Token allocation

Six numbers per bucket, and only one of them gets discussed

A row in an allocation table is not finished when it has a percentage. It needs the token count that percentage converts to, the portion released at the token generation event, the cliff before any further release begins, the length of the vest running after that cliff, and how often the vest pays out. Six fields. A table missing any of them cannot be modeled, which means nobody can tell you what it will do.

The rows have to sum to 100 with nothing tucked inside another row. Advisor tokens carved quietly out of the team bucket, a liquidity provision that is really an exchange listing fee, a marketing reserve living inside treasury: every one of those is a place where a table sums to 100 on the slide and to something else onchain. Give each bucket its own row even when it is small, because the small rows are the ones that show up unannounced in month nine.

The interaction between rows is the entire exercise. Two buckets that each look restrained on their own can still put their first release in the same month, and the market prices the sum of what arrives, not the intent behind each piece. Allocation is built as one model or it is not built at all.

From percentage to token count to launch-day float

Run the arithmetic on an illustrative 1,000,000,000 token supply. Community takes 40 percent, so 400,000,000 tokens, with 10 percent of that bucket unlocked at launch. Investors take 18 percent, 180,000,000 tokens, nothing at launch, 12 month cliff. Core contributors take another 18 percent on the same 12 month cliff. Treasury takes 16 percent, 160,000,000 tokens, releasing monthly from day one over 36 months. Liquidity takes the remaining 8 percent, 80,000,000 tokens, fully unlocked at launch, because a pool that is not funded does not function.

Launch-day float is therefore 40,000,000 tokens from the community bucket plus 80,000,000 for liquidity. That is 120,000,000 tokens, 12 percent of supply. This one number sets the relationship between market capitalization and fully diluted valuation on day one, and it is the number most tables never actually compute before the terms are signed.

Now look at month 13. Two buckets holding 360,000,000 tokens between them begin releasing in the same month. Nothing in the percentage column shows that. It appears only once all six fields are filled in and every row is laid on one calendar, which is the argument for building the vesting schedule and the allocation table as a single artifact.

Why nobody can tell you what a normal allocation looks like

Here is the part almost every competing page gets wrong. No institutional dataset of token allocations across launched projects exists. We checked directly: a16z crypto's State of Crypto reports, Messari research, Electric Capital's developer report, CoinGecko and CoinMarketCap. None of them publish an averaged insider versus community allocation benchmark drawn from a sample of launches. The tidy ranges in circulation, community somewhere in the twenties, insiders held under forty, are folklore. They persist because each article cites the last one that said it.

What does exist is the set of allocations individual projects have disclosed themselves. That is a sample of real decisions, not an average, and the distinction matters the moment an investor asks whether your table is normal.

So the defensible answer to that question is not a percentage. It is: here is our split, here are the disclosed splits from four comparable launches, and here is why ours differs. In our view that answer holds up in a diligence call and a benchmark claim does not, because the benchmark claim collapses the second someone asks for the source. Being the one team in the room that can say where its comparison came from is worth more than a number that sounds authoritative.

What four launches actually disclosed

Uniswap's genesis supply was 1,000,000,000 UNI: 60 percent to the community, 21.51 percent to the team, 17.80 percent to investors, 0.69 percent to advisors, with the team, investor and advisor tokens vesting over four years.¹ Optimism started at 4,294,967,296 OP, split across 20 percent token treasury, 20 percent retroactive public goods funding, 19 percent user airdrops, 19 percent core contributors, 17 percent investors and 5 percent partner funds.²

Aptos launched 1,000,000,000 APT at roughly 51 percent community, 19 percent core contributors, 16.5 percent foundation and 13.5 percent investors.³ dYdX issued 1,000,000,000 DYDX at about 50 percent community, 27.73 percent to past investors, 15.27 percent to founders, employees and consultants, and 7 percent reserved for future employees.⁴

Two of those pages render their tables client side, so read the decimals as the projects' own published figures rather than values we re-extracted from raw markup. The shape of the spread survives any rounding you care to apply.

Plenty of projects publish no bucket table at all. Arbitrum's documentation covers unlock mechanics and circulating supply rather than allocation percentages.⁵ Sui's official tokenomics page states the 10,000,000,000 SUI supply cap and the one year investor cliff without publishing bucket percentages anywhere.⁶ The Sui splits that circulate widely appear on neither the docs page nor the economics whitepaper. A number being everywhere is not the same as a number being sourced.

The insider read, and what it is really measuring

Group core contributors and investors together and the four disclosed launches above land at 32.5 percent for Aptos, 36 percent for Optimism, 40 percent for Uniswap, and 50 percent for dYdX once its future-employee reserve is counted. That is a 17.5 point spread across four launches that all reached major exchanges.

An investor reading your table is not scoring that number against a norm. They are asking three things. How much supply sits with people who paid a fraction of the public price. When that supply becomes sellable. And what has to be true about the business for those holders to want to keep it past the date they could exit.

Which is why the insider percentage alone is a weak signal and the insider percentage paired with cliff and vest length is a strong one. Forty percent insider supply on a four year vest behind a one year cliff is a different instrument from 25 percent insider supply that goes fully liquid in nine months. Read the two columns together or do not read them at all.

The table cannot manufacture demand

Allocation decides who holds supply and when they are able to sell it. It has no view on whether anyone wants to buy. Across 100+ projects and $100MM+ in combined raises, the pattern that keeps repeating is a table engineered to the decimal sitting on a business with no revenue path, where the allocation work is being asked to do a job it structurally cannot do.

The token is infrastructure. The business is the engine. A community bucket is a distribution decision, and distribution only compounds when the thing being distributed has a reason to be held.

The order we build the table in

Total supply first, because every other number is derived from it. Launch-day float second, because that is what the market meets on day one and it constrains everything upstream. Locked buckets third, sized against what the business needs to fund rather than against what each counterparty asked for. Cliffs and vest lengths last, laid on a single calendar and checked specifically for the months where two or more buckets land together.

Reverse that order, which is what happens by default when allocation is handled as a sequence of deal terms, and you end up with a table that sums to 100 and still surprises you in month 13. The arithmetic was never the hard part. The sequencing is.

One caution covering all of it. Nothing here is a recommendation to buy, sell or hold any token, and whether a particular allocation or offering raises securities questions in your jurisdiction is fact-specific and belongs with your counsel.