Supply analysis: emissions, unlocks, float and market makers

Circulating supply is a declaration

The tokenomics guide covers the four numbers every token page shows: circulating supply, total supply, max supply and fully diluted valuation. The first thing an analyst learns is where the most important of them comes from. CoinGecko’s published methodology says it plainly: circulating supply is obtained from the token teams and verified by CoinGecko, and for smart contract tokens it is calculated by deducting locked tokens from total supply, using the addresses of those locked tokens as supplied by the team. Premined coins count as circulating, because whoever controls them can sell them.

So the figure is only as good as the list of addresses the team hands over. Anything not on the list counts as circulating, whether or not it trades, and anything held under an agreement that never touches the list does not appear at all:

None of this implies bad faith. It means the number on the dashboard answers "what has the team declared unlocked?", which is a different question from "what could be sold next month?"

Float: the supply that actually trades

Analysts borrow the equity market idea of free float: circulating supply minus balances held by insiders, treasuries, vesting contracts and other holders that are not realistically trading. Float is what sets the price, and it can be far smaller than circulating supply.

A low float, high FDV launch prices a small tradable supply at a valuation that assumes the whole supply is worth the same per token. Binance Research’s May 2024 report "Low Float & High FDV: How Did We Get Here?" put numbers on the trend: tokens launched early in 2024 had an aggregate market cap of just 12.3% of their FDV as of April 2024, recent launches had circulating supplies as low as 6% with none above 20% in its sample, and roughly 155 billion dollars of tokens were due to unlock between 2024 and 2030. It estimated that about 80 billion dollars of new demand would be needed just to hold the 2024 launches at their prices as that supply arrived.

Modeling an unlock against real volume

A cliff releases nothing until a date; linear vesting then releases an equal amount each period. What matters is not the size of an unlock in tokens but how it compares with what the market can absorb. Here is a worked model for an illustrative token (every number below is invented for the example): 1 billion total supply, 15% liquid at launch, ecosystem grants of 30% vesting linearly over 48 months, investors with 25% on a one year cliff then 24 months linear, the team with 20% on a one year cliff then 36 months linear, and a 10% treasury with no schedule.

# Illustrative token: 1 billion total supply, 15% liquid at launch.
TOTAL = 1_000_000_000
LAUNCH = 0.15 * TOTAL

def released(month):
    eco  = min(month, 48) / 48 * 0.30 * TOTAL                  # linear from launch
    inv  = max(0, min(month - 12, 24)) / 24 * 0.25 * TOTAL     # 1 year cliff, then 24 months
    team = max(0, min(month - 12, 36)) / 36 * 0.20 * TOTAL     # 1 year cliff, then 36 months
    return LAUNCH + eco + inv + team                           # treasury 10% is unscheduled

for m in (1, 12, 13, 24, 36, 48):
    unlock = released(m) - released(m - 1)
    print(m, released(m), unlock, unlock / released(m - 1))

Month 13 is the event: both cliffs end and monthly supply growth jumps from 2.9% to 9.9% of circulating supply. Now set that against the market. Give each recipient a sell assumption:

Suppose the token trades at 0.40 dollars with 2 million dollars of daily volume, which is 5 million tokens a day or 150 million a month. The month 13 unlock alone equals 4.4 days of all trading. Expected selling of 8.47 million tokens is 5.6% of the month’s volume, which means someone has to be a net buyer of about 3.39 million dollars a month just to hold the price where it is.

Two refinements make the model honest. First, use organic volume: reported volume includes wash trading, which absorbs nothing. Second, look at order book depth, the dollars it takes to move the price by 2%, because a seller in a hurry meets the book, not the daily average. The unlock and dilution calculator gives the mechanical dilution of any single unlock; this model adds who is selling and into what.

Emissions, burns and real dilution

Unlocks release supply that already exists. Emissions create new supply, usually as staking rewards or liquidity incentives, and burns destroy it. What matters is the net:

A protocol issuing 50 million tokens a year on a supply of 1 billion while burning 20 million from fees has net inflation of 3%, not 5%. Whether the burn holds up depends on fees, which fall when activity does, so model the burn under a quiet market as well as a busy one.

Staking rewards are dilution with an opt out. Stakers receive new tokens roughly in proportion to what they hold, so their share of the network stays about constant, while everyone who does not stake is diluted:

With an 8% staking yield and 6% supply growth, the real yield is 1.89%, and a holder who does not stake sees their share of the supply shrink by 5.66% a year. A headline yield is only income to the extent that it beats supply growth. Buybacks work in the opposite direction, and trackers such as Tokenomist now report buybacks and burns next to unlocks, so the three can be netted in one place.

Market maker agreements: supply you do not see

New tokens need someone quoting both sides on exchanges from the first minute. The arrangements usually take one of two broad forms: a retainer, where the market maker uses its own inventory and charges a fee, or a loan and call option structure, where the project lends tokens to the market maker and grants it options to buy some or all of them later at set prices. The second has three consequences for a supply analyst.

Take an illustrative case: 20 million tokens lent for 12 months at a 0.50 dollar listing price, with calls on 10 million at 0.60 and 10 million at 0.75.

The loan is circulating supply from day one. In the model above, 20 million tokens is 13.3% of the launch float, and it appears in no unlock schedule. The market maker is protected on the way down. It owes back a number of tokens, not a dollar value, so tokens sold at 0.50 can be bought back at 0.30 to repay the loan, while the calls capture a rally. Hedging shapes the price path. A market maker hedging long calls sells tokens as the price rises and buys as it falls, which dampens moves in both directions. The option pricing behind that is in the guide to options and the Greeks.

Not every firm calling itself a market maker provides liquidity. On October 9, 2024 the SEC charged three firms it described as "so-called market makers", ZM Quant, Gotbit and CLS Global, with selling wash trading as a service to token promoters, including for a token that the FBI had created as part of a parallel investigation. Terms of legitimate agreements are rarely published, so look for disclosures in token documentation and exchange listing announcements, and watch for large transfers from project wallets to trading firms shortly before a listing.

Treasury and foundation wallets

Treasuries and foundations often hold the largest balances of all, with the least structure: no vesting, spending decided by a team or a governance vote. Find them in the documentation, confirm them onchain, and check their custody and signers. Then watch what they do. Transfers from a treasury to exchange addresses are the clearest early signal of selling, and the techniques in following the money apply directly.

Who is unlocking matters as much as how much. Arbitrum’s monthly unlock in mid 2026 was 92.65 million ARB, just under 2% of circulating supply, and it goes to two groups: the team, future team and advisors, and the investors. Those two groups have different costs, different horizons and different constraints, and a serious model gives each its own assumption rather than one blended sell rate.

A supply due diligence routine

1. Rebuild the allocation table from primary documents

   Take the percentages, cliffs and vesting periods from the project’s own documentation or token contract, not from a summary. Note anything with no schedule at all, such as a treasury or ecosystem fund.

2. Map allocations to addresses

   Find the vesting contracts, multisigs and treasury wallets, and check who can move them. A multisig with a low threshold or an admin key that can release vesting early changes what "locked" means.

3. Reconcile reported circulating supply with your float

   Subtract insider, treasury and contract balances from the reported figure. The gap is the part of the market cap that is not really trading.

4. Model the schedule against real volume

   Compute each month’s unlock as a share of circulating supply and of organic trading volume, with a range of sell assumptions per recipient.

5. Net out emissions and burns, then watch the wallets

   Add scheduled emissions, subtract burns, and set up alerts for transfers from insider wallets to exchange addresses, which the guide to onchain SQL shows how to query.

It pairs naturally with the broader process in researching a token, and the design side of the same problem, how emissions and incentives should be built in the first place, is the subject of token mechanism design. None of this is a view on any token’s price: it is a way to see the supply side clearly before you form one.

Common questions

Where to go next