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.
When a market making contract ends, quoting stops and the order book reverts to whatever natural interest exists, which for most growth-stage tokens is very little. Spreads widen within minutes, depth thins across the first day, and exchange monitoring systems begin recording the deterioration immediately. The commercial unwind runs in parallel: loaned inventory and capital are returned on the notice schedule, any options survive or expire according to the contract, and exchange API keys have to be revoked and reissued. A transition planned in advance takes days and is largely invisible on the chart; one that is not takes weeks and is visible to everyone, including listing teams.
Founders spend considerable effort on the start of a market making relationship and almost none on the end of it. That is understandable, and it is also where a disproportionate share of the damage occurs, because the termination clause is usually read for the first time at the moment it is needed and by then its terms are fixed.
This is an account of what actually happens, arranged as a sequence rather than a list of clauses. It applies whether the engagement is ending because the term expired, because the project chose to change desks, or because the relationship broke down.
What happens to the order book
The first hour
Quoting is a continuous activity, and when it stops the effect is immediate rather than gradual. The resting orders that constituted your depth are cancelled, and the book falls back to whatever limit orders unaffiliated participants happen to have left. On a large-cap asset that residue is substantial. On a token with a market capitalisation in the tens of millions, it is frequently a handful of orders at prices nobody expects to trade at.
The visible consequences are a spread that widens by a multiple rather than a margin, and a book in which an ordinary retail sell moves the price several percent. Nothing about the project has changed; the mechanism that was translating its price into an orderly market has simply been switched off.
The first day
Within a session the secondary effects appear. Arbitrage between venues weakens because the spread no longer justifies the capital, so the token's price on the affected exchange starts drifting away from its price elsewhere. Volume falls, partly because trading has become expensive and partly because algorithmic participants that were interacting with the book withdraw when the counterparty disappears.
This is also the point at which exchange monitoring begins recording. Venue systems measure spread and depth continuously, and several of the published thresholds are defined over consecutive-day windows. MEXC, for example, treats an average daily spread above two percent sustained for fifteen consecutive days as grounds for a warning tag. A gap of a few days between desks is survivable; a gap of three weeks starts a clock that is difficult to stop. The full set of thresholds is covered in why tokens get delisted.
The first month
If no replacement is quoting, the deterioration compounds rather than stabilising. Wider spreads deter the discretionary buyers who would otherwise supply natural depth, which widens spreads further. Holders who try to exit at any size discover the price impact and post about it. The chart develops the characteristic pattern of low-volume gaps that is legible to anyone assessing the token, and by the end of a month the book usually looks materially worse than the underlying project's situation warrants.
This is why the question worth asking is not what happens when a contract ends but how long the interval is between one desk stopping and the next one starting. The answer should be zero, and it is achievable with roughly two weeks of planning.
What happens commercially
The unwind runs on the terms you signed, and the terms differ sharply by pricing model. Each of the three common structures ends differently, which is one of the reasons we compared them in retainer versus profit-sharing versus the loan model.
Under a retainer
This is the cleanest ending. The inventory and capital deployed for quoting were yours throughout, so termination is a matter of the desk ceasing to quote, returning control of the accounts, and providing a final reconciliation. What to check is the fee schedule for the final period, whether any notice-period retainer remains payable, and the exact date after which quoting stops, so that it can be matched to the date the successor desk starts.
Under a loan and options structure
This is where most disputes live. The tokens were lent to the desk, and the return schedule is defined by the contract rather than by your preference. Options granted as part of the arrangement may survive termination, which means the desk can continue to hold a call on your supply after it has stopped providing any service. Whether that is acceptable is a question to settle when signing rather than when leaving.
The specific items to establish are the return date for loaned tokens, whether return is in kind or in value, what happens if the desk is short at termination, whether options survive and for how long, and what disclosure you receive about the desk's position at the point of handover. A desk exercising options into a book that no longer has a market maker is a scenario worth modelling before it occurs.
Under profit sharing
The main questions are the final accounting period, how open positions are valued at termination, and who bears the cost of unwinding them. Ambiguity here tends to surface as a disagreement about a number that neither party can independently verify, which is an argument to specify the valuation methodology in the contract.
The operational handover
Four items account for most of the friction, and all four are administrative rather than commercial.
API keys. The outgoing desk held keys to your exchange accounts, or was quoting through its own accounts on your behalf. Keys must be revoked at a defined moment and new ones issued to the successor, and the two events should be minutes apart rather than days. Establish before termination who holds which keys and who has authority to revoke them.
Inventory location. Tokens and quote currency sitting in sub-accounts need to move, and exchange withdrawal processes are slower than most timelines assume. Start this before the final quoting day rather than after it.
Exchange relationships. If the outgoing desk was your named market maker with a venue, the exchange needs to be told about the change, ideally by you rather than by rumour. Listing teams notice when a book's behaviour changes character and prefer to be informed in advance.
The historical record. Request the full performance history before the relationship ends, not after. Spread, depth, uptime and inventory data become considerably harder to obtain once the commercial relationship has closed, and the successor desk will want a baseline.
How to run a transition with no gap
The mechanics are straightforward when sequenced properly. Select and contract the successor before serving notice on the incumbent, so that the notice period doubles as the successor's onboarding period. Set the successor's start time to the incumbent's stop time on the same day, with a short overlap where both are quoting if the venue and the contracts permit it. Move inventory during the notice period rather than after it. Revoke keys at the switchover moment. Tell the exchange in advance.
Run this way, a change of desk is invisible on the chart, which is the correct outcome. The token's market should not carry a visible record of your procurement decisions.
One caution about overlap: two desks quoting the same pair without coordination can trade against each other, which produces volume that looks organic and is not. If an overlap is used it should be brief, explicitly agreed, and structured so that responsibilities are separated by venue or by time rather than shared.
What to put in the agreement before you need it
Everything above is easier to manage when it was written down at the start. The clauses that matter are the notice period on both sides, the return schedule for loaned tokens and capital with dates rather than descriptions, whether options survive termination, who holds and who can revoke API keys, the obligation to provide full historical performance data on request, a cooperation requirement during handover to a successor desk, and confidentiality terms that do not prevent you from briefing the incoming desk properly.
A desk with clean exit mechanics is signalling that it does not expect you to want to use them. One that resists specificity here is telling you something about how the ending will go, and it is worth listening. Our full list is in what belongs in a market making agreement, and the diligence questions that surface these issues before signing are in crypto market maker red flags.
FAQ
What happens to a token's price when a market making contract ends?
Price does not move mechanically at termination, but the conditions around it change immediately. Spreads widen, depth falls to whatever unaffiliated participants have left resting, and ordinary trades produce much larger price moves. The resulting volatility is a liquidity effect rather than a change in valuation.
How quickly does the order book deteriorate?
Within minutes for spread, within a session for depth and cross-venue price alignment, and over subsequent weeks for volume and holder confidence. The compounding is the dangerous part: worse execution deters the natural buyers whose orders would otherwise have partially replaced the desk's.
Can a project switch market makers without a gap in coverage?
Yes, and it should. Contract the successor before serving notice, use the notice period for onboarding, and set the handover so that the new desk begins quoting as the incumbent stops. Planned this way the transition is invisible on the chart.
What happens to loaned tokens when the contract ends?
They are returned according to the schedule in the agreement, which may be immediate or staged over weeks. The points to confirm are whether return is in kind or in value, what happens if the desk holds a short position at termination, and what disclosure you receive about its position during the handover.
Do a market maker's options survive termination?
That depends entirely on the contract, and both arrangements exist. Options that survive mean the desk continues to hold a call on your supply after it has stopped providing service, which is a materially different situation from one where they expire with the engagement. Establish this before signing.
Who controls the exchange API keys after termination?
The project should, and the agreement should say so explicitly along with who is authorised to revoke them and when. Revocation and reissue to the successor desk should be separated by minutes, which requires the successor to be contracted and onboarded beforehand.
Can ending a market making contract lead to delisting?
Indirectly, yes. Exchanges measure spread and depth continuously, and several published thresholds are defined over consecutive-day windows. An extended period without quoting can breach those windows and start a review process, which is why the interval between desks matters more than the change itself.
How much notice is normal?
Thirty to ninety days is common, and the figure matters less than what it is used for. Treat the notice period as the successor's onboarding window rather than as a waiting period, and the length becomes an operational convenience rather than a risk.
Should two market makers overlap during a handover?
A brief, explicitly coordinated overlap can be useful, with responsibilities separated by venue or by time. An uncoordinated overlap is worse than a short gap, because two desks quoting the same pair can trade against each other and generate activity that resembles manufactured volume.
