When you open Reddit, the New York Times or Spotify, much of what loads comes from a nearby Fastly server rather than the company's own computers far away.
The company · 2 · What exactly does it sell?
Delivery is three-quarters; security is growing
Each $1 of sales, second quarter 2026
73%23%4%
Delivering web trafficBlocking attacksRunning customers' code
Delivery is billed by the traffic carried. Security is mostly yearly subscriptions paid ahead, which makes it steadier.
The company · 3 · How does it make money?
Paid by the traffic it carries, under yearly contracts
95%of sales from its biggest customers
$333.5Msigned but not yet delivered
39%of sales from just ten customers
Contracts run about a year, with a monthly minimum. Without Fastly, customers would serve from their own cloud or use Akamai or Cloudflare.
Is it a good business? · 4 · What gives it an edge?
Customers add products, but keep a second network
$1.19
spent now for each past dollar
2 or more
networks large customers run, moving traffic between them
Rules written in Fastly's tools make leaving work. But customers can shift traffic to their second network without rewriting anything, so delivery prices fall every year.
Amazon, Google and Microsoft also bundle their own networks with cloud accounts.
Is it a good business? · 6 · What does its record show?
Cheaper to grow, costlier in servers
Sales, $M
354.3
432.7
506.0
543.7
624.0
719.1
Dec 2021Dec 2022Dec 2023Dec 2024Dec 2025last 12 mo
Sales+20%sales in the latest quarter
Selling costs43 → 31 centsselling costs per $1 since 2021
Servers2 → 10 centsserver spending per $1 since 2024
Every running cost takes a smaller share of sales, and share pay per dollar halved. But server spending eats the cash gain.
Is it a good business? · 7 · How much room is there to grow?
A slow-growing delivery market
5% to 6%a year, delivery market growth, its finance chief says
3%of network-using websites on Fastly
$22Bthe market, management's own figure
No outside body counts the spending, so the market's size is the seller's number.
Is it a good business? · 8 · Who is in charge?
A new chief, and a founder selling steadily
Since 2025Kip Compton as chief executive
About 4%of the shares owned by insiders
4.67Mshares sold by insiders in a year
The finance chief sold outside a set plan near the high.
The story right now · 9 · What does AI do to web traffic?
AI agents bring new traffic, and new ways around it
About 6.5xhow fast AI traffic grows against human traffic
-18%the shares when Anthropic skipped edge networks
+187%the shares in a year, mostly a higher price per dollar of revenue
AI agents visit websites like people, which could bring Fastly traffic. But agents run inside big clouds can skip edge networks altogether.
The story right now · 10 · Where did the growth come from?
The headline 20% hides who drove it
Sales grew 20%The headline.
Ten customers drove it39% of sales.
Everyone else grew 8%Down from 12% in June.
Broad growth is the test18% would prove us wrong.
Growth that rests on a few big customers can vanish if one shifts traffic away. Broad growth is what makes it last.
The story right now · 11 · What changed lately?
Sales rose 20%, but only 8% outside the top ten
$190.2Mquarterly sales, above the forecast
+32%security sales
39%of sales from ten customers
About 16%growth the next forecast implies
Mixed quarter
In September management set 2029 sales targets of $1.1B to $1.3B.
The story right now · 12 · What is everyone arguing about?
Is growth broad, or a few big customers?
Many investors think
Growth is back, driven by security and AI
We think
Real but narrow growth, on cheapening delivery
Outside the ten largest customers sales grew about 8%. We are too harsh if that returns toward 18%.
The price and our view · 13 · What is a share worth?
No single value here: it is priced as a bet
1.2×if it works
100%kept if it fails
Multiples of the company's value when the bet was last struck, 5 Oct 2026.
We put no single value on a share here, because the business does not yet earn a lasting profit. Instead we ask what the shares pay if the plan works, and what is left if it fails.
The price and our view · 14 · How could we be wrong?
A few customers, falling prices and new shares
One customer groupBiggest
One customer is 12% of sales.
Prices keep fallingShowing now
Rivals sell at larger scale.
AI may route around itWatch
Agents run on clouds skip the edge.
AI traffic's worthUnknown
Whether it pays like human traffic.
Share pay is a fifth of sales, and its loans that turn into shares would add about 12% more shares.
The price and our view · 15 · How will we know?
February shows whether growth broadens
Sales of $198M or more, others growing 12% or moreGood sign
Others' growth reaches 18% in a yearWe turn positive
Others' growth 6% or less in a yearWe turn negative
FEB 202718
Fourth-quarter 2026 results. Companies report every three months; each report is a test of the story.
Our view moves on these business events, and turns negative if the business is valued at 4.5 times a year's sales or more.
The price and our view · 16 · So, what do we think?
Faster growth, but narrow, on a network priced down
NeutralMedium conviction
Fastly runs a network that makes big websites load fast and keeps attackers out. Sales grew 20%, mostly from its ten largest customers, while delivery prices fall every year and share pay adds shares. Neutral.
↑ ↓ to move · C for contents
What it does
More about the company
Who it is
When you open Reddit, the New York Times or Spotify, much of what loads comes from a nearby Fastly server rather than the company's own computers far away. Fastly keeps copies of pages, video and software close to users so they load fast, checks each request to block attacks and bad bots, and runs small pieces of customer code there too.
About 634 customers pay over $100,000 a year and bring 95% of revenue. Without Fastly they would serve from their own cloud or use Akamai or Cloudflare.
How it makes money
Delivery, 73% of revenue, is billed by the amount of traffic carried, above a monthly minimum, under contracts of about a year. Security, 24%, is mostly yearly subscriptions paid in advance, which makes it steadier. The rest is running customers' code. Contracts already signed but not yet delivered come to $333.5M, most of it due within a year.
Growth and profit
What growth costs
Fastly now spends far less of each sales dollar to grow: selling costs fell from 43 cents in 2021 to 31, and share pay from 40 cents to 20. But it is spending more on servers, 10 cents of each sales dollar against 2 in 2024, so cash grows more slowly than profit.
Its record
The numbers
Is it growing?Steady
Sales grew 21% over the twelve months to Sep 2026, faster than the 15% in the year to Dec 2025.
Is it profitable?Losing
Operating margin is −9.0%, up from −39% in the year to Dec 2023.
Does growth pay off?Pays off
Running costs take 72¢ of each sales dollar, down from 115¢ in the year to Dec 2021.
Does it earn good returns?Poor
Shareholders' money lost 6.5% a year, up from −21% in the year to Dec 2021.
Is it making cash?Solid
It turned 7.9% of sales into free cash, $56.8M over the last twelve months.
Is the balance sheet safe?Net cash
It holds $365M of cash against $324M of debt, $40.9M more cash than it owes.
Are shares being diluted?Diluting
The share count rose 38% in 4.8 years, and pay in shares costs 19% of sales.
Why it wins
The evidence
Engineers write their caching and security rules in Fastly's own tools, so moving everything elsewhere is work, and customers tend to add products rather than leave: for every dollar existing customers spent a year ago they now spend $1.19.
But most large customers also run a second network and can shift traffic without rewriting anything. That caps what Fastly can charge: delivery prices fall by a mid-single-digit percentage every year, by management's account, while traffic grows about 20%.
Who could beat it
The rivals in full
Cloudflare is the giant: it sits in front of a quarter of all websites, up from a fifth a year earlier, against under 1% for Fastly. Akamai sells the same delivery and security to the same large companies, and Amazon, Google and Microsoft bundle their own networks with the cloud accounts customers already pay for.
the web firewall, bot management and attack defence, sold with their own larger delivery networks
F5 and Thales' ImpervaFFIVHO.PA
the web firewall and API protection, from vendors whose boxes and services large companies already run
How big it could get
The market and who has signed
The market
About a third of all websites use a network like Fastly's, and Fastly is on 3% of those, mostly very large ones. Its finance chief says the delivery market grows 5% to 6% a year. Management puts the whole market at $22B, but that is the seller's number. No outside body counts what companies spend on delivery and web security, so the true size is not known.
Who runs it
Leaders, pay and the cash
Who runs it
Kip Compton became chief executive in June 2025, promoted from running products after his predecessor resigned, and Richard Wong joined as finance chief that August. Insiders own only about 4% of the company, most of it the founder Artur Bergman's, and he has sold steadily under a set plan. Nobody inside bought shares in the past year. Bonuses pay on revenue and, from 2026, on committed revenue and free cash, never on profit after share pay.
The last report
The latest report
Fastly's Q3 2026 sales rose 20% to $190.2M; outside the top ten, about 8%
What it did to the thesis Mixed quarter
It pressed on the test we wrote down: growth outside the largest customers slowed instead of broadening. Retention and security improved, but through the largest customers. The shares then fell below our price line, so the view moves to Neutral.
The report
Sales beat the top of the $184M to $190M guide, security grew 32% and existing customers spent 19% more than a year before, the most since 2022. But the ten largest customers rose to 39% of revenue and grew about 47%, so everyone else grew about 8%, down from 12% in June.
One customer is 12% of revenue, contracted revenue not yet delivered grew 24% against 38% in June, and the fourth-quarter guide of $198M to $202M implies about 16% growth.
What's happening now
The background
The shares nearly tripled in a year, almost all of it buyers paying more for each dollar of sales rather than sales per share rising. The path was violent: they fell 18% in April when Anthropic launched a way to run AI agents without edge networks, and 38% after the May results.
In September management set 2029 targets of $1.1B to $1.3B of sales. The Q3 results on November 4 were the first test after the World Cup: they beat the guide, but growth outside the ten largest customers slowed. Fourth-quarter results are due in February.
Where we disagree
Both sides in full
The market
Growth is back for good, with security and AI traffic driving it: analysts have warmed sharply in a year.
We believe
Growth is real but narrow and delivery keeps getting cheaper, but the price now pays for part of the plan rather than all of it.
The argument
The headline 20% hides where it came from: outside the ten largest customers revenue grew about 8%, down from 12% in June, and customers paying over $100,000 a year grew by only seven in a year.
The case against us is that existing customers keep spending more, now $1.19 for each past dollar. If growth outside the largest customers returns toward 18%, we are too harsh.
What it's worth
How we value it
No fair value: the analysts' forecasts show no lasting operating profit after stock pay, so a value built on earnings would rest on a guess; the bet prices it instead.
The bet
The bet is on security alone. Delivery and the rest are counted separately, at what Akamai is worth for each dollar of sales.
If security keeps growing near the top of management's range for two more years, the whole company ends up worth only a little more than today, so it is an expensive bet. If security stalls, delivery and the cash still cover nearly all of today's price, so the downside looks small, though Akamai's value is generous to a business still losing money.
The bet
Seen as a bet
The bet is on security alone. Delivery and the rest are counted separately, at what Akamai is worth for each dollar of sales.
If security keeps growing near the top of management's range for two more years, the whole company ends up worth only a little more than today, so it is an expensive bet. If security stalls, delivery and the cash still cover nearly all of today's price, so the downside looks small, though Akamai's value is generous to a business still losing money.
What could go wrong
Every risk
The risks
One customer group
One customer brings 12% of revenue, and the ten largest customers 39%; if one moves traffic, a big piece of growth goes.
Prices keep falling
Delivery prices fall every year, and Cloudflare, Akamai and the cloud companies sell the same service at larger scale.
New shares
Share pay is a fifth of sales and both convertible notes are worth converting, adding about 12% more shares.
Insiders are selling
Insiders sold 4.67M shares in a year and bought none; the finance chief sold outside a set plan near the high.
AI may route around it
AI companies running agents on their own clouds could take that traffic off edge networks, the fear behind April's fall.
What we still do not know
How much of Fastly's traffic and revenue comes from AI agents and crawlers, and whether it pays as well as human traffic.
Whether TikTok's group plans to move traffic to its own network or a rival.
How much of last year's revenue customers keep before adding more; Fastly stopped reporting it.
What changes our view
Every route to a new view
Every route to a new view
→ Unattractive
The business is valued at 4.5 times a year's sales or more, up from about 3.4.
→ Unattractive
Revenue outside the ten largest customers grows 6% or less in a year.
→ Attractive
Revenue outside the ten largest customers grows 18% or more in a year. Customers paying over $100,000 a year reach 680, from 634.
The next test
It supports the case if
Growth outside the top ten back at or above 12% and revenue at or above $198M.
It pressures the case if
Growth outside the top ten below 8% again, or revenue below $198M.
We would also be wrong if
Security grows above 30% a year while existing customers keep spending at least $1.15 per past dollar.
The verdict
Scores and how we weighed it
The scores
Growth
3 of 4
Business quality
2 of 4
Financial position
3 of 4
Execution record
3 of 4
Per share capture
2 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
No fair value is drawn: the forecasts show no lasting profit after stock pay, so the bet prices it.
BusinessNegative
A mixed business: delivery prices fall each year and share pay adds new shares.
GrowthMixed
Growth returned, but it comes from the largest customers rather than broadening.
RisksNegative
One customer group carries much of revenue, larger rivals price hard, convertible notes may add shares and insiders are cashing out.
RivalsMixed
It sits between its two listed rivals, so the group says little about the price.
The groupMixed
It sits in none of our ranked groups; internet traffic and security spending move it.
Latest newsMixed
The last report pressed on the test: growth outside the largest customers slowed, though retention improved.