Salesforce, Inc.Full report →1 / 14
CRMNYSEThe short version

Salesforce, Inc.

Salesforce is the largest maker of customer-relationship software, sold by subscription. Its mostly recurring revenue throws off $14.4 billion of free cash flow a year, and the shares have halved from their December 2024 peak.

From a December 2024 peak of $367.87 the shares halved, bottoming near $150 in June 2026 and trading at $170.77 by mid-July — a compression in the multiple, not the cash flow.
$170.77
Share price (17 Jul 2026)
$140B
Market cap
$41.5B
Revenue (FY2026)
10.3%
Free-cash-flow yield
SwipeScroll▾
The statements

A decade of hyper-growth turned into a high-margin cash engine

FY2019 → FY2026as reported · $
Revenue$41.5B+10%
Gross margin77.7%+0.5pp
Operating margin20.1%+1.0pp
Net income$7.5B+20%
Free cash flow$14.4B+16%
As-reported figures; fiscal year ends January 31.
  • Revenue. Sales compounded at roughly 19% a year over the decade, from $8.4 billion in FY2017 to $41.5 billion in FY2026, though growth has since cooled to about 10%.
  • Margins. GAAP operating margin rose from about 3% in FY2023 to 20% in FY2026 after a deliberate cost reset — operating income climbed from $0.5 billion to $8.3 billion.
  • Cash. Free cash flow reached $14.4 billion, up from $1.7 billion nine years earlier, and has risen every single year in between.
Valuation

Priced for no growth — but the cash yield carries a stock-comp asterisk

10.3%
FCF yield on equityclears the 8% floor
8.5%
FCF yield on enterprise value
7.8%
Owner-yield after stock compjust below 8%
~2.1x
Net debt / free cash flow
  • Cheap on cash. At $170.77 the equity yields 10.3% on free cash flow and 8.5% on enterprise value — both above the 8% a cash-focused buyer wants; a reverse-DCF implies the market expects roughly 0% FCF growth forever.
  • 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.
  • Buyback math. The March 2026 accelerated repurchase struck $198.34 a share — 22% below the year's $254.21 average and its best-priced tranche, yet still above today's ~$171.
The durability question

The engine meant to replace displaced seats is still a rounding error

Agentforce ARR vs the per-seat base it must defend ($B)
Pure Agentforce ARR is about 2% of revenue; the per-seat base is about 95% — a gap near fiftyfold.
  • The engine meant to replace displaced seats — pure Agentforce ARR of about $800M, roughly 2% of revenue once $1.1B of acquired Informatica is stripped out — is still a rounding error against the ~95% of $41.5B revenue that is per-seat, even as Salesforce's own 10-K newly names AI-native rivals that can 'displace established user interfaces' and Benioff proves the mechanism by cutting his own support heads.
  • The counter. That $800 million grew 169% year over year, with 362 customers re-upping on Agentforce usage in one quarter versus three a year earlier — a base compounding at triple digits could reach materiality in two to three years.
  • What's missing. Net dollar retention on the Agentforce and Data 360 line is undisclosed, so whether usage revenue is replacing lost seats or merely adding to them cannot yet be read from outside.
The moat

Customers don't leave: attrition has held near 8% for four years

Revenue growth vs attrition (%)
The base is sticky; the growth rate is what decelerated.
  • A 92% floor. An 8% annual attrition rate means the installed base renews at about 92% of its value before a single new seat is sold — steady through a downturn, a restructuring, price rises and the AI transition.
  • Why they stay. Switching costs run deep: years of integration, AppExchange apps that don't travel, and multi-cloud data gravity. Customers running more clouds churn less than the company average.
  • The limit. The moat holds the base, not the growth rate — revenue growth still halved from the mid-20s to about 9%. Attrition would be the first number to move if the switching cost thins.
What you're buying

Five product families, none more than a quarter of the mix

Subscription revenue by product family, FY2026
Subscription and support is about 95% of the $41.5B total.
  • Recurring by design. Customers pay ahead and revenue is booked ratably over multi-year contracts, which is why the cash line is so steady.
  • Old clouds still lead. Service and Sales — the two oldest franchises — remain the largest lines; the fastest grower is Platform, Slack and Data, up about 23%, where Agentforce revenue sits.
  • Global, Americas-led. The Americas are about two-thirds of revenue ($27.2B), Europe $10.0B and Asia-Pacific $4.3B.
Cash quality

Real, consistent cash — with a stock-comp asterisk and new debt

Total debt and net debt ($B)
The March 2026 buyback traded a decade of net cash for ~2.1x FCF of net debt.
  • Twice net income. FY2026 operating cash flow of $15.0 billion was 2.0x net income, and free cash flow has risen ten years straight; capex is only about 1.4% of revenue.
  • The wedge. About $3.5 billion of the cash is stock compensation added back — a real dilution cost. Strip it out and the owner-yield falls from about 9% to roughly 7.8%.
  • Float, not debt. $24.3 billion of customer prepayments funds the business interest-free; the new $25B buyback debt is covered at roughly 5x interest, not stretched.
Growth

Growth halved to ~10%; the re-acceleration bet rides on AI

Revenue growth: mid-20s to high-single-digits (%)
The $60B FY2030 target implies holding roughly 10% for five years.
  • An expansion problem. Attrition held near 8%; what slowed is new-logo and upsell growth against a $41 billion base. Sales and Service each grew about 8%, Marketing barely 3%.
  • The AI layer. Agentforce and Data 360 ARR reached $2.9 billion by January 2026, roughly tripling year over year — but that is only about 7% of revenue, and the pure agentic slice is nearer 1%.
  • New and uncertain. The offset to seat loss is consumption pricing, which the company itself calls new and uncertain — it bills agent activity, not headcount, but makes revenue harder to forecast.
Competition

A widening CRM lead, an unsettled fight over agents

Where the agent contest stands
PlayerDisclosed AI tractionStructural edge
Salesforce (Agentforce)~$1B Agentforce ARR; $2.9B with Data 360Sits on the system of record and 92%-retained data
Microsoft (Copilot)~15M paid seats; part of a ~$37B AI businessShips agents into ~450M existing M365 seats
ServiceNow (Now Assist)Fast-growing AI ACV off a 98%-renewal baseOwns IT, HR and security workflow
AI-native startupsVenture-funded, narrow use casesNo legacy interface to defend
Metrics are not comparably defined across vendors; shown as reported.
  • Dominant where it's scored. Salesforce holds about 20% of worldwide CRM-applications revenue — more than four times its nearest rival, and its 13th straight year at the top.
  • The real threat. Not a share war in CRM but a shift in what customers buy: for the first time the 10-K names AI-native rivals that can 'displace established user interfaces.'
  • Proof it bites. Salesforce cut its own support headcount using Agentforce — evidence the technology works, and that the same math could shrink the seats customers buy from it.
Capital allocation

From serial acquirer to buying back its own stock

Capital returned: buybacks and dividends ($B)
About $35B returned since FY2023 — roughly 83% of free cash flow.
  • The pivot. After a decade buying growth — $58 billion of goodwill from Tableau, Slack, MuleSoft and Informatica — Salesforce began repurchasing stock in FY2023 and started a dividend in FY2025.
  • Returns caught up. Goodwill-inclusive return on capital rose from about 1% through FY2022 to roughly 10% — at its cost of capital, no more — as operating income grew fifteenfold from FY2022.
  • Bought above today. Every tranche was struck above the recent ~$171 quote, including the $198.34 accelerated buyback — the count is shrinking, but not timed to the lows.
The ten-year test

Three of four paths leave revenue higher a decade out

Revenue in FY2036 by scenario
Only the Displacement path — recapture treading water — roughly matches today's ~$170 price.
  • The asymmetry. A sustained decade-long decline is the minority outcome: it needs seat displacement to run unchecked for ten years against a book that renews at 92% and $35 billion contracted for next year.
  • Priced near the bear. At about $170 the enterprise trades near 11.8x free cash flow — close to 0% perpetual growth. Even the Fade path implies free cash flow near $25 billion a decade out.
  • What settles it. Whether AI agents erode the per-seat base faster than usage pricing recaptures it — the one variable the whole case turns on.
Options and volatility

Liquid, long-dated options — priced for a bumpy ride

Dec 2028
Longest listed option
~48%
At-the-money implied vol
$2.4B
Traded per day
~608k
Options open interest (contracts)
  • Clears the screen. Salesforce offers liquid, at-the-money options out to December 2028 — about two and a half years — against a stock trading roughly $2.4 billion a day.
  • Not a cheap regime. Implied vol near 47-48% sits toward the top of the stock's one-year realized range — the level you would expect after a de-rating of more than 50%, not before it.
  • No fear premium. Put and call implied vols are about equal and skew is near zero; positioning is if anything call-heavy, with a put/call open-interest ratio near 0.63.
The price

The stock halved while the cash flow it buys grew 52%

Analyst price targets vs today ($)
About 50 analysts; the $160–$475 spread is the bull-bear debate in one range.
  • A multiple, not the cash. The move from $367.87 to $170.77 is de-rating: price/FCF compressed from about 28x at the December 2024 peak to under 10x today, while free cash flow rose 52%.
  • What went wrong. The peak priced mid-20s growth and fast AI monetization; instead growth settled near 10% and the AI franchise reached only about 7% of revenue. The rating fell from growth to value.
  • The bear's answer. If agentic AI compresses the seat model, growth does not just decelerate — it inverts, and a price implying 0% growth is correct, not cheap. That is the fact the read most depends on.
What to watch

A cash machine priced for no growth — durable at the base, unproven at the slope.

Read the full report →

This distills a guided study of Salesforce built chapter by chapter — the statements, the moat, the price, and what would change the read.

Compiled from the full report · 2026-07-21 · For information, not investment advice.