Helen is Marketing Team Lead at Motion Trade with 4+ years in Web3 and crypto marketing. Before joining Motion Trade, she built and led the marketing function at CLS Global and managed social media campaigns for a portfolio of crypto clients at Ninja Promo. She specializes in turning complex trading products into clear stories.
There is no universal figure, because liquidity requirements are set by sell pressure rather than by market capitalisation. The workable method is to size the book so that the largest single sell you can realistically expect during launch week moves the price by less than two to three percent, and to size any on-chain pool so that a trade equal to one percent of the quote-side reserve moves the marginal price by roughly two percent. For most launches between ten and one hundred million dollars of fully diluted valuation, that calculation lands somewhere between one hundred fifty thousand and one million dollars of deployable liquidity across all venues combined.
The question arrives in almost every first call with a founder preparing for a token generation event, and it is usually phrased as a request for a number. How much liquidity do we need? The honest answer is that the number is an output rather than an input, and that two projects with identical valuations can need budgets that differ by a factor of five depending on how their tokens are distributed and where they intend to trade.
What follows is the method we use to arrive at that number, written so that a founder can run it independently before speaking to any desk. It has three inputs, one piece of arithmetic for order books, one for automated market makers, and a set of assumptions that are worth arguing about openly rather than burying in a proposal.
Start with what the book has to absorb, not what the token is worth
Fully diluted valuation is the wrong anchor because almost none of it is tradeable on day one. What determines whether your chart holds together in the first week is the relationship between two quantities: the tokens that can actually reach the market, and the depth standing ready to buy them.
So the first input is the circulating float at the moment trading opens, expressed in dollars rather than tokens. Count everything unlocked: the public sale, the airdrop or points allocation, any liquid portion of team or investor allocations, and the tokens seeded into on-chain pools. If your token launches with a fifty million dollar fully diluted valuation and eight percent circulating, the market is four million dollars in size, and every subsequent calculation should be run against four million rather than fifty.
The second input is the share of that float you expect to be sold in the first days. This is where most launch plans quietly break. Airdrop recipients who received tokens for activity rather than capital behave very differently from participants who paid for an allocation, and a distribution weighted towards the former should be modelled with sell-through assumptions of fifty to seventy percent within the first week. A sale-heavy distribution to buyers who chose the asset at a set price will typically show much lower immediate turnover. Model the two cohorts separately, because averaging them produces a number that describes no one.
The third input is your venue mix. A token opening simultaneously on two centralised exchanges and one decentralised pool needs liquidity in three places at once, and capital that is committed to one venue is not available to defend another. Deciding this before the sizing exercise rather than during it saves a great deal of retrospective argument, and our comparison of where a token should start building liquidity covers the trade-off in detail.
The order book calculation
Exchanges and traders both assess a book the same way, by looking at how much can be bought or sold within a defined distance of the mid price. The standard measure is depth at one percent and two percent from mid, on each side, in dollars.
Set your target by taking the largest single sell you consider realistic during launch week and requiring that it clears within your acceptable price impact. If your model says a single recipient cohort might dump one hundred and twenty thousand dollars of tokens in one action, and you are willing to accept a two percent move, you need roughly one hundred and twenty thousand dollars resting on the bid side within two percent of mid on the venue where that sell will land. Half of that should sit within one percent, because depth concentrated at the two percent boundary is depth that lets the price fall one and a half percent before it does anything useful.
Two adjustments matter. The first is replenishment: depth is not a static wall but a stock that gets consumed and rebuilt, so a desk quoting properly will refill the bid ladder within seconds and can therefore absorb considerably more than the instantaneous snapshot suggests, provided it has inventory. The second is that quoting depth requires both sides. A book with a deep bid and a thin ask invites exactly the volatility you are trying to avoid, and any rally will look violent and unconvincing.
There is also a floor set by the exchanges themselves rather than by your ambitions. MEXC, for example, publishes ST Warning Rules under which a token whose average daily bid-ask spread exceeds two percent for fifteen consecutive days becomes eligible for a warning tag and the delisting process that follows it. Whatever your own targets, the book has to clear the venue's minimum standard continuously, not on average.
The automated market maker calculation
On-chain pools follow a different and considerably more predictable rule. In a constant product pool, the marginal price moves approximately twice the proportion of the reserve that a trade consumes. A buy equal to one percent of the quote-side reserve moves the marginal price by roughly two percent; a buy equal to five percent moves it by roughly ten.
This gives you a direct way to size the pool. Decide the largest on-chain trade you want to accommodate at an acceptable impact, then multiply. A pool intended to absorb a twenty thousand dollar swap within two percent needs roughly two million dollars on the quote side, which is far beyond what most launches deploy. A more common target is to accept five to eight percent impact on a trade of that size, which implies a quote-side reserve in the range of two hundred fifty to four hundred thousand dollars, matched by an equivalent value of tokens.
Two practical notes. Pool liquidity is symmetric, so half of every dollar you commit is denominated in your own token and is exposed to the price decline it is there to cushion. And concentrated liquidity positions can achieve the same depth with substantially less capital within a defined range, at the cost of falling out of range entirely if price leaves the band, which during price discovery it frequently does.
Three worked examples
A points-distribution launch at forty million fully diluted valuation
Twelve percent circulating gives a float of four point eight million dollars, most of it distributed to farmers. Modelling sixty percent sell-through over the first week produces roughly two point nine million dollars of expected supply, arriving unevenly with the heaviest concentration in the first forty-eight hours. This launch is capital-intensive relative to its valuation: it needs meaningful depth on two venues, an on-chain pool in the three hundred thousand range, and a desk with enough inventory to keep absorbing without exhausting its position. Total working liquidity of six hundred thousand to one million dollars is a realistic band, and a launch of this shape attempted with two hundred thousand will produce the chart everyone recognises.
A sale-led launch at twenty million fully diluted valuation
Six percent circulating gives a float of one point two million dollars held mainly by participants who bought at a known price. Expected first-week sell-through of twenty to thirty percent implies around three hundred thousand dollars of supply. One centralised venue plus a modest pool is defensible here, and total working liquidity of one hundred fifty to three hundred thousand dollars is proportionate. Spending a million on this launch would not improve the outcome; it would simply idle.
A tier-one listing at one hundred fifty million fully diluted valuation
The constraint is no longer sell pressure alone but the venue's expectations. Larger exchanges assess quoting quality continuously and compare your book against everything else in the same category, which means committed depth on several pairs, tight spreads maintained through volatility, and coverage across all trading sessions. Budgets here are set by the number of venues and the depth commitments written into the agreement rather than by float arithmetic, and they generally begin in the high six figures. Our breakdown of what a listing actually costs covers the surrounding budget lines.
What the money is actually for
A common source of confusion is treating the liquidity figure as a fee. In most structures it is not spent, it is deployed. Inventory committed to quoting remains yours in a retainer arrangement, and what you pay for is the desk's capital efficiency, infrastructure and coverage rather than the inventory itself. In a loan structure the tokens are lent rather than sold, and the cost is expressed through options on your supply. The three models produce very different balance sheets from the same headline number, which is why we set them side by side in retainer versus profit-sharing versus the loan model.
The part that genuinely is spent is thinner than founders expect: the retainer, the exchange-side costs, and the losses a desk absorbs while quoting into a falling market. That last item is real, and it is the reason a serious desk will ask about your unlock schedule before quoting a price.
The number changes after launch, and usually downwards
Launch-week sizing is a peak requirement rather than a run rate. Once the first distribution cohort has finished selling and the token has found a range, the depth needed to maintain a healthy book typically falls, sometimes by half. What replaces it is a continuity requirement: the book has to remain within venue thresholds every day, including the days when nothing is happening, because those are the days exchange monitoring systems are measuring.
The failure mode we see most often is a project that funds launch generously, watches the chart stabilise, and then reduces liquidity provision at exactly the point where the next unlock cliff arrives. The sizing exercise is therefore worth running three times: for launch week, for the ninety days after, and for the first major unlock. The reasons that sequencing goes wrong are set out in our analysis of why most tokens fail in the first ninety days after TGE.
What to put in the agreement
Whatever number the model produces, it belongs in the contract as a set of measurable commitments rather than as a headline. Specify minimum depth in dollars at one and two percent from mid, per venue and per pair. Specify a maximum spread and the uptime percentage at which it must be maintained. Specify what happens during defined volatility events, because every desk widens at some point and you want the threshold written down rather than discovered. Our guide to what belongs in a market making agreement sets out the full list.
FAQ
How much liquidity does a token need at TGE?
Enough to absorb the largest realistic single sell of launch week within an acceptable price impact, which for most launches between ten and one hundred million dollars of fully diluted valuation works out between one hundred fifty thousand and one million dollars across all venues. The determining factor is circulating float and how it was distributed, not valuation.
Is fully diluted valuation a useful basis for sizing liquidity?
No. Locked tokens cannot be sold, so they place no demand on the order book. Size against circulating float and expected sell-through instead, and revisit the calculation before each unlock cliff.
How deep should the order book be at one and two percent from mid?
As a starting point, depth within two percent should cover the largest single sell you consider realistic in launch week, with at least half of it concentrated within one percent. Depth stacked at the outer boundary allows the price to fall most of the way before it provides any support.
How large should the DEX pool be?
In a constant product pool, a trade consuming one percent of the quote-side reserve moves the marginal price by roughly two percent. Decide the trade size and impact you will accept, then size the reserve accordingly. Accommodating a twenty thousand dollar swap at five to eight percent impact implies a quote side of roughly two hundred fifty to four hundred thousand dollars.
Does an airdrop change the liquidity requirement?
Substantially. Recipients who received tokens for activity rather than capital sell at much higher rates and much faster than buyers who chose the asset at a price, so a points-led distribution can require several times the depth of a sale-led launch at the same valuation.
Is the liquidity budget money that gets spent?
Mostly it is deployed rather than spent. Inventory used for quoting remains the project's asset under a retainer, and what is genuinely consumed is the retainer itself, exchange-side costs, and the inventory losses incurred while supporting a falling market.
Can a project provide its own liquidity at launch instead of hiring a desk?
It can seed an on-chain pool without help, and many do. Quoting a centralised order book continuously to venue standards is a different task, requiring infrastructure, staffed coverage across every session, and inventory management under volatility. Exchanges assess the result rather than the intent, and thin books are penalised regardless of who was running them.
How long does launch-level liquidity need to be maintained?
Peak requirements usually persist for the first two to four weeks, after which the run rate falls as the initial distribution cohort finishes selling. What does not fall is the continuity requirement: venue monitoring runs daily, so the book has to stay above thresholds on quiet days as well as active ones.
What happens if a token launches under-liquid?
Spreads widen, ordinary sells move price disproportionately, and the resulting chart discourages the buyers who would otherwise provide natural depth. The condition compounds, and on several venues it also triggers formal monitoring: MEXC treats a sustained spread above two percent as grounds for a warning tag, and other exchanges apply comparable tests.
