Twilio sends the texts and calls behind other apps
Texts
log-in codes, alerts and chats
Calls
billed by the minute
AI agents
answering customers on the phone
Developers add a few lines of code and their app can text, call or email customers through Twilio, which rents phone networks around the world.
The company · 2 · What exactly does it sell?
Three quarters of sales are fees per use
Each $1 of sales, third quarter of 2026
76%24%
Fees per text and per minuteSubscriptions for email and customer data
Most customers pay as they go and can cut their use at any time, but none is large.
The company · 3 · How does it make money?
Paid per text, per minute and per month
76%of sales from fees per use
6.8%profit per $1 after share pay
About 10%of sales paid to staff in shares
Sales are large and growing, but after paying staff in shares little profit is left.
Is it a good business? · 4 · What gives it an edge?
Code written to Twilio is costly to move
$1.14
spent a year later per $1 before
9%
of sales from the ten biggest customers
Once a company's apps call Twilio, switching means rewriting and retesting code and moving phone numbers, so customers stay and spend more.
Is it a good business? · 5 · What is the catch?
Rivals offer the same service, some cheaper
Infobipranked first by one research firm
Bandwidthowns its own US network
Amazonsells parts for AI call centres
Listed rivals: SINCH.ST · ERIC · BAND · AMZN · MSFT
Twilio rents most of the networks it sells, so it competes on software and reach rather than owning the wires.
Is it a good business? · 6 · What does its record show?
Sales rose by two fifths and losses turned to profit
Sales, $B
2.8
3.8
4.2
4.5
5.1
5.8
Dec 2021Dec 2022Dec 2023Dec 2024Dec 2025last 12 mo
Sales$4,154M → $5,832Msales, 2023 to the past year
Profit−$876.5M → $397.6Moperating profit, 2023 to the past year
Staff5,587staff in 2025, up 0.9%
Profit came from cutting selling and office costs while sales grew and the staff count stayed flat.
Is it a good business? · 7 · How much room is there to grow?
Americans send trillions of texts a year
2.2 trillionUS text messages in 2024
2.4 trillionUS minutes on phone calls
About 2.7 millionUS customer-service jobs
Nobody counts how much of that traffic businesses send, so Twilio's share of it cannot be measured.
Is it a good business? · 8 · Who is in charge?
A long-time insider runs it, with an activist watching
Since 2024Khozema Shipchandler, chief executive
Under 1%of shares owned by insiders
3.2 millionshares sold by the activist's fund this year
Bonuses pay on growth and on profit before share pay; part of long-term pay rides on cash generated.
The story right now · 9 · Why does AI matter to Twilio?
AI agents need a phone line to answer customers
About 1%of Twilio's voice calls handled by AI so far
$0.07a minute to connect an AI model to a call
About 2.7 millionUS customer-service jobs
Businesses are starting to let AI answer customer calls. Each AI agent still needs a phone connection, which Twilio sells by the minute.
The story right now · 10 · Why does Twilio's profit look so different?
Paying staff in shares eats half the profit
Sales$1.56B in the third quarter.
Profit as Twilio shows it21% of sales.
Less pay in shares9.9% of sales.
Profit under accounting rules9.5% of sales.
Shares given to staff are a real cost, because each new share shrinks every owner's slice. We count them; the profit Twilio leads with does not.
The story right now · 11 · What changed lately?
Sales grew 20%, but next quarter is slower
+20%sales
+16%without carrier fees and purchases
$1.14spent a year later per $1 before
11% to 12%growth expected next quarter
Solid quarter
The shares fell 15.2% the day after the report because the forecast slowed.
The story right now · 12 · What is everyone arguing about?
Can AI agents justify the price?
Many investors think
An AI winner whose growth will speed up
We think
A good business priced for AI sales nobody has reported
We are wrong if Twilio reports AI-agent sales near $500M a year while growth stays above 15%.
The price and our view · 13 · What is a share worth?
The price sits far above even our sunny case
$63.63what we think a share is worth · the last close is of that
$63.63our value
Attractive below $50.90
Neutral
Unattractive above $76.36
$41.90gloomy: growth halves and AI adds little
$63.63our estimate: analysts' growth, share pay a cost
$77.37sunny: faster growth as AI agents arrive
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 · What if it all works?
As a bet on AI agents, it is priced for success
Priced for successthe kind of bet this is
0.7×if it works
56%kept if it fails
Multiples of the company's value when the bet was last struck, 13 Nov 2026.
Even if AI-agent sales grow several times over, the shares would be worth less than today's price, because the price already assumes they arrive.
The price and our view · 15 · How could we be wrong?
Customers can walk, and rivals are cheaper
Customers can walkBiggest
Most pay as they go.
Same service, cheaperShowing now
Rivals list lower prices.
Rented networksWatch
Two carriers hit trouble.
What AI agents earnUnknown
No sales reported.
Fees passed on from phone carriers add to sales but not to profit, so growth looks faster than it is.
The price and our view · 16 · How will we know?
February tests whether growth holds
Growth of 11% or more without carrier feesGood sign
A 2027 forecast below 10% growthBad sign
AI-agent sales near $500M a yearWe turn neutral
FEB 202711
Fourth-quarter 2026 results, 11 February 2027. Companies report every three months; each report is a test of the story.
Our view also changes on price alone: we turn neutral if the shares fall to within a fifth of our estimate.
The price and our view · 17 · So, what do we think?
A good business at a price that assumes too much
$63.63our value
Attractive below $50.90
Neutral
Unattractive above $76.36
UnattractiveHigh conviction
Twilio grows fast, customers spend more each year and it now makes a profit. But after paying staff in shares little is left, and the price sits far above our value. Unattractive.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
Twilio sells building blocks for talking to customers. A bank's app that texts you a log-in code, a delivery app that calls the driver, a shop that emails a receipt: often the message goes through Twilio, which has contracts with phone networks around the world and wraps them in simple code.
Its newest product connects a business's AI model to a live phone call, so an AI agent can answer customers. An outside research firm ranks Twilio among the top three sellers of this kind of service.
How it makes money
About three quarters of the money is fees per text sent and per minute of calls; the rest is subscriptions for email and customer data. Most customers can cut their use any time, but none is large: the ten biggest are 9% of sales. Part of the fee is the phone carriers' own charge, passed on with no profit on it.
Growth and profit
What growth costs
Growth has got cheaper: from 2023 to 2025 sales rose 22% while spending on selling and offices fell and the staff count stayed flat, turning a loss into a profit. Research spending still rises, and share pay still takes about 10 cents a dollar.
Its record
The numbers
Is it growing?Steady
Sales grew 19% over the twelve months to Sep 2026, faster than the 14% in the year to Dec 2025.
Is it profitable?Thin
Operating margin is 6.8%, up from −21% in the year to Dec 2023.
Does growth pay off?Pays off
Running costs take 42¢ of each sales dollar, down from 81¢ in the year to Dec 2021.
Does it earn good returns?Low
Each dollar put into the business earns 4.2% a year after tax, about where it was (1.1%) in the year to Dec 2025.
Is it making cash?Strong
The business brought in $1.19B of cash from operations, 20% of sales, before capital spending, which the filer does not tag.
Is the balance sheet safe?Net cash
It holds $2.89B of cash against $994M of debt, $1.90B more cash than it owes.
Are shares being diluted?Shrinking
The share count fell 7.1% in 4.8 years, and pay in shares costs 10% of sales.
Why it wins
The evidence
Once a company's apps are written to call Twilio, moving means rewriting and retesting that code and moving its phone numbers, so customers tend to stay and grow. Existing customers spent $1.14 a year later for each dollar, about $1.10 without the carriers' fees.
For AI agents the price is small next to what they save: a minute of AI voice through Twilio lists at about 8 cents, against about 36 cents a minute for a customer-service worker's median wage.
Who could beat it
The rivals in full
Several rivals sell the same thing: Infobip, Sinch and Vonage rank beside Twilio, and Bandwidth owns its own US phone network. Amazon and Microsoft sell the parts for businesses building their own AI call centres. Twilio rents most of its networks, so it competes on software and reach, not on owning the wires.
Each rival
Infobipprivate
Twilio's messaging and voice interfaces, sold to the same developers and enterprises worldwide
SinchSINCH.ST
messaging and voice interfaces, the same CPaaS product, strongest in carrier messaging
Vonage, owned by EricssonERIC
voice, video and messaging interfaces sold to developers
BandwidthBAND
US calling and texting interfaces, sold on owning its own phone network
Amazon Connect and Microsoft's contact-centre toolsAMZNMSFT
Conversation Relay and Twilio's voice for a business building its own AI contact centre
How big it could get
The market and who has signed
The market
Americans sent nearly 2.2 trillion texts and spent over 2.4 trillion minutes on calls in 2024, and about 2.7 million work as customer-service agents. Nobody counts how much of that traffic businesses send, so the size of Twilio's market cannot be pinned down; one research firm puts the field at about $28B a year worldwide.
Who runs it
Leaders, pay and the cash
Who runs it
Khozema Shipchandler, the former finance chief, has run Twilio since January 2024; Aidan Viggiano is finance chief. Insiders own under 1%. Bonuses pay on growth and on profit before share pay. An activist fund, Sachem Head, has a director on the board and sold 3.2 million shares this year.
What they did with the money
It bought back about $3.3B of shares since 2024, cutting the count about 12%, though lately buybacks only match the new shares it pays staff. It bought Stytch, a small company that checks the identity of AI agents, for about $104M.
The last report
The latest report
Twilio's sales rose 20% but it expects slower growth next quarter
What it did to the thesis Solid quarter
It showed the business improving, with customers spending more and profit rising, but the slower forecast removes the speed the price was paying for.
The report
In the third quarter of 2026 sales rose 20% to $1.56B, or 16% without carrier fees and purchases. Existing customers spent $1.14 for each dollar a year before, up from $1.09. Profit under accounting rules was $148.6M, about 10% of sales.
The forecast for the fourth quarter is 11% to 12% growth without fees, slower than the third, and the shares fell 15.2% the next day.
What's happening now
The background
The shares rose 64.5% in a year, mostly because buyers paid more for each dollar of sales. They joined the S&P 500 in October 2026, then fell 15.2% after the third-quarter report forecast slower growth.
The next check is the fourth-quarter report in February 2027, with a forecast for 2027 and an Investor Day in April where AI-agent sales may be shown for the first time.
Where we disagree
Both sides in full
The market
An AI winner whose turnaround is done and whose growth will speed up as AI agents need phone lines.
We believe
A real business getting better, but the price pays for AI-agent sales nobody has reported, on profit share pay mostly uses up.
The argument
Most analysts rate it a buy, more than in March. They read the profit Twilio reports before share pay; we count share pay as a cost.
Where we could be wrong: if Twilio reports AI-agent sales near $500M a year while growth stays above 15%, the business would be worth far more than we think.
What it's worth
How we value it
Valued on the analysts' revenue path for three years, then each year closing 30% of the gap to 4%, on GAAP operating profit with stock pay a cost and tax at a normal 25%, since past losses leave no filed rate. The bridge adds cash, short-term securities and the Syniverse stake at carrying value and subtracts the notes.
The three cases
Bear case$41.90
Growth runs at half the analysts' path, 6.5%, 5.4% and 3.0%, the margin after stock pay settles at 6.9%, a fifth under the base, and new capital earns only the 8% rate: the test if AI agents add little.
Our value$63.63
Revenue grows 12.9%, 10.8% and 6.0%, the analysts' path (C44), then fades toward 4%; the margin moves to 8.6%, the analysts' 18.6% for 2028 less stock pay of 10.05% (C09); new capital earns 12% for an average moat.
Bull case$77.37
Growth runs a quarter above the analysts' path, 16.1%, 13.5% and 7.5%, the margin reaches 9.4%, and new capital earns 15%, as if AI agents add a second leg of growth.
The bet
Seen as a bet
The bet is on Twilio's AI-agent products: the connection that puts an AI model on a live phone call, and checking the identity of AI agents. Today they are a sliver of its voice calls and report no sales.
Even if they grow several times over, the shares would still be worth less than today. Most of today's price already assumes the AI products succeed, so this bet is priced for success.
What could go wrong
Every risk
The risks
Customers can walk
About three quarters of sales are fees most customers can cut any time, with no long contract.
Same product, cheaper
Rivals sell the same service, some at lower list prices, and one owns its network.
Rented networks
Twilio relies on phone carriers it does not own; two of them ran into trouble this year.
Fees that flatter growth
Carrier charges passed through add sales but no profit.
Insiders selling
No insider bought shares this year; the activist on the board sold heavily.
What we still do not know
How much AI agents earn Twilio. It reports no sales by product.
How much customers grow their spending without the carriers' fees, quarter by quarter.
Why Twilio stopped reporting its customer count in 2026.
What changes our view
Every route to a new view
Every route to a new view
→ Neutral
The share price falls to about $76.36, 20% above what we think the business is worth.
→ Attractive
The share price falls to about $50.90, 80% of what we think the business is worth.
→ Neutral
Twilio reports AI-agent sales of $500M a year or more. Sales grow 15% or more for a full year without fees or purchases.
The next test
It supports the case if
Organic growth of 11% or more in Q4 and a 2027 guide of at least 10% organic, with net expansion of 110% or more.
It pressures the case if
Organic growth below 11% in Q4, a 2027 guide under 10% organic, or net expansion below 110%.
We would also be wrong if
Sales growth without fees or purchases falls below 8% for two quarters running.
Existing customers fall below $1.05 a year later for each dollar, without carrier fees.
The verdict
Scores and how we weighed it
The scores
Growth
3 of 4
Business quality
3 of 4
Financial position
4 of 4
Execution record
3 of 4
Per share capture
3 of 4
Potential
3 of 4
Risk profile
3 of 4
Valuation
1 of 4
Evidence quality
3 of 4
How we weighed it
ValueNegative
The price is several times our base value and above our sunny case.
BusinessPositive
Sticky with developers, customers spending more each year, net cash and growing free cash flow.
GrowthMixed
Growth sped up this year, but the guide slows it and carrier fees flatter it.
RisksMixed
Customers can cut usage at any time and it rents the networks it sells, but no customer is large.
RivalsMixed
Ranked near the top, but rivals offer the same product, some cheaper, and the cloud giants offer the parts.
The groupMixed
Software that AI might help or replace trades on mood, and joining the index lifted it.
Latest newsNegative
The last report grew fast but guided slower, and the shares fell on it.
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.