Chainlink is a data company that sells one product: the current price of things.

Not to you. To software. There is a large and growing category of financial programs that run on blockchains, and those programs need to know what gold costs, what a share costs, what one currency is worth against another. They cannot find out on their own. A blockchain is a shared record that thousands of computers keep identical copies of, and the only way to keep those copies identical is to let nothing in from outside. A program living on one can see every transaction ever made on that blockchain and nothing else in the world.

So the price has to be carried in. Chainlink is who carries it.

That is the whole business. A lending platform needs to know when the collateral behind a loan has fallen too far, and it asks Chainlink. An insurance contract needs to know whether a flight landed late, and it asks Chainlink. The programs pay for the answer because without it they cannot function at all, in the same way a shop cannot open without a till.

Chainlink is not itself a blockchain. It is software that runs on other people's blockchains, the way an app runs on a phone somebody else built. That distinction matters more than it sounds, and it comes back later.

The answers do not fetch themselves. Independent businesses called node operators run Chainlink's software, take the requests, go out and get the prices, and deliver them. They are companies, in different countries, with staff and electricity bills, and they do this for money.

LINK is the money.

It is what a program pays when it asks for a price, and it is what the operator receives for supplying one. That is a real job for a token, and it is a rarer thing than the industry's marketing suggests. Most tokens are attached to a project without ever being necessary to it, the way a shop might print its own gift vouchers that nobody has to use. LINK is necessary. The work does not happen without it changing hands.

There is a second job. An operator who wants to serve has to lock LINK away first, as a bond. It sits there, unspendable, and if the operator delivers a bad price or stops delivering at all, part of it is taken. A landlord asks for a deposit for the same reason, and it works for the same reason: the money at risk is the operator's own.

Both jobs create demand for LINK that has nothing to do with anyone expecting the price to rise.

Chainlink earned $62,653,663 in the year to 24 August 2026.

Every day of that year is measured. Coverage is complete and the records are current, which sounds like a technicality and is not: for most projects this figure has to be estimated from a partial window, and an estimate can be stretched to flatter. This one cannot. It is a total.

Two channels make it up. $58,890,899 came through staking, and $3,762,764 came directly from operators being paid for individual requests. The larger number is the one to watch, because it is the network keeping a share of what passes through it rather than simply passing payments along.

A single year's figure tells you nothing about direction, and on Chainlink the direction is the more interesting number.

The record starts in August 2023. Across its whole life the measurement has captured $69,435,581, and $62,653,663 of that fell in the last twelve months. About nine dollars in every ten Chainlink has ever earned arrived in the past year.

That is a business accelerating hard, not a business coasting. And it is worth being careful about why it is credible: the year's total and the rate implied by the last ninety days agree with each other to within about one part in a hundred. A recent burst of activity stretched across a whole year is the ordinary way a revenue figure gets inflated. Here the recent months and the full year tell the same story, which means the growth is the trend rather than a spike.

The measurement covers what happens on the blockchains Chainlink runs on, and nothing else.

Chainlink's own documentation says the company has already earned hundreds of millions of dollars, substantially off the chain, from large enterprises paying for access under ordinary commercial contracts. Nothing measures that, and nothing can. A contract signed in an office leaves no public trace.

So there is a gap of roughly ten times between what Chainlink says it earns and what can be independently counted, and the two figures are measuring different things rather than contradicting each other. If Chainlink's account of itself is accurate, the business is considerably larger than the part visible here.

This is the only significant claim on this page that cannot be checked from outside. It is Chainlink's own claim, and it runs in Chainlink's favour. Worth holding at arm's length for exactly that reason, and worth knowing that every number below is the conservative version.

Owning LINK and earning from LINK are two different things, and the difference is not hidden.

Chainlink keeps a share of what flows through the network and spends it buying LINK on the open market. Nothing is destroyed and nothing is paid out like a dividend. The bought tokens go to the operators who have locked LINK away as a bond, as their reward for doing so.

LINK sitting in a wallet earns nothing. Not a small amount. Nothing.

That is the whole distinction between owning a share of a business and owning the thing a business happens to be paid in.

A shareholder is owed part of the profit by law. If the company earns a dollar, some fraction of that dollar is theirs.

A LINK holder is owed nothing of the sort. What they own is the token that Chainlink's earnings are used to buy on the open market. If the buying is heavy enough, the price of the token rises with it.

That is a different claim, and a weaker one. A share of profit is guaranteed. A rising token price depends on the buying continuing, at scale, in the same way it does now.

Anyone who wants a direct share has to put their tokens to work as a bond and accept the conditions that come with it.

One rate is worth knowing before moving on. A network taking a small cut of what passes through it is charging for a service. A network taking a large cut is running a business off it. Chainlink's documentation describes lending markets on Aave, where people borrow against tokens they deposit, in which 35% of the value recovered when a loan is closed out is collected by Chainlink.

A third is not the rate of a toll booth. It is closer to a landlord's share. For a project whose entire case rests on being neutral infrastructure that everyone can rely on, that is a number worth sitting with, and it only became visible once somebody measured it.

A price on its own means nothing. Set against what the thing earns, it starts to mean something.

Chainlink's market value is $6,544,748,170, and its fees are running at about $63,434,266 a year. That is a price of roughly 103 times the annual takings.

Put it in ordinary terms: buying the whole of Chainlink at today's price would cost about a hundred years of what it currently earns. Whether that is expensive depends entirely on the growth, which is why the ninety-percent figure above matters so much. A hundred years of takings is a lot to pay for a business standing still and is not obviously much to pay for one that earned most of its lifetime revenue last year.

For scale, the case this test was written against: a wireless network project once traded at roughly 87,000 times its annual takings. Chainlink is nearly a thousand times closer to its own earnings than that. A hundred and three is a modest number in this market, and it can only be called modest because there is something to compare it to.

And remember what the divisor leaves out. The fees measured here exclude everything Chainlink earns away from the blockchains it runs on. If the company's account of its enterprise business holds, the real multiple is lower than a hundred and three, not higher. There is no version of that omission that makes Chainlink look worse.

A token's release schedule is where a project can quietly take value from the people already holding it. Print more, pay them out, and every existing holder owns a slightly smaller piece without anyone announcing anything. It is the most common way a holder gets diluted and the least visible.

Chainlink's schedule is finished.

There will never be more than a billion LINK. Nothing at all is set aside for paying people in newly created tokens. The release schedule has run to its end: a hundred and twenty-six scheduled releases, all of them behind us, none pending, nothing left to fall.

That is uncommon and it removes a whole category of risk in one go. Nobody holding LINK is waiting for a release to land on the price. None of the network's growth is being paid for by printing more of the thing they hold. The rewards paid to operators come out of money the network actually earned.

The split between insiders and the public is 53.2% to node operators and the company, 46.8% to the public sale. The larger share went to insiders, which is worth knowing plainly. What matters as much is that the figure is final rather than current: the allocation as it stands is the allocation as it ends. Insiders cannot increase their share from here.

One quantity is still open, and it is large. 251,990,421 LINK, about a quarter of every LINK there will ever be, has no stated destination. Chainlink has not said where it goes.

It is disclosed rather than concealed, and everything else in the schedule is settled, which makes it a question about the future rather than a problem in the present. It is also the single largest unresolved thing in Chainlink's token economics, and anyone holding LINK for years should be watching what happens to it.

This is where Chainlink fails, and it fails clearly.

Chainlink Labs decides which node operators are permitted to serve the network. Since the operators are the network, choosing them is choosing what the network is. The company also controls what gets built next, controls the software upgrades, and controls the rules governing how those rules change. The founder remains at the front of the project rather than having stepped back from it.

Galaxy Research, reviewing the project independently, describes perceived centralization around a curated set of node operators and foundation stewardship.

Holding LINK buys no say in any of this. There is no vote attached to the token, none at all.

That produces an odd result worth stating honestly rather than burying. One of the standard tests asks whether a hostile party could buy up enough votes to seize control of a network. Chainlink passes it, and it passes because there are no votes to buy. The danger is genuinely absent, and it is absent because the thing that would have been captured was never built.

So Chainlink has traded one risk for another. Nobody can take it over from outside. The company can steer it from inside, change what it does and what it charges, without asking anyone who holds the token. Someone comfortable with that is comfortable owning a company's decisions, and the only real question is whether they trust that company for as long as they intend to hold.

There is a way to build these systems so that trusting anybody is unnecessary. Mathematics can prove a piece of data genuinely came from where it claims to, prove a user is a real and distinct person, perform calculations on figures that stay encrypted the whole way through. Nothing has to be taken on faith because the proof travels with the claim.

Chainlink does not work that way. It works because operators are paid well to behave, lose the money they have bonded if they do not, and have reputations to protect. That is a night watchman rather than a lock. It is a real defence and it has held for years, and it is a fundamentally different kind of thing from a guarantee that cannot fail.

Two events show what the difference costs.

A Chainlink price feed was paused during the collapse of LUNA in 2022. Lending platforms went on lending against a price that had stopped moving, and roughly eleven million dollars of bad debt landed on Venus Protocol and Blizz Finance, two of the businesses relying on that feed. In 2020, attackers flooded Chainlink's operators with requests designed to burn through the fees they pay to run anything at all.

Notice where the cost fell. Not on Chainlink. On the businesses that were using it.

That is the shape of the whole arrangement in one sentence: it works, it has worked for years, and on the occasions when it does not work the loss shows up somewhere else. This is the same failure the section above describes, seen from the other end. There it is a fact about a company holding the keys. Here it is a fact about what the network's promises are made of.