How to read a token's supply and emission schedule

A token's supply schedule is the set of rules that decide how many units exist, how many more will be created, when they become available to sell, and who receives them — and almost every serious misreading of a crypto project starts by confusing two of those things.

The three supply numbers

Every token has at least three supply figures and they are routinely quoted as if they were one. Getting them apart is most of the work.

Circulating supply

The number of units that exist and are not locked, reserved or otherwise unavailable. This is the figure that should sit behind any market capitalisation calculation, and it is also the softest of the three, because whoever compiles it has to make judgement calls: does a treasury holding count? Tokens in a multi-year vesting contract? Coins provably lost?

Different data providers answer those questions differently, which is why circulating supply for the same asset can vary between sources. When it matters, find out what the provider counted.

Total supply

Everything that currently exists, including tokens that nobody can move yet. Locked team allocations, escrowed treasuries, unvested investor tranches: all counted here, none counted in circulating supply.

The gap between circulating and total is one of the most informative numbers in tokenomics, and one of the least quoted. A project where circulating supply is 20% of total has 80% of its eventual supply waiting somewhere, and that supply will arrive on a schedule somebody has already written down.

Maximum supply

The ceiling the protocol will ever permit. Three things can be true here, and knowing which one you are looking at matters more than the number itself.

• A protocol constant. Enforced by consensus rules that every node checks. Bitcoin's 21 million works this way, as does Avalanche's 720 million and Cardano's 45 billion. Changing it requires changing what the network accepts as valid.
• A fixed initial mint with no issuance function. Everything was created at once and there is no mechanism to make more. XRP and LINK are in this category. The ceiling holds because there is no code path to exceed it.
• A programme target. A number an entity is working toward through discretionary action, such as BNB's stated floor of 100 million. It holds because someone keeps doing the thing. That is a different kind of assurance from a consensus rule, and it deserves to be described differently.

And of course a maximum supply may simply not exist. Ethereum has none. Dogecoin has none. That is not automatically bad, and the section on emission explains why.

Market cap and the number that is not market cap

This is the single most consequential confusion in the subject, so it gets its own section.

Here is the shape of the problem. A project launches with 5% of its tokens circulating. At a given price, its market cap looks modest — perhaps small enough to sound like an opportunity. Its fully diluted valuation, on the same price, is twenty times larger.

Neither number is wrong. They answer different questions. Market cap tells you what the available float is worth. Fully diluted valuation tells you what the whole thing would be worth if every token existed today and the price did not move.

That last clause is where the second error lives. Fully diluted valuation assumes constant price across a supply expansion, which is precisely the assumption that a supply expansion tends to break. It is not a forecast. It is a way of asking: how much of what I am buying has not been created yet?

The practical habit is simple and worth building. Whenever you see a market cap, look up the ratio of circulating to total supply. If it is near one, the two valuations are close and you can move on. If it is far from one, you are looking at a project whose supply story is still mostly in the future, and you should go and read the emission schedule before anything else.

Reading an emission schedule

An emission schedule tells you how new tokens come into existence. There are a handful of common shapes and each one behaves differently over time.

Step decay

A fixed issuance rate that drops by a fixed proportion at fixed intervals. Bitcoin's halving is the canonical example: the block subsidy halves every 210,000 blocks. The result is a staircase, and because each step is half the last, the great majority of supply arrives in the early steps. By the time you get to the fifth or sixth step, the remaining issuance is a rounding error against what already exists.

Rate decay

An inflation rate that declines continuously toward a floor. Solana works this way: a starting rate, an annual reduction applied to the rate, and a long-term floor it approaches but does not go below.

The trap here is applying the decay to the wrong quantity. A −15% disinflation rate does not mean supply falls 15%. It means the inflation rate is 15% lower next year than this year: 8% becomes 6.8%, then 5.8%, and so on. Supply keeps growing throughout.

Balance decay

A fixed fraction of a shrinking pot, released each period. Cardano does this: every epoch takes a set percentage of whatever remains in an unissued reserve. Mathematically this converges the same way rate decay does, but it is driven by a balance rather than by a percentage, and the pot never quite empties.

Fixed absolute issuance

The same number of new tokens every period, forever. Dogecoin issues 10,000 per block with no end. This produces something people find counter-intuitive: the inflation rate falls every year without any halving, because a constant numerator over a growing denominator shrinks. Five billion new coins against a base of 100 billion is 5%; against 200 billion it is 2.5%.

Polkadot deliberately moved to this shape in 2024, swapping a constant rate for a constant amount, and then reduced the amount again by a later vote — which brings us to the last shape.

Governed issuance

Issuance decided by an on-chain vote and changeable at any time. If a network's monetary policy is a governance parameter, then no forward projection of its supply is more reliable than your guess about future votes. That is not a criticism; it is a property you need to know about, and it means checking the current rule rather than a number in an article.

Cliffs, vesting and why the shape matters

Emission covers tokens being created. Unlocks cover tokens that already exist becoming movable. For a holder the second is often the more immediate concern, because unlocked supply can be sold today.

Two release shapes dominate, and they behave nothing alike.

Linear vesting releases a steady trickle over a period. The market absorbs it continuously, and by the time anyone notices, it has been happening for months.

A cliff releases a block all at once on a date. Nothing, nothing, nothing, then a large tranche becomes transferable in a single moment. Cliffs are usually used to keep early participants committed for a minimum period, which is a reasonable design goal, and the side effect is a supply event with a known date.

Most real schedules combine them: a cliff at the end of year one, then linear release across years two and three. Reading such a schedule means asking three questions.

1. How big is the tranche relative to circulating supply? A release worth 2% of float is noise. One worth 40% is a different asset afterwards.
2. Who receives it? An unlock to a team under a further internal lock behaves nothing like one to a fund whose mandate is to distribute.
3. Is it publicly known? A scheduled unlock that everyone can see has had time to be priced in. A discretionary release nobody expected has not.

The archive's unlock calendar collects the scheduled releases for the assets recorded here, as a dated snapshot with a source on each row.

How burns fit in

A burn destroys tokens by sending them somewhere unspendable. Burns run against emission, and the useful question is never “does it burn?” but “what drives the burn, and how does that compare with what drives issuance?”

• Fee burns destroy part of what users pay. The rate is a function of network usage, so it varies continuously and cannot be projected from the protocol alone. Ethereum burns a computed base fee; TON burns half of transaction and storage fees; Avalanche burns the entire fee, both components.
• Programme burns destroy tokens on a schedule set by an entity, sometimes with a formula. BNB's quarterly auto-burn is the clearest example in this archive.
• Anti-spam burns destroy a small amount per transaction as a cost of using the network. XRP does this; the monetary effect is incidental to the purpose.

What a burn is not is a guarantee of a shrinking supply. Where issuance and burn both run, net supply change is the difference between them, and that difference can have either sign. A project that burns tokens while issuing more of them is not deflationary; it is a network with two flows.

Who gets the new tokens

The last question, and one that supply charts never show: where does the new issuance go?

To miners, in proof-of-work networks, who face real electricity costs and therefore sell a significant share of what they receive. To validators and stakers, in proof-of-stake networks, where a good deal of it is often restaked rather than sold. To an on-chain treasury, in networks like Polkadot and Cardano, where it is spent on approved proposals. Or to a company's balance sheet, in projects that retained a large allocation.

Two networks with identical issuance rates can have completely different sell pressure depending on who receives the tokens and what those recipients need to do with them. If you are trying to reason about supply hitting the market, the recipient matters at least as much as the amount.

The tokens that exist but cannot move

There is a category between locked and liquid that neither supply figure captures well, and it is large enough to matter.

Staked tokens exist, are counted in circulating supply, and are not available to sell today. On Polkadot that gap is twenty-eight days wide, because unbonding takes that long. On Cardano it is zero, because delegation does not lock anything. Both networks have high staking participation and the practical consequence is completely different.

Tokens sitting in contracts are similar: collateral backing a loan, liquidity in a pool, assets bridged to another network. All counted, none straightforwardly available.

And then there are coins that are simply gone. Keys lost, addresses that provably cannot spend. Nobody knows how many, estimates vary widely, and no data provider deducts them consistently. For an asset with a long history this is not a rounding error.

None of this makes circulating supply useless. It makes it a starting point rather than an answer, and it means two assets with the same headline float can have very different amounts genuinely loose.

A worked example, with made-up numbers

The figures below are invented to show the arithmetic. They are not any real project.

A token has a maximum supply of one billion. Two hundred million circulate. The price is one dollar.

• Market capitalisation is 200 million dollars. That is what the available float is worth.
• Fully diluted valuation is one billion dollars, at the same price.
• The ratio of circulating to maximum supply is 20%. Four fifths of the eventual supply does not exist yet or cannot move yet.

Now read the schedule. Say a hundred million tokens unlock in a single tranche next quarter. That is 50% of the current float arriving at once. For the price to hold, demand has to absorb an amount equal to half of everything currently trading.

None of this predicts anything. It just tells you what has to happen for a number to stay where it is, which is a more useful thing to know than a forecast. And it is entirely computable from published information, which is why the schedule is worth ten minutes before anything else.

Where to actually check

Everything above is only worth anything if you go and look. In rough order of reliability:

1. The protocol's source code. Consensus constants are in the code. Bitcoin's halving interval is a line in a file. This is the highest-confidence source there is.
2. A block explorer or the chain itself. Issued supply, balances, unlock contracts: all readable directly. Not a claim, an observation.
3. Official documentation. Good for mechanism and design intent. Watch for stale pages — an official page can be out of date, and some carry warnings saying so. Check when it was last revised.
4. Official announcements and governance records. Where policy changes are recorded. For governed issuance, this is where the current rule actually lives.
5. Data aggregators. Convenient, and they make judgement calls you cannot see. Fine for a rough figure, not for a claim you are going to rely on.

One habit worth adopting: when you write down a supply figure, write down the date you checked it. Any number that moves is only true as at a moment, and a figure without a date is a figure you cannot audit later. Every record in this archive does exactly that, which is why they say “checked August 2026” rather than pretending to be live.

Six ways people get this wrong

Collected from the fourteen records in this archive, in rough order of how often they appear in general coverage.

1. Quoting fully diluted valuation as market cap, or the reverse. Whichever number supports the point being made. Check which supply figure was used.
2. Treating a burn as proof of scarcity. A network that burns while issuing has two flows and a net that can go either way. Ethereum's record works through this properly.
3. Treating escrowed tokens as destroyed. Escrow paces market access. The tokens still exist and are still counted.
4. Applying a disinflation rate to supply instead of to the rate. A 15% annual reduction of an 8% inflation rate gives 6.8%, not −7%. Supply still grows.
5. Calling every ceiling a hard cap. A consensus rule, a fixed initial mint, a programme target and a governance decision are four different kinds of assurance. The supply structure table separates them into their own column for exactly this reason.
6. Repeating a number without its date. Anything that moves is only true as at a moment. A figure with no date attached cannot be checked and cannot be trusted later.

If you have worked through all of this for a coin you are interested in and reached the point of wanting to actually buy some, the mechanics of doing that — the account, the verification, choosing the right pair, and the question of where the coins live afterwards — are covered in the practical walk-through.

Primary references

• Bitcoin Core's mainnet consensus parameters for the 210,000-block subsidy interval.
• Anza's implemented Solana inflation schedule for rate decay and the terminal range.
• Cardano's supply overview for the reserve-pot example.

This guide teaches a method rather than reporting figures, so it carries no dated numbers of its own. Every specific example links to the record where that project's figures are set out with their sources and the month they were checked.

Next: the supply structure table puts all fourteen assets side by side · the unlock calendar collects the scheduled releases · all fourteen records