Crypto Market Making Explained: How It Works, What It Costs, and How to Choose a Provider

A market maker is a firm that continuously quotes both a bid and an ask on your token, on the venues where it trades, so that anyone who wants to buy or sell can do so immediately at a predictable price. Without one, an order book is a queue of hopeful limit orders with air in between. A buyer who wants size has to walk that air, and the resulting price jump is what traders call slippage. Slippage is the tax your holders pay for having no liquidity provider.

The market maker is not there to push your price up. Its job is narrow and mechanical: keep the spread tight, keep depth present on both sides, and stay online through volatility. It earns from the spread between its own bid and ask, from exchange maker rebates, and from whatever fee or option structure you have agreed. Any provider whose pitch centres on price appreciation rather than book quality is describing a different, and usually non-compliant, activity.

The provider connects to each venue via API, prices your token off a reference — usually the deepest venue or a composite index — and posts layered orders around that mid-price. As trades hit those orders the firm's inventory drifts long or short, so it re-hedges, often on a correlated perpetual or on another venue where your token trades. The quoting engine adjusts spread width and size dynamically as volatility rises, which is why a thin book widens during a market flush rather than disappearing entirely.

Performance is measured, not asserted. The standard metrics are time-weighted spread, depth at defined bands such as ±0.5%, ±1%, and ±2% from mid, two-sided uptime as a percentage of the measurement window, and the share of maker volume the firm contributes. A serious agreement writes these into a KPI schedule with a reporting cadence. If you cannot see daily numbers, you cannot tell whether you are paying for liquidity or for a screenshot.

Three forces converge. Exchanges require credible liquidity as a listing condition because a dead book damages their own users' experience. Treasuries and larger holders will not build a position they cannot exit. And retail sentiment is shaped almost entirely by chart quality — a book that gaps 8% on a modest sell order reads as abandoned regardless of what the project ships.

Founders often discover this in the wrong order: they list first, watch the first week of erratic candles, then scramble for a provider from a position of weakness. Arranging liquidity before the listing application is submitted is both cheaper and more credible, because you are negotiating with time on your side rather than with a visibly broken order book as your only leverage.

Under the loan model, you lend the provider a tranche of tokens plus a stablecoin counterpart as working inventory, and grant call options over that tranche at strikes above the reference price. The provider quotes with your inventory and profits if the token appreciates past the strikes. Cash cost is low; the risk is that the option structure and the loan size are misaligned with your float, which can produce persistent sell pressure at exactly the strikes you set.

Under the retainer model you pay a monthly fee in stablecoins and the provider deploys its own capital. There are no options, no borrowed float, and the incentive is simple — meet the KPIs or lose the mandate. Retainers cost more in cash but keep your cap table clean. Hybrids exist: a smaller loan with a reduced retainer, or a retainer plus a performance component tied to measured depth and uptime rather than to price. We break the arithmetic down in the pricing and structure guides linked below.

On a centralised exchange, liquidity is an order book maintained by firms posting and cancelling quotes. On a decentralised exchange it is a pool of paired assets governed by a curve, where anyone providing liquidity is passively quoting at all prices in range. The two behave differently under stress: an order book can be pulled, a pool cannot, which is why pool providers carry divergence loss while order-book makers carry inventory risk.

Most tokens need both. A DEX pool gives permissionless access and a price reference from day one; a CEX book gives depth, size, and the fee tiers that professional flow requires. What matters is that the two are arbitraged properly, otherwise your token trades at two prices and every quote source publishes a different number.

Providing two-sided quotes is a recognised, legitimate market function performed on regulated venues worldwide. What is not legitimate is generating trades between accounts you control to inflate reported volume, layering orders you intend to cancel, or coordinating quotes to move a price to a target. Those behaviours are wash trading and spoofing, and regulators from the SEC to VARA, MAS, and the FCA treat them as market abuse regardless of whether a token is a security in that jurisdiction.

The practical test for a founder is documentary. Does the agreement define quoting obligations rather than volume targets? Does the provider report depth and spread rather than notional traded? Will it accept a clause prohibiting self-matching? A provider that hesitates on any of those is selling you a metric, not a market.

Start with venue coverage: the firm should already be an approved maker on the exchanges you care about, with its own fee tier. Then examine capital — a provider quoting only with your loaned inventory has no skin in the game. Then references, specifically from projects of similar market cap rather than the two blue-chip logos on the deck. Then the contract: notice period, KPI definitions, reporting, termination, and what happens to loaned tokens on exit.

Finally, look at how they answer hard questions. A good desk will tell you what it cannot do, will push back on unrealistic spread targets for a low-float token, and will decline a mandate where the free float cannot support the depth you want. That refusal is the strongest positive signal in this market.

Six to eight weeks before listing, define your liquidity budget and free float. Four to six weeks out, shortlist providers and request indicative terms on the same specification so the quotes are comparable. Three weeks out, sign, fund, and complete venue onboarding and API keys. One week out, run the book in a test configuration on your DEX pool. On listing day, the provider is already quoting when the pair opens rather than warming up in public.

Xavion Capital sits on the founder's side of that process. We do not run a book. We structure the mandate, benchmark the terms, and hold the provider to the KPI schedule — which means the numbers you are shown each month are the numbers we asked for.

Take a token launching with a $2m stablecoin-equivalent circulating float and a target of 1% total supply reserved for liquidity. Under a retainer, a two-venue mandate with ±1% depth of $150k per side and a maximum 40bps spread commonly prices in the low tens of thousands of dollars per month, scaling with venue count and how demanding the uptime target is. Under a loan, the issuer might instead post 1% of supply plus a matching stablecoin tranche, granting call options at strikes 50%, 100%, and 200% above the reference price over a twelve-month term.

Model both to a common basis. Twelve months of retainer at a representative rate is a known, bounded cash number. The loan's cost is contingent: at a flat price it costs almost nothing beyond the opportunity cost of locked float; if the token triples, the provider captures meaningful upside through the exercised strikes, and that value should be compared against the retainer total, not against zero.

The comparison usually surprises founders in one of two directions. Early-stage teams with thin treasuries often find the loan is the only structure they can afford, and rightly accept the contingent cost as the price of getting listed at all. Teams with investor capital and a longer runway frequently discover the retainer is cheaper in every plausible price scenario once the option value is properly modelled, and switch models before the loan term begins.

The most frequent failure is a provider quoting to the letter of a KPI while missing its spirit — for example holding depth exactly at the measurement instant each hour while letting the book thin out between checks. This is why serious agreements specify continuous or high-frequency sampling rather than point-in-time snapshots, and why founders should ask a candidate provider how its own monitoring works before signing.

A second failure mode is inventory mismatch: a provider under-capitalised relative to the depth it has promised will widen sharply in the first real volatility event, precisely when holders are watching most closely. This is difficult to see from a term sheet alone, which is why reference calls with projects of a similar market cap matter more than the size of the provider's headline client list.

A third is silent scope creep in reverse — a provider that quoted two venues at signing and quietly reduces effective coverage to one once volume disappoints, without renegotiating the fee. Monthly reporting that names each venue's measured spread, depth, and uptime individually, rather than a blended average, is the simplest defence against this.