overturn that. If cRPO growth stays below reported revenue growth for several quarters, the backlog would be signaling that renewals are shrinking. And if multi-cloud attach at the top of the account list stops rising while customer counts hold, the expansion engine would be running out of room. Neither appears in the figures available so far.
Questions worth following
- Can Salesforce turn Agentforce pilots into paying production deployments?
Agentforce is the company's answer to seat-based pricing that AI could erode. If trials stay in test mode, Salesforce keeps selling seats to customers who need fewer of them, and the 10% growth rate has no replacement engine. Conversion would change both the revenue base and how the business is valued.
- Are existing Salesforce accounts still adding clouds at the same rate?
Growth now comes from deepening installed accounts, not from new logos. Attach rates show whether the suite still compounds switching costs or whether procurement is narrowing scope onto fewer modules, which sets the ceiling on revenue per customer and the durability of renewals.
- Is Salesforce losing renewal pricing power to Microsoft's bundle?
Microsoft can price Dynamics alongside Teams and Azure in a single enterprise agreement, which gives buyers a benchmark the account team has to beat. Deeper concessions would cap revenue per customer and slow the margin story even if logos stay in place.
Financial quality
Financial quality: strong cash, a newly levered balance sheet
Salesforce earns good-quality operating profit and converts it into cash unusually well. Over five fiscal years it lifted GAAP operating margin from 7.0% to 21.5% while gross margin rose 4.2 points to 77.7%. Fiscal 2026 produced $15.0 billion of operating cash flow against $7.46 billion of net income, and free cash flow of $14.4 billion was 34.7% of revenue on capex of just $0.6 billion. Two things complicate that picture: a meaningful slice of the cash is compensation paid in stock, and a debt-funded buyback has changed the balance sheet without changing the business.
| USD bn unless noted | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 | FY2027 Q2 | Assessment |
|---|---|---|---|---|---|---|---|
| Revenue | 26.49 | 31.35 | 34.86 | 37.89 | 41.52 | 11.35 | Adequate |
| Revenue growth (%) | — | 18.3 | 11.2 | 8.7 | 9.6 | 10.8 | Adequate |
| Gross margin (%) | 73.5 | 73.3 | 75.5 | 77.2 | 77.7 | 76.7 | Exceptional |
| Operating margin, GAAP (%) | 7.0 | 9.3 | 14.4 | 19.0 | 21.5 | 20.5 | Strong |
| Net margin (%) | 5.4 | 0.7 | 11.9 | 16.4 | 18.0 | 31.1 | Strong |
| ROE (%) | 2.9 | 0.4 | 7.0 | 10.3 | 12.4 | 9.7 | Adequate |
| ROIC (%), after tax, pre-interest | 2.3 | 3.7 | 6.5 | 9.3 | 10.6 | 10.5 | Adequate |
| Operating cash flow | 6.00 | 7.11 | 10.23 | 13.09 | 15.00 | 1.27 | Exceptional |
| Free cash flow | 5.28 | 6.31 | 9.50 | 12.43 | 14.40 | 1.10 | Exceptional |
| FCF margin (%) | 19.9 | 20.1 | 27.3 | 32.8 | 34.7 | 9.7 | Strong |
| Operating cash flow / net income (x) | 4.2 | 33.9 | 2.5 | 2.1 | 2.0 | 0.4 | Strong |
| Capex / revenue (%) | 2.7 | 2.6 | 2.1 | 1.7 | 1.4 | 1.5 | Exceptional |
| Liabilities / assets (%) | 38.9 | 41.0 | 40.3 | 40.6 | 47.3 | 65.0 | Weak |
| Current ratio (x) | 1.05 | 1.02 | 1.09 | 1.06 | 0.76 | 0.84 | Adequate |
| Goodwill / assets (%) | 50.4 | 49.1 | 48.7 | 49.8 | 51.6 | 54.0 | Weak |
| Operating income / interest (x) | 8.5 | 10.1 | n/d | 26.7 | 27.9 | 5.0 | Adequate |
*Single quarter ended July 31, 2026. Flow ratios are quarterly and not comparable with full years; ROIC is annualized. n/d = not disclosed. Fiscal 2023's near-zero net income makes that year's ROE and cash-conversion ratios meaningless.
Margins: real progress, partly paid in shares
The operating margin gain is mostly cost discipline, not pricing: gross margin added 4.2 points while operating margin added 14.5. The gap between accounting bases is wide, though. Salesforce reported a 34.1% non-GAAP operating margin against 20.5% GAAP in the July 2026 quarter, so roughly $1.5 billion of the quarter's revenue sits in stock compensation and amortization of acquired intangibles. That is why operating cash flow runs at about 2x net income: the add-back is genuine cash, but the expense returns through the share count, and the buyback is what neutralizes it. Treating the 2x conversion as pure operating superiority overstates the case.
Leverage: the buyback changed the risk, not the business
Salesforce issued $25 billion of notes in March 2026 to fund an accelerated repurchase. Gross debt went from about $9.0 billion in October 2025 to $39.8 billion by April 2026, against $8.31 billion of cash in July. Interest expense was $0.47 billion in the July quarter alone versus $0.32 billion for all of fiscal 2026, a run rate near $1.9 billion and coverage of roughly 5x. Equity fell from $59.1 billion in January to $38.4 billion in July; with goodwill at $59.3 billion, tangible equity is about negative $21 billion. ROE will drift up simply because the denominator shrank, so ROIC near 10.5% is the better gauge, and it sits only a couple of points above a plausible cost of capital.
Working capital: a sub-1 current ratio that is mostly deferred revenue
The current ratio fell to 0.76 at fiscal year-end and 0.84 in July; the quick ratio is identical because there is no inventory. Current liabilities of $26.3 billion include deferred revenue that is settled by delivering service, not cash, so a ratio below 1 is structural here. Seasonality confirms it: operating cash flow was $6.7 billion in the April 2026 quarter and $1.27 billion in the July quarter, reflecting annual billings. Receivables turnover has held at 3.0–3.2x, about 115 days, for four years, so collections are not stretching. A drop below roughly 2.5x, or a shrinking deferred revenue balance, would change that read.
| USD bn unless noted | Q1 FY26 (Apr 2025) | Q2 FY26 (Jul 2025) | Q3 FY26 (Oct 2025) | Q4 FY26 (Jan 2026) | Q1 FY27 (Apr 2026) | Q2 FY27 (Jul 2026) | Assessment | Trend |
|---|---|---|---|---|---|---|---|---|
| Revenue | 9.83 | 10.24 | 10.26 | 11.20 | 11.13 | 11.35 | Adequate | Rising about 11% |
| Net income | 1.54 | 1.89 | 2.09 | 1.94 | 2.11 | 3.53 | Strong | Inflated by investment gains in Q2 FY27 |
| Operating cash flow | 6.48 | 0.74 | 2.32 | 5.46 | 6.70 | 1.27 | Strong | Seasonal, first-quarter heavy |
| Total assets | 98.61 | 97.57 | 95.14 | 112.30 | 106.68 | 109.62 | Adequate | Stepped up $17bn in Q4 FY26 |
| Total liabilities | 37.94 | 36.24 | 35.12 | 53.16 | 72.45 | 71.24 | Weak | Jumped with new debt |
| Net PP&E | 5.26 | 5.18 | 5.10 | 5.12 | n/d | n/d | Adequate | Flat, capital-light |
| Current ratio (x) | 1.07 | 1.12 | 0.98 | 0.76 | 0.79 | 0.84 | Adequate | Fell with debt and deal funding |
Acquisitions: the capital goes out the deal door
Capex of 1.4% of revenue makes Salesforce look asset-light, but investing cash outflow was $8.59 billion in fiscal 2026, and the company has since closed Informatica (which added $456 million of July-quarter revenue) and Fin, agreed to buy Contentful, and on October 20, 2026 agreed to buy Braze for $5.66 billion in cash, subject to shareholder and regulatory approval. Reported revenue growth of 10.8% in the July quarter leans on Informatica; organic growth is closer to 7%, an estimate from the disclosed contribution. The return on this spending is hard to audit because Salesforce does not disclose acquired-unit profitability, and it lands in goodwill rather than capex.
What to watch
- Non-operating gains. July-quarter net income of $3.53 billion exceeded operating income of $2.33 billion because of $2.61 billion of gains on strategic investments, which Salesforce reported. Similar gains in coming quarters would keep flattering EPS in the $16.67–16.71 fiscal 2027 non-GAAP guide; their absence leaves a cleaner, lower base.
- Guidance versus cash. Management guided fiscal 2027 free cash flow growth of 4–5% on $46.1–46.4 billion of revenue. If cash generation stalls while the buyback and Braze cash payment proceed, net debt near $28 billion rises further. The 10-Q note on maturities is where to check the refinancing calendar.
- Goodwill headroom. Goodwill at 54% of assets with negative tangible equity means impairment testing is the place to see whether acquired businesses are meeting underwriting assumptions; an impairment would not affect cash but would confirm the deals were priced richly.
Questions worth following
- Can Salesforce fund buybacks and acquisitions from free cash flow?
Free cash flow of $14.4 billion in fiscal 2026, guided to grow 4–5% in fiscal 2027, must absorb a $25 billion buyback, the $5.66 billion Braze deal and roughly $1.9 billion of annual interest. Whether cash generation carries that load decides how much more debt the balance sheet takes on.
- How much of Salesforce's fiscal 2027 profit is recurring?
Fiscal 2027 guidance of $16.67–16.71 in non-GAAP EPS leans on gains from strategic investments that will not repeat on a schedule. Separating operating profit from mark-to-market gains decides whether the earnings base is a 34% non-GAAP margin business or a lower-quality one.
- Are Salesforce's acquisitions earning their cost of capital?
Goodwill is 54% of assets and the company has agreed four acquisitions in five months. Returns on that spending decide whether acquired revenue is worth its price. Organic growth near 7% against 11% reported suggests deals, not the base business, are carrying the growth.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Key risksWhat could break the investment case
- Valuation and margin of safetyWhat the current price assumes, and how much room for error it leaves
- SummaryOur overall judgement and what to watch next
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Salesforce Inc
Salesforce Inc
Business, moat and key questions
Open company page