How to Launch a Token on a DEX

Is a DEX-only launch the right choice for my token?

A DEX-only launch is the right choice when three things hold at once: a strong community, a narrative that fits the moment, and a low market cap. Our read is that those conditions are not independent preferences but a single test. A low cap means the float can be absorbed by organic demand rather than requiring a venue to manufacture it, and a strong community with a narrative people can repeat is what supplies that demand in the first hours. Where one of the three is missing, a DEX-only opening tends to expose the gap faster than a centralised listing would, because there is no exchange order book and no exchange audience standing in for interest the project has not yet earned.

The structural case for going DEX-first has also improved. DEX-to-CEX volume ratios have risen roughly sixfold over five years, which is why we treat on-chain liquidity as a credible primary venue rather than a holding pattern until a tier-one listing arrives. Five years ago a pool was where a token waited. Now it can be where a token trades for its whole first phase, with the depth, the fee tier and the routing doing the work an exchange market maker would otherwise do. The trade-offs between the two venues differ in execution mechanics more than in legitimacy, and we have written separately on how market making differs between centralised and on-chain venues.

Where we push back is on teams treating DEX-only as the cheap option. It is the honest option. A pool quotes continuously against whoever shows up, and a thin community is visible in the first chart rather than hidden behind a listing announcement. That is why the community side of the test is the one we spend most time on before a date is set, and why building the community before the launch is a sequencing question rather than a marketing one.

For projects that fail the test, the usual answer is not to abandon the DEX but to change what the DEX is being asked to do: a smaller opening float, a tighter range, and a second venue added once a returning bid exists. The decision point comes back at the moment a listing is offered, and at that point the pool's own trade history is the evidence a team has about whether the demand was ever there.

How much liquidity do I need to fund and who covers the cost?

The issuer covers the cost. A market maker decides how depth is concentrated and where it sits in the range, but the capital funding that depth comes off the issuer's balance sheet, in tokens on one side and a quote asset on the other. Thin DEX depth is the market maker's problem to solve, with concentrated liquidity at the price points where execution actually happens, and our read is that negotiations go wrong when an issuer assumes the mandate includes the inventory as well as the placement.

On sizing, what we watch is where holders trade rather than a headline number. Depth gets split between centralised venues and DEX pools according to where the flow actually lands, because capital sitting in a pool nobody routes through does nothing for spread or slippage on the venue that matters. We have written in more detail on how the two venue types differ in practice and on how much depth a launch needs and where it belongs.

There is one route to making the cost partly self-funding. Where a perps market exists on the token, hedged liquidity earns funding carry, and that carry offsets part of what the issuer puts up to hold depth in place. Our position is that this changes the arithmetic of a DEX launch, because the same inventory that absorbs flow is also generating carry while it sits there.

That offset only exists once there is a perp to hedge into. A token launching DEX-first with no derivatives market has no carry on day one, and the full capital cost of depth sits with the issuer until a perp venue lists the pair and the hedge can be put on.

Not investment advice.

What must be locked down before the token goes live?

Supply, vesting and launch valuation have to be settled roughly 48 hours before the token goes live, because after that point market making is the only lever still available. Once the contract is deployed and the schedule is published, the distribution curve is immutable. What remains adjustable is market structure: where depth sits, how it is maintained across venues, and how quoting responds to flow. That lever does not reach back into a supply schedule that shipped wrong.

Float is the part of the schedule we spend the most time on before the window closes. Our read, and the figures below are illustrative of where we have seen market reaction change rather than a target for any particular token, is that a float below roughly 6 to 7% tends to put a token back in the low-float high-FDV cohort that buyers have learned to discount on sight. Below roughly 5 to 6%, the mechanism gets worse: each cliff is large relative to the circulating base, so the unlock schedule starts breaking supply-shock thresholds on its own arithmetic regardless of how the token is marketed. How much float is right for a given token depends on how much real demand exists at listing, which is why we work the number alongside the rest of the schedule in our tokenomics design framework rather than in isolation.

Launch valuation is the second thing that has to be fixed before the lock. Our view of this market is that capital is selective rather than absent, so the gap between listing price and what the first stretch of unlocks will ask the book to absorb is what we watch. Launches priced above fair value take harsher post-listing drawdowns, and launches priced conservatively relative to their own unlock schedule tend to hold up through the cliffs. The question we put to a team is whether the price at listing leaves room for the supply arriving behind it.

Alongside those, the operational pieces have to be signed and funded before the 48 hour mark: the market making agreement and its parameters, the inventory and quote capital sitting on the venues where it is needed, and the pool configuration on each one. A team still negotiating a market maker term sheet in the final week is negotiating against a deadline, and that is where the weaker proposals get signed. From TGE onward the conversation is execution only: spread, depth, venue coverage, and what the daily reporting shows.

Not investment advice.

What happens to price after unlocks, and is that manipulation?

Price usually falls when a cliff hits, and our read is that this is the schedule doing exactly what it said it would do rather than anyone manipulating the market. The supply was published before launch. Everyone trading the token could read the date. When new tokens become transferable and the order book has to absorb them, the clearing price moves to where the standing bids are, and on most DEX pairs those bids are thinner than the size arriving.

Our position is that unlock schedules move price more than launch-day hype does. The first week of trading is noisy and reversible. A cliff is neither. It is a known increase in circulating supply landing against whatever demand the project actually built in the months it was locked. If buyers did not grow while tokens were locked, price has one direction to go on unlock day, and no amount of market making on the day changes the arithmetic of more sellers meeting the same depth.

That is why we treat the question of manipulation as mostly a distraction. The behaviour that gets labelled manipulation after a cliff is normally just early holders selling into a book that was never deep enough for them, plus a chart that makes it look coordinated because the dates were coordinated. What is worth examining is the gap between the schedule and the demand curve, which is the kind of structural problem covered in our write-up of the tokenomics mistakes that surface after TGE.

What we watch in the run-up to a cliff is whether the holder count and the organic volume on the pair have grown since the last one, and whether depth at the levels where unlocked holders are likely to sell has been funded by anyone. Those two measures say more about the next unlock than the current price does. After the cliff clears, the float is permanently larger, and the following unlock date is already published and arrives against that same base of buyers.

Not investment advice.

Should I run DEX and CEX market making together after launch?

Yes, and in the first 90 days after TGE we run DEX and CEX market making together rather than treating them as sequential phases. The two venues do different jobs. DEX liquidity is the fallback that cannot be revoked: the pool sits onchain, it quotes at whatever depth it was funded to, and no exchange risk team can switch it off during a bad week. CEX is where distribution and price discovery usually sit, because that is where the order flow, the listing audience and the reference price for everyone else arrive.

Running only one of them leaves a gap that shows up at the worst moment. A token with CEX depth and a thin pool has no quoted price at all if a venue pauses deposits or delists, and holders who hold onchain have nowhere to transact. A token with a deep pool and no CEX presence is discoverable only by people who already know the contract address, and the price formed there is the price a small number of participants are willing to make. Quoting both in parallel means the onchain pool and the book are referenced against each other instead of drifting, which is the practical version of the problem we wrote about in the real cost of fragmented liquidity.

The structure of the mandate matters as much as the venue mix. Our market making mandates are retainer-only: the client retains all capital and PnL, and we trade through API keys with no custody, no token loan and no options. There is structurally nothing to rug, because no tokens leave the client's control and no counterparty ends up holding a call on supply it can exercise against the chart. That is a different arrangement from the loan plus options deal that is still common, and the trade-offs between the two are set out in our comparison of loan plus options and retainer structures.

In practice the parallel setup also changes what happens when something breaks. If a CEX pauses withdrawals, quoting continues onchain and the team can point holders somewhere real on the same day. If the pool is drained by a sharp move, the book absorbs flow while the pool is rebalanced. The first 90 days are when both of those events are most likely, and when the two venues are being watched against each other most closely by anyone deciding whether to list the token next.

Not investment advice.

What if an exchange asks for a token grant before listing?

Our position is that a venue asking for a token grant to itself gets that demand converted into a user campaign, or the listing waits. The tokens still leave the treasury, but they go to traders who do something to earn them rather than to the venue's own balance sheet.

The reason we hold that line is what a grant to the venue creates on day one. It is a holder with no cost basis, no lockup it negotiated, and no reason to care where the price sits when it decides to move. That supply lands in the same book our quotes are in, and the team that gave it away has no visibility into when it gets sold. A user campaign puts the same tokens in front of people who trade the pair, and the distribution is tied to activity the team can see.

When a venue will not accept the redirect, we are willing to delay the listing rather than concede. Our read is that the cost of waiting is usually smaller than teams assume at the point of negotiation, because the venue is asking for the grant precisely when it expects the listing to bring it volume. A team that has sequenced the rest of its launch properly, which is the part we cover in the full TGE checklist, is not forced to accept a term on someone else's timetable.

What we watch after the redirect is agreed is whether the campaign terms are written down with the same specificity as the grant would have been: how many tokens, over what period, against what measured activity, and who holds the undistributed balance in the meantime. A campaign with the size fixed and the schedule left open is the grant again under a different name, and the venue gets the discretion the team was trying to keep.

Can I use a treasury buyback to support the DEX market after launch?

A treasury can buy its own token on a DEX after launch, and whether that exercise does anything depends almost entirely on what pays for it: funding from the native token is circular, funding by cutting runway trades operational survival for a price signal, and neither does anything useful or fixes an underlying product problem.

The circularity is the part issuers underweight. Using the same asset to create the bid as the one the bid is meant to affect means the treasury is selling into itself, and the net position after the print is worse than before. Cutting runway is a different failure with the same ending, because the operating budget is the thing that produces the reason to hold the token in the first place. Our read is that a buyback layered on top of weak demand buys a short window and nothing past it, and that the diagnosis work belongs with how the non-native side of the balance sheet is held and drawn, which we treat as part of treasury management rather than market operations.

On execution, the only problem to solve is the immediate front-run on the print. Issuer foreknowledge is not itself an issue, because the buyback is announced on purpose so that others can buy alongside it. What we watch is the window between the announcement and the fills, where a bot sitting in front of the pool takes the move the treasury is paying for, which on an AMM shows up as the buy walking the curve into someone else's inventory rather than into the treasury's.

The legal footing is thinner than most issuers assume. There is no crypto buyback safe harbour. MAR Article 5 covers equity only, so any issuer buyback defence rests entirely on the conduct not being manipulative on the specific facts. That puts the weight on the contemporaneous record: the stated rationale, the funding source, the venues and the sizes, and who held the information before the announcement and for how long. Our position is that this record gets assembled while the buyback runs, because a regulator reading it two years later is reading what was written at the time.

Not investment advice.

Does launch-day work end once the token is trading?

No. Supply and vesting are fixed by the time the token trades, which is why tokenomics design belongs pre-TGE, but market structure stays live after the first fill. At TGE the lever that remains is market making. After TGE the ongoing work is market-health telemetry.

That split matters for how a team staffs the weeks around a DEX launch. Any argument about emission curves, cliff dates or allocation sizes has to be settled while the contract and the vesting schedule can still change. Once the pool is seeded and the first unlock calendar is public, the questions that can still be answered are about quoting, venue coverage and how depth behaves when a seller arrives. Our read is that teams who treat launch day as the finish line end up with a design they cannot revise and no visibility into the market they just created.

Telemetry is the part most often skipped, because it has no deadline attached. Spread, depth at a given distance from mid, fill quality and where volume actually sits across venues are all things that drift week to week, and a team only knows whether its market maker is doing what the agreement says if those numbers arrive on a regular cadence. We have written separately about what a market maker dashboard should put in front of an issuer.

Founders can bring us in at any stage of that lifecycle, including after the token is already trading, and work already done with other firms can be carried into the engagement rather than redone. There is no requirement to start from scratch and no requirement to use us exclusively. A team that has a tokenomics model from one vendor, a market maker on a signed term sheet and an unlock calendar already published can hand all three over and have the telemetry layer built on top of what exists.

Not investment advice.