Skip to main content

This site is for educational purposes only. Nothing here constitutes financial advice.

Topic 85 of 189

Market Makers and Spreads

How crypto market makers (Wintermute, Jump, GSR, Cumberland) build inventory positions, manage risk, and price spreads — and why understanding MM economics explains so much about market structure.

Beginner
8 min readUpdated July 2026Block Clarity Hub Editorial Team

The basics

What Market Makers Do

Market makers provide liquidity by quoting bid and ask prices simultaneously. When someone wants to buy, the MM sells from inventory; when someone wants to sell, the MM buys into inventory. The spread between bid and ask is the MM's profit margin. They're essentially professional resellers of liquidity. In crypto, major MMs include Wintermute, Jump Trading, GSR, Cumberland, Amber Group, and many smaller firms. They handle a substantial fraction of total crypto trading volume across CEXs, DEXs, and OTC.

Why Spreads Are What They Are

MM spreads are calibrated to cover risks and costs. Inventory risk: holding crypto inventory exposes the MM to price moves. Adverse selection: some traders have information the MM doesn't — those trades cost the MM more. Cost of capital: capital tied up in inventory could earn yield elsewhere. Operational costs: infrastructure, salaries, fees. Wider spreads compensate for higher risks; tighter spreads reflect lower risks. BTC/USDC at major venues might trade with 1-2bps spread; obscure tokens might have 100-1000bps.

Why You Should Care

Spreads are a transparent cost you pay on every trade. A 50bps round-trip spread (25bps each side) means a $10K trade costs $50 just to enter and exit. Trading frequently against wide spreads is one of the largest drags on retail crypto returns. Understanding which assets and venues have tight vs wide spreads is essential for managing trading costs. Major-pair CEX spreads are tight; long-tail and DEX spreads can be substantially wider.

  • MMs quote bid/ask simultaneously, profiting from the spread
  • Major crypto MMs: Wintermute, Jump, GSR, Cumberland, Amber
  • Spread reflects inventory risk + adverse selection + cost of capital + operational costs
  • Spreads are a major retail-trading cost; understand which markets are cheap vs expensive

Key Takeaways

  • Market makers provide liquidity in exchange for spread
  • Spread size reflects multiple structural risks, not arbitrary
  • BTC/USDC at major venues has tight spreads; long-tail assets and DEXs can be wider
  • Spread is a transparent cost; trading frequently amplifies the drag

Going deeper

Adverse Selection and Toxic Flow

MMs lose money on 'informed' trades — trades by counterparties who know the price is about to move. This is 'toxic flow.' If a buyer knows good news is about to come out, they buy from the MM at the ask. After the news, the price rises, and the MM has effectively sold the asset to someone who knew better — locking in a loss. To avoid this, MMs widen spreads when they suspect informed flow. The mix of informed vs uninformed flow at any venue determines spread economics. Retail flow is mostly uninformed; institutional or sophisticated flow has higher informed-fraction.

Inventory Management

MMs aim to hold roughly neutral inventory — not directionally exposed. When they end up with too much of one asset (because clients sold more than bought), they hedge by selling on another venue or shorting futures. Net result: MMs are largely directionally agnostic. When MMs can't hedge cleanly (small venues, illiquid pairs), they end up holding inventory longer and need wider spreads to compensate. This is why long-tail or new-venue spreads tend to be wide initially.

MM Competition and Spread Compression

Where multiple MMs compete, spreads tighten. Each MM is willing to take less margin to win flow. BTC/USDC at major CEXs has many MMs competing; spreads are 1-2bps because no one can charge more without losing business. New tokens or new venues have fewer MMs; spreads start wide and compress as more MMs enter. Watching spread compression on new venues is a real-time signal of how mature the market is becoming.

DEX vs CEX Market-Making Economics

MMs operate on both CEXs (traditional orderbook market making) and DEXs (typically via aggregators and intent protocols rather than naive AMM LP). DEX market making is different because of MEV — sandwich attacks make traditional 'quote on an orderbook' MM less effective. MMs increasingly route to DEXs via intent-based protocols (CoW Swap, UniswapX) where they fill orders directly without exposing their quotes. Watching MM flow patterns reveals how DEX market structure is evolving.

Key Takeaways

  • Adverse selection (informed flow) is the biggest MM risk; toxic flow widens spreads
  • MMs aim for neutral inventory; can't hedge cleanly = wider spreads
  • Competition compresses spreads where it exists
  • DEX market making differs from CEX due to MEV; intent-based protocols are MMs' DEX path

Under the hood

Latency and Co-location

Top-tier MMs operate with sub-millisecond latency to exchange engines. Servers co-located in the same data centers as exchange matching engines. Custom networking, FPGA-based execution paths, optimised software stacks. The latency advantage means they can update quotes in response to new information faster than competitors. In tight-spread markets where the first MM to update wins the trade, latency is everything. Smaller MMs operate at higher latency tiers and accept lower trading volumes.

Risk Management Systems

MM operations need real-time risk management. Inventory limits across assets and venues. P&L caps that pull quotes if losses exceed thresholds. Correlation tracking across positions. Stress-testing against market scenarios. Hedging programs running continuously. A modern MM operation is essentially a hedge fund focused on liquidity provision rather than directional bets. Top MMs make billions in revenue annually; the operational sophistication required is substantial.

Privileged Market Maker Programs

Some exchanges offer 'designated' or 'privileged' MM programs with fee rebates, priority access, or other benefits in exchange for commitments (quoting minimum size at minimum spread for minimum percentage of time). These programs are common at major venues and are how exchanges incentivise tighter spreads. Smaller traders can't access these programs directly. Knowing whether a venue uses designated MMs vs general competition affects how reliable the quoted liquidity actually is.

MM and FTX — The Lesson

Many crypto MMs lost substantial capital in the FTX collapse — they had inventory and working capital on FTX that became unrecoverable. The lesson: MM operations have substantial venue counterparty risk. Modern MM operations carefully distribute capital across venues, hold minimum operational balance on any single CEX, and use reputable custody (Fireblocks, BitGo) when possible. The FTX collapse fundamentally reshaped MM risk management practices and is part of why several MMs shifted toward on-chain trading where custody is self-managed.

Understanding MM Behavior Predicts Market Behavior

Many market events are explainable as MM responses to changing conditions. Spreads widening before a major event = MMs pricing in elevated risk. Spread compression after consolidation = MMs settling into a stable view. Quotes pulled during volatility = MMs reducing exposure. Reading MM behavior gives a much richer view of market structure than just price action. Several Twitter/X accounts (e.g., @JonCryptoBee) provide MM-flow analysis worth following.

Key Takeaways

  • Top MMs have sub-millisecond latency advantages via co-location
  • MM operations are sophisticated risk-management businesses, not directional trading
  • Designated MM programs at major venues drive tight spreads via commitment
  • FTX collapse reshaped MM risk practices; custody and venue concentration are now actively managed
  • Reading MM behavior (spread changes, quote pulls) gives richer market-structure insight than price alone

Related Content

Portfolio Rebalancing Systematics

How systematic rebalancing keeps portfolio weights at target allocations, the math of why rebalancing produces an excess return in volatile markets, and on-chain rebalancing protocols.

Prediction Market Mechanics

How Polymarket, Augur, and Kalshi structure event-betting markets, what makes a 'good' prediction market, and the structural questions around accuracy, manipulation, and information aggregation.

Basis Trade Deep Dive

How the cash-and-carry basis trade works mechanically across CME, ETFs, perpetuals, and spot — and why basis trading accounted for part of 2024-2025 ETF inflows.

Options vs Perpetuals — Picking the Tool

When to use options vs perpetuals for expressing the same trade view, what each product's structural strengths are, and the framework for product selection across different strategies.

Token Velocity

Why MV=PQ predicts that high-velocity tokens (used and immediately dumped) capture less value than low-velocity tokens (held), and what this means for token design.

Real Yield

How protocols generate sustainable yield from real fees (vs Ponzi-like token emissions), why this category emerged post-2022, and the protocols leading it (GMX, Synthetix, dYdX).

Vote-Escrowed Tokenomics (ve)

How Curve's veCRV model pioneered lock-and-earn governance with vote weighting, why ve has been so widely copied (Convex, Velodrome, others), and the gauge wars it enabled.

Bonding Curves

How bonding curves automate token pricing without orderbooks, the math behind linear/exponential/sigmoid curves, and the pump.fun-style use cases driving memecoin distribution.

References & further reading