IBM runs the core computer systems of banks and governments
Software
almost half of sales, most of the profit
Consulting
people who build and run big projects
Mainframes
giant computers for card payments and bookings
When you pay by card or book a flight, the transaction often passes through an IBM mainframe. IBM sells those machines, the software on them and the people who connect it all.
The company · 2 · What exactly does it sell?
Software is the biggest of four businesses
Each $1 of sales in 2025, by business
44%31%23%1%
SoftwareConsultingMainframes and serversLoans to clients
Software includes Red Hat, which lets programs run on any cloud, and the programs that run on mainframes. Consulting and hardware make up most of the rest.
The company · 3 · How does it make money?
Most software is paid for again every year
About 80%of software sales repeat each year
$3 to $5of software sold for each $1 of mainframe
$69Bof signed work not yet delivered
Clients pay yearly for software licences and support, by the project for consultants, and in waves for each new mainframe model.
Is it a good business? · 4 · What gives it an edge?
Moving off the mainframe is slow and risky
About 85%
of mainframe capacity is kept or growing, says IBM
Three of ten
critical US government systems replaced in six years
Banks and governments have written decades of programs for IBM machines. Rewriting them could break the systems that move their money, so they keep paying.
Is it a good business? · 5 · What is the catch?
Consulting is stuck and rivals target the mainframe
AI may let clients do work they once paid consultants for, and cheaper Indian firms press on prices.
Is it a good business? · 6 · What does its record show?
Sales grew slowly, but profit per dollar rose
Sales, $B
57.4
60.5
61.9
62.8
67.5
69.9
Dec 2021Dec 2022Dec 2023Dec 2024Dec 2025last 12 mo
Sales$57,350M → $67,535Mrevenue, 2021 to 2025
Profit8.4% → 15.3%profit before tax per $1, 2021 to 2025
Cash$14.7Bcash left after investment in 2025
As software grew to almost half of sales, IBM kept more of each dollar. It pays a dividend every quarter and has not bought back shares since 2019.
Is it a good business? · 7 · How much room is there to grow?
Software and IT services are a huge, growing market
$3.3 trillionspent on software and IT services in 2026, says Gartner
15.1%growth in software spending
About 2%IBM's share of that spending
The market is vast, but nobody measures the slice of big-company spending IBM can actually win.
Is it a good business? · 8 · Who is in charge?
A lifer runs IBM and is paid on growth and cash
1990Arvind Krishna joined IBM
2020became chief executive
$11.3Bspent buying Confluent in 2026
He bought Red Hat, spun off the old outsourcing arm and keeps buying software firms. Bonuses pay on revenue, earnings per share and cash generated.
The story right now · 9 · How much does AI matter here?
AI work lifts software but has not grown consulting
About 30%of consulting sales are AI work, says IBM
Flatconsulting sales last quarter
Companies want AI built on top of systems they cannot easily replace, which helps IBM's software. In consulting, AI work so far replaces older work rather than adding to it.
The story right now · 10 · Why do mainframe sales swing so much?
Mainframe sales rise and fall with each new model
A new model launchesz17 arrived in June 2025.
Clients upgrade in a rushIBM Z sales rose by half in 2025.
Sales fall as the model agesDown 20% last quarter.
Software keeps comingClients pay for software on the capacity they use.
So a falling quarter for mainframes is normal between models. What matters is whether clients keep or grow their capacity, which pulls software sales behind it.
The story right now · 11 · What changed lately?
Software grew, mainframes fell, consulting stood still
$17.1Bquarterly revenue, up 5%
8%software growth
25.2%one-day share fall in July
Mixed quarter
In July about $400M of software deals slipped and IBM cut its growth outlook. Last quarter cash flow was strong, but consulting signings fell again.
The story right now · 12 · What is everyone arguing about?
Can software growth carry a slow-growing giant?
Many investors think
A steady software and AI winner after one bad quarter
We think
A durable franchise growing slowly, priced just above its worth
Against us: if software grows near 10% without purchases and consulting signings turn up, the business would be worth more than the price.
The price and our view · 13 · What is a share worth?
The price sits at our line, below our sunny case
$204what we think a share is worth · the last close is of that
$204our value
Attractive below $163
Neutral
Unattractive above $225
$104gloomy: growth half what analysts expect
$204our estimate: growth as analysts forecast, then slowing
$264sunny: growth a quarter faster than forecast
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, with debt taken off. Our rule: buy only at least a fifth below our estimate.
The price and our view · 14 · How could we be wrong?
Cheaper exits, stuck consulting and heavy debt
Leaving the mainframeBiggest
AI tools make moving off cheaper.
Consulting signingsShowing now
Down again last quarter.
DebtWatch
$63.8B owed, ahead of owners.
Quantum salesUnknown
Not reported separately.
Our estimate also assumes IBM keeps paying a low tax rate; a normal rate would cut it.
The price and our view · 15 · How will we know?
Late January brings the 2027 outlook
Growth of 5% or more in the 2027 outlookGood sign
Consulting signings growing againGood sign
Software growth slipping below 7%Bad sign
JAN 202727
Fourth-quarter 2026 results and 2027 outlook, around 27 January. Companies report every three months; each report is a test of the story.
Our view also changes on price alone: if the shares fall to about a tenth above our estimate, we move to Neutral.
The price and our view · 16 · So, what do we think?
A durable business at a price still above its worth
$204our value
Attractive below $163
Neutral
Unattractive above $225
UnattractiveMedium conviction
Clients rarely leave and cash is strong. But growth is slow, partly bought with debt, and the price is about 119% of our value, so it is not cheap enough to buy.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
When you pay by card, book a flight or file a tax return, the transaction often passes through an IBM mainframe, a large computer built to process huge numbers of small transactions without stopping. IBM sells those machines, the software that runs on them, and Red Hat, which lets companies run programs across any cloud.
It also sells consultants who design and run projects. Its rivals are Microsoft, Oracle and Amazon in software, Accenture and Indian firms in consulting, and Dell and HPE in hardware.
How it makes money
Software brings in about 44 cents of each dollar, mostly paid yearly for licences and support. Consulting brings about 31 cents, billed by project. Mainframes, servers and storage bring about 23 cents, in waves as each new model ships, and every mainframe sold pulls more software sales behind it.
Growth and profit
What growth costs
Growth has made IBM more profitable. As software became almost half of sales, the share of each dollar kept after making the product rose, and selling and admin costs fell, so official profit before tax nearly doubled as a share of revenue between 2021 and 2025.
Its record
The numbers
Is it growing?Slow
Sales grew 6.8% over the twelve months to Sep 2026, about the same pace as the 7.6% in the year to Dec 2025.
Is it profitable?Healthy
It keeps 16% of sales as profit, up from 12% in the year to Dec 2023.
What does it keep on each sale?Solid
It keeps 58¢ of each sales dollar after making what it sells, and spends 29¢ of it on selling and admin, its biggest running cost.
Does it earn good returns?Strong
Shareholders' money earned 35% a year, up from 29% in the year to Dec 2021.
Is it making cash?Strong
It turned 21% of sales into free cash, $14.7B over the last twelve months.
Is the balance sheet safe?Heavy debt
It owes $61.6B more than the cash it holds, about 3.9 years of the cash its operations bring in.
Are shares being diluted?Creeping
The share count rose 5.4% in 4.8 years, and pay in shares costs 2.8% of sales.
Why it wins
The evidence
Banks and governments have run their core programs on IBM mainframes for decades, and rewriting them is slow and risky. The US government's own auditor found agencies had replaced only three of ten critical old systems in six years. So clients keep paying: about 80% of software revenue repeats every year.
The threat is Amazon, which now sells an AI tool it says moves mainframe programs to the cloud in months.
Who could beat it
The rivals in full
Cloud providers want the mainframe's work, Oracle and Microsoft sell rival software, and Accenture and cheaper Indian firms compete for consulting. AI tools may also let clients do work they once paid consultants for. In quantum, five other teams are in the same final stage of a US defence test.
Each rival
Amazon Web Services, Microsoft Azure and Google CloudAMZNMSFTGOOG
a client moves programs off the mainframe and IBM software to a public cloud, now with AI rewriting tools
Oracle, SAP, Microsoft, Broadcom (VMware) and SalesforceORCLSAPMSFTAVGOCRM
a client buys database, automation or platform software from a rival
BMC Softwareprivate
a client buys mainframe management tools from BMC
AccentureACN
a client hires Accenture for the transformation project
Capgemini and India-based providers such as Infosys and WiproCAP.PAINFYWIT
a client hires a cheaper offshore firm
Clients' own AI tools doing work consultants didin house
a client uses AI coding and agents in-house instead of buying hours
How big it could get
The market and who has signed
The market
Gartner, a research firm, expects the world to spend about $1.4 trillion on software and $1.9 trillion on IT services in 2026, growing 9% to 15%. IBM's revenue of about $70B is roughly 2% of that. Nobody outside measures the slice of large-company spending IBM can actually win.
Who runs it
Leaders, pay and the cash
Who runs it
Arvind Krishna, at IBM since 1990, has run it since 2020; he led the Red Hat purchase, spun off the old outsourcing business as Kyndryl and bought Confluent. Directors and officers own a tiny fraction of the company.
Bonuses and long-term shares pay on revenue, operating earnings per share and free cash flow, so pay rewards growth and cash. IBM pays a dividend every quarter and has not bought back shares since 2019.
The last report
The latest report
IBM's third-quarter revenue rose 5%, with software up and mainframes down
What it did to the thesis Mixed quarter
It supported the business: software and cash flow held up and the cash outlook rose. It did not answer the doubts on growth: Consulting stayed flat, signings fell and mainframe sales dropped. Our view holds.
The report
In the quarter to September 2026 revenue rose 5% to $17.1B. Software grew 8%, led by Red Hat and data products, while Consulting was flat and mainframe sales fell 20%. Cash flow was strong and IBM raised its free cash flow outlook for the year.
Consulting signings fell again, and the year's growth outlook stayed at the lower range set in July.
What's happening now
The background
The shares fell a quarter in one day in July when about $400M of software deals slipped and IBM cut its growth outlook. They have recovered part of the fall, and over the year the price fell more than sales per share rose.
The next test is the fourth-quarter report, expected around 27 January 2027, with the 2027 outlook.
Where we disagree
Both sides in full
The market
A steady software and AI winner hit by one slipped quarter.
We believe
A durable franchise growing slowly with heavy debt, priced just above its worth.
The argument
About half of analysts rate IBM a buy, seeing an AI software winner after one bad quarter. We think the price needs growth about twice what analysts expect for a decade.
The evidence against us: if Software grows near 10% without purchases and Consulting signings turn up, the value would rise above the price.
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 pre-tax profit with stock pay a cost; the bridge takes the filed debt of $63,817M, adds $1,825M of marketable securities and subtracts $669M of minority interest.
What the price assumes
Sales growth in the first year, then slowing as ours does, at the same profit per sale.
The price assumed
10.4%
Our estimate
4%
Measured at the close of
16 Nov 2026
The three cases
Bear case$104
Growth runs at half the analysts' path, 1.9%, 2.6% and 2.5%, as Consulting shrinks and the mainframe turns down; the margin settles at 17.7% and new capital earns only the 8% rate.
Our value$204
Revenue grows 3.9%, 5.3% and 4.9% over three years, the analysts' path (C30), then fades toward 4%, closing 30% of the gap each year; the margin moves to 22.1%, the analysts' 2028 figure less stock pay of 2.8% of sales.
Bull case$264
Growth runs a quarter above the analysts' path, 4.9%, 6.6% and 6.1%, as Software reaches 10% growth, the margin reaches 24.3%, and new capital earns 15%.
The bet
Seen as a bet
This bet is on IBM's quantum computers. IBM plans to spend more than $10B on them over five years and was picked by the US defence research agency for the final stage of its test, but it reports no quantum revenue.
The rest of IBM, valued alone, is worth about today's price, so almost nothing in the price rests on quantum. If it reaches the big win, the shares gain little; if it fails, little is lost, which makes it a safe and expensive bet.
What could go wrong
Every risk
The risks
Leaving the mainframe gets cheaper
AI tools from Amazon and others pitch moving old programs in months, not years.
Consulting keeps shrinking
Signings fell in 2025 and again last quarter, and AI may replace paid hours.
Heavy debt
Debt of $63.8B and a pension shortfall come ahead of shareholders.
A low tax rate
Our value assumes IBM keeps paying tax in the mid-teens.
The mainframe cycle
Hardware sales fall between models, as they did by a fifth last quarter.
What we still do not know
How much revenue quantum computing brings; IBM gives contracts signed since 2017 but no revenue.
How much of software's growth came from buying Confluent.
How fast clients are moving work off the mainframe.
What changes our view
Every route to a new view
Every route to a new view
→ Neutral
The share price falls to about $225, a tenth above what we think the business is worth.
→ Attractive
The shares fall to 80% of what we think the business is worth.
→ Neutral
Software grows 10% or more in a quarter without help from purchases. Consulting signings grow from a year earlier.
The next test
It supports the case if
2027 growth guided at 5% or more, Software growing near 10% at constant currency, Consulting signings up and free cash flow above $15.7B for 2026.
It pressures the case if
2027 growth guided below 4%, Software below 7%, or IBM Z still falling sharply.
We would also be wrong if
Mainframe sales keep falling more than 10% after the next model launches.
Software growth falls below 5% for two quarters.
The verdict
Scores and how we weighed it
The scores
Growth
2 of 4
Business quality
3 of 4
Financial position
2 of 4
Execution record
3 of 4
Per share capture
2 of 4
Potential
2 of 4
Risk profile
3 of 4
Valuation
2 of 4
Evidence quality
3 of 4
How we weighed it
ValueNegative
The price has slipped a hair inside the line above our value, on a value that already allows IBM's low tax, though below our bull case.
BusinessPositive
Clients rarely leave the mainframe, most Software is recurring and cash flow is large and steady.
GrowthNegative
Growth is modest and partly bought; the price needs about twice what the analysts expect.
RisksNegative
Debt is high, the pension is short, and our value already assumes a low tax that may not last.
RivalsMixed
Dearer than Accenture and Oracle on profit, while Amazon now attacks the cost of leaving the mainframe.
The groupMixed
It moves with enterprise software, where AI spending lifts some names and threatens others.
Latest newsMixed
The last report beat on cash and Red Hat, but Consulting stayed flat and the mainframe fell.
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.