Salesforce keeps a company's record of its customers
Sales
staff track every deal in it
Service
agents answer customers' cases
Slack
moves information between systems
Salesforce rents big companies the software they run sales, service and marketing on. Its newest product, Agentforce, adds AI agents to the same software in pricier editions.
The company · 2 · What exactly does it sell?
Two-thirds is the main apps; a third is data tools
Each $1 of sales, second quarter fiscal 2027
63%32%5%
Sales, service, marketing, SlackData and connection toolsSetup and consulting
The main apps grew 8%. The data tools grew 20%, mostly because Salesforce bought Informatica.
The company · 3 · How does it make money?
Customers rent it per user, a year ahead
$10.8Bof the quarter's $11.3B from subscriptions
$33.5Balready contracted for the next 12 months
Customers pay per user, sign contracts usually for a year or more, and pay ahead. No customer is a tenth of sales.
Is it a good business? · 4 · What gives it an edge?
Years of records make leaving costly
About 8%
of contract value lost a year
20.0%
of its market in 2025, IDC
20.7%
its share the year before
A large customer builds years of records, custom fields and links inside Salesforce. Leaving means rebuilding all of it. AI coding makes that rebuild cheaper.
Is it a good business? · 5 · What is the catch?
The new rival is the customer itself
$600,000a year Curative stopped paying, building its own in two months
Salesforce's own annual report now lists AI-native companies among its rivals.
Is it a good business? · 6 · What does its record show?
Profit now grows only as fast as sales
Sales, $B
26.5
31.4
34.9
37.9
41.5
43.9
Jan 2022Jan 2023Jan 2024Jan 2025Jan 2026last 12 mo
Sales$26.5B → $43.9BJan 2022 to the last 12 months
Profit per $1Near 20 centsoperating profit per $1, flat since 2025
Interest$473Ma quarter, after borrowing to buy
Cutting selling costs in fiscal 2025 made it far more profitable. Since then each cost takes the same share of each dollar.
Is it a good business? · 7 · How much room is there to grow?
It already holds a fifth of its market
2,666,000US customer service jobs in 2025
-5%the government's expected shrink by 2035
The main business grows with its market. Fewer service jobs means fewer users to charge, and nobody yet counts the market for AI agents.
Is it a good business? · 8 · Who is in charge?
A co-founder still in charge, paid on beating the market
Since 2001Marc Benioff as chief executive
About 3.5%of the shares owned by insiders, mostly Benioff
103.3%of target bonus paid, on subscription sales and adjusted profit
Robin Washington, once Gilead's finance chief, took over finance in March 2025.
The story right now · 9 · What does AI do to office software?
AI agents could do work people now do in the software
AgentforceSalesforce's own AI agents
AI agents can answer cases and update records by themselves. That could sell pricier editions, or cut the number of people paying per user. Fear of the second nearly halved the shares from December to June.
The story right now · 10 · How much of the growth is bought?
Growth from buying a company is not the same as growth
Sales rose 11%The headline figure.
Part came from InformaticaA company it bought.
Without it, about 6.4%What customers added.
Buying cost interest$473M a quarter.
Buying a company adds its sales overnight, but the business underneath grows no faster. That is why we watch growth without acquisitions.
The story right now · 11 · What changed lately?
Sales rose 11%, about 6.4% without Informatica
$11.3Bsecond-quarter sales, up 11%
+8%the main apps, no faster than a year ago
$1.5BAgentforce, on a wider definition
+22.6%the shares in one day after results
Mixed quarter
The quarter showed users still growing, which eased the AI fear for now. Part of the profit beat came from gains on its Anthropic stake.
The story right now · 12 · What is everyone arguing about?
Is Salesforce cheap, or fair once costs count?
Many investors think
Cheap on adjusted earnings, with AI fears overdone
We think
A good business at a fair price, not a cheap one
The cheap reading includes Anthropic gains and leaves out share pay and new interest. Against us: growth without acquisitions climbing toward 9%.
The price and our view · 13 · What is a share worth?
The price sits close to our value
$218what we think a share is worth · the last close is of that
$218our value
Attractive below $174
Neutral
Unattractive above $261
$104gloomy: growth halves and profit per sale slips
$218our estimate: experts' forecasts, then slowing
$293sunny: growth runs a quarter faster
Value per share in three futures. None is a prediction of the share price.
We estimate what a share is worth from the cash the business should earn over many years. Our rule: buy only when the price is at least a fifth below our estimate, as a cushion for being wrong.
The price and our view · 14 · How could we be wrong?
Customers building their own, and agents cutting users
Customers build their ownBiggest
AI coding replaced one contract in two months.
Agents cut paid usersWatch
It charges per user.
Growth is partly boughtShowing now
About 6% without acquisitions.
Paid usersUnknown
Management gives no number.
Agentforce's $1.5B is a spoken figure on a widened definition, not a filed one.
The price and our view · 15 · How will we know?
December shows whether growth speeds up
Growth without acquisitions near 7%Good sign
The main apps growing 8% or moreGood sign
The main apps growing less than 6% a yearWe turn negative
DEC 20262
Third-quarter fiscal 2027 results. Companies report every three months; each report is a test of the story.
Our view also changes on price alone: we turn positive if the shares fall a fifth below our estimate, and negative if they rise a fifth above it.
The price and our view · 16 · So, what do we think?
The customer-software leader, at a fair price
$218our value
Attractive below $174
Neutral
Unattractive above $261
NeutralMedium conviction
Salesforce rents the software big companies run sales and service on, and customers rarely leave. But growth without acquisitions is about 6%, and AI tools let some customers build their own. Neutral.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
Salesforce keeps a company's record of its customers and runs the work around them. Sales staff track deals in it, service agents answer cases, marketers send campaigns, and Slack and its data tools move information between systems.
Customers rent it per user per year. No customer is a tenth of revenue. The alternatives are cheaper suites from HubSpot and Freshworks, Microsoft's Dynamics bundled with Office, or building their own. Its newest product, Agentforce, adds AI agents to the same software, sold in pricier editions.
How it makes money
Almost all of the money is subscriptions: $10.8B of the $11.3B Salesforce took in Q2 FY2027. Customers sign contracts, usually for a year or more, and pay ahead, so $33.5B of revenue is already contracted for the next twelve months.
Two-thirds comes from the sales, service, marketing and Slack apps, which grew 8%. The rest is data and connection tools, which grew 20%, mostly because of Informatica.
Growth and profit
What growth costs
Salesforce became far more profitable when it cut selling costs while sales kept growing. Since then growth has stopped doing that work: each cost takes about the same share of every sales dollar, so profit grows only as fast as sales. Selling is still the largest cost, and share pay is expected to rise.
Its record
The numbers
Is it growing?Steady
Sales grew 11% over the twelve months to Jul 2026, about the same pace as the 9.6% in the year to Jan 2026.
Is it profitable?Healthy
Operating margin is 20%, up from 14% in the year to Jan 2024.
Does growth pay off?Flat
Running costs take 57¢ of each sales dollar, down from 71¢ in the year to Jan 2022.
Does it earn good returns?Low
Each dollar put into the business earns 9.2% a year after tax, up from 0.9% in the year to Jan 2022.
Is it making cash?Strong
It turned 34% of sales into free cash, $15.2B over the last twelve months.
Is the balance sheet safe?Light debt
It owes $27.9B more than the cash it holds, about 1.8 years of the cash its operations bring in.
Are shares being diluted?Shrinking
The share count fell 16% in 4.5 years, and pay in shares costs 8.3% of sales.
Why it wins
The evidence
A large customer builds years of records, custom fields, automations and links to other systems inside Salesforce, and trains its staff on it. Leaving means rebuilding all of that, so most customers renew: Salesforce loses about 8% of its contract value a year, the same as a year earlier.
An outside research firm, IDC, still ranks it first with 20.0% of the market in 2025, though that slipped from 20.7%. The protection is weakening at the edges: AI coding makes the rebuild cheaper.
Who could beat it
The rivals in full
The new threat is the customer itself. Curative, a health start-up, cancelled a $600,000-a-year Salesforce contract after building its own replacement in two months with AI coding tools. Salesforce's own annual report now lists AI-native companies that may bypass its workflows as competitors.
The old rivals remain: Microsoft, HubSpot, Freshworks and ServiceNow. Anthropic is both partner and possible rival for the agents.
Each rival
Customers' own developers using AI coding agentsin house
renting a CRM, by building a company's own customer database and workflow app with AI coding tools
Sales and Service Cloud seats, bundled with the Microsoft licences and Copilot agents a company already pays for
HubSpot and FreshworksHUBSFRSH
Sales, Service and Marketing Cloud for smaller and mid-sized buyers, with simpler, cheaper suites
ServiceNow customer service workflows and AI agentsNOW
Service Cloud and Agentforce service agents where a company already runs its internal workflows on ServiceNow
Anthropic's and OpenAI's own agents working on company dataprivate
Agentforce, by having a general model act on customer records directly; Anthropic is also Salesforce's partner in Claudeforce and an investment
How big it could get
The market and who has signed
The market
Salesforce already holds about a fifth of a market it has led for 13 years, so its main business grows with what companies spend on this software rather than by taking much more of it.
The new part, AI agents, is where it could grow faster, but nobody outside counts that market. The work they would do is large: 2,666,000 customer service jobs in the US in 2025, which the government expects to shrink 5% by 2035.
Who has signed
Customers had $66.3B of contracts not yet billed at the end of July 2026, half of it due within a year. The largest named agent deal is the US Department of Veterans Affairs: one firm year with two optional years, worth up to $1.6B in all.
Who runs it
Leaders, pay and the cash
Who runs it
Marc Benioff co-founded Salesforce in 1999 and has run it since 2001. Robin Washington, on the board since 2013 and once Gilead's finance chief, took over finance in March 2025. Insiders own about 3.5%, mostly Benioff. The bonus rides on subscription sales and adjusted profit; Benioff's shares pay only if the stock beats the Nasdaq-100 and margins plus growth hit targets, or if AI sales do. None of it rides on cash left after share pay.
What they did with the money
Management borrowed $25B in March 2026 and spent it buying back shares, cutting the share count about 15% in a year. The cost is interest: $473M in Q2 against $67M a year earlier. It also keeps buying companies for cash, including Fin for $3.6B and Contentful for about $1.5B.
The last report
The latest report
Salesforce's Q2 FY2027 revenue rose 11% to $11.3B, helped by Informatica
What it did to the thesis Mixed quarter
It answered the fear that customers were leaving or cutting users, for one quarter. It did not show faster growth without acquisitions, and the earnings beat came from an investment gain, so the case for a better business than the price assumes stayed unproven.
The report
In Q2 FY2027, the three months to July 2026, revenue rose 11% to $11.3B. Without Informatica it grew about 6.4%. The main apps grew 8% after currency effects, no faster than a year earlier.
Management said the number of users grew, few customers left, and Agentforce passed $1.5B of yearly recurring revenue, though on a wider definition than before. Earnings per share jumped, but most of the jump was a gain on investments, chiefly Anthropic, not the business.
What's happening now
The background
The shares fell from $269.11 in December 2025 to $146.32 in June 2026, on fear that AI would shrink the number of users companies pay for. Then Q2 showed users growing, and the shares rose 22.6% in one day.
So the fear has eased but not gone: analysts have grown a little less positive since June. The next answer comes on 2 December 2026, when Salesforce reports Q3 FY2027 and should show growth without acquisitions near 7%.
Where we disagree
Both sides in full
The market
Salesforce is cheap: about 14 times this year's adjusted earnings, with AI fears overdone. Most analysts rate it a buy.
We believe
It is a good business at a fair price, not a cheap one, once share pay, investment gains and new interest are counted.
The argument
The cheap reading uses adjusted earnings that include gains on the Anthropic stake, and cash flow before share pay and before a full year of interest on the new debt.
The case against us: if growth without acquisitions climbs toward 9%, as management promises, today's price would prove cheap after all.
What it's worth
How we value it
Valued on the analysts' revenue path for three years, each later year closing 30% of the gap to 4%, on official operating profit with stock pay a cost and tax paid in full; the thesis's own evidence is the check on that path.
What the price assumes
Sales growth in the first year, then slowing as ours does, at the same profit per sale.
The price assumed
12.5%
Our estimate
10%
Measured at the close of
6 Nov 2026
The three cases
Bear case$104
Growth runs at half the analysts' path, 5%, 5% and 5%, the margin settles a fifth lower at 21.1%, and new capital earns only the 8% rate: the downturn test.
Our value$218
Revenue grows 10%, 10% and 10% over three years, the analysts' path (C54), then fades toward 4%, closing 30% of the gap each year; the margin moves to 26.4%, the analysts' year-three figure after stock pay and on the filed basis (8.9% less).
Bull case$293
Growth runs a quarter above the analysts' path, 13%, 12% and 12%, the margin reaches 29.0%, and new capital earns 15%.
The bet
Seen as a bet
The bet is on Agentforce alone, the AI agents. The rest of Salesforce is valued separately, at what its rivals pay for each dollar of spare cash, turned into a price per dollar of sales.
That existing business, growing steadily, is worth about today's price on its own, so the agents come free. But even if they roughly triple in two and a half years, the whole company is worth only a little more than today: a defensive bet with little upside.
What could go wrong
Every risk
The risks
Customers build their own
AI coding let Curative replace a $600,000-a-year contract in two months; if larger companies follow, renewals fall.
Agents cut paid users
Salesforce charges per user, so if AI agents shrink sales and service teams, revenue falls unless pricier editions make up for it.
Growth is partly bought
Informatica, Fin and Contentful add revenue; without them growth is about 6%, and buying companies spends shareholders' money.
Debt now costs real money
Interest rose to $473M a quarter after the $25B borrowing, which slows growth in the cash left for shareholders.
AI figures are management's
Agentforce's $1.5B is a spoken figure on a widened definition, not a filed one.
What we still do not know
How much of Agentforce's $1.5B is new spending on agents rather than older products now counted in it.
Whether the total number of paid users is rising or falling; management says it grew but gives no number.
How large IDC's customer software market is in dollars, and how much of Salesforce's revenue it counts.
What changes our view
Every route to a new view
Every route to a new view
→ Attractive
The share price falls to about $174, 80% of what we think the business is worth.
→ Unattractive
The share price rises to about $261, 20% above what we think the business is worth.
→ Unattractive
The main apps grow less than 6% a year after currency effects. Revenue growth without acquisitions falls below 5%.
The next test
It supports the case if
Revenue at or above $11.46B with Informatica at most 4.5 points of growth, and Agentforce Apps at 8% or more.
It pressures the case if
Growth without acquisitions below Q2's 6.4%, or Agentforce Apps below 8% in constant currency.
We would also be wrong if
Salesforce loses more than about 9% of its contract value in a year.
The cash left after share pay falls in fiscal 2028.
The verdict
Scores and how we weighed it
The scores
Growth
3 of 4
Business quality
3 of 4
Financial position
3 of 4
Execution record
3 of 4
Per share capture
3 of 4
Potential
2 of 4
Risk profile
3 of 4
Valuation
3 of 4
Evidence quality
3 of 4
How we weighed it
ValueMixed
The price is about at our value, with the bear case well below.
BusinessPositive
Software big companies run sales and service on, and customers rarely leave.
GrowthMixed
Growth without purchases is slow, and the deals carry it.
RisksNegative
Customers can build their own, agents may cut paid users, growth is partly bought and debt costs real money.
RivalsMixed
Dearer than rivals on sales and cheaper on spare cash; the whole group has fallen.
The groupMixed
Software names move together on fears that AI shrinks them.
Latest newsMixed
One quarter eased the fear of customers leaving, but growth did not speed up and the beat came from an investment gain.
In plain words
Our estimate of value
We forecast how much cash the business can earn over the next ten years and beyond, then ask what that stream is worth today to an investor who wants an 8% yearly return. We do it three times: a gloomy future, an expected one and a sunny one. We call a share Attractive only when its price is at least 20% below the expected estimate, and Unattractive when it is 20% above.