The Problem with CLOBs (6 minute read)

Prediction markets hit $6.5B weekly volume but their central limit order book architecture concentrates liquidity with professional market makers, leaving long-tail markets illiquid and locking out passive DeFi capital.

What: A critique of how Polymarket and Kalshi use central limit order books (CLOBs) that require professional market makers to provide liquidity, which works for high-profile political and sports markets but leaves thousands of niche markets in entertainment, science, and culture without support.

Why it matters: The architecture explains why 85-90% of prediction market volume remains concentrated in politics and sports despite apparent demand for diverse markets, and why $550M in prediction market capital can't tap into $100B of passive DeFi liquidity sitting in lending and yield protocols.

Deep dive

Polymarket switched from AMM to CLOB in late 2022 after liquidity providers lost money on every resolved market; Kalshi launched with order books from day one
Combined weekly volume between platforms reached $6.5B in April 2026, with Bernstein projecting the industry hits $1T by 2030
Kalshi has only 23 active market makers, with the top 3 providing 70% of liquidity in election contracts
CLOBs require active professional management to quote spreads and respond to information in real-time, eliminating passive participation entirely
Market makers show up for presidential elections and major sports but ignore thousands of potential long-tail markets
The architecture is structurally incapable of serving permissionless markets because by the time a market maker evaluates whether to provide liquidity for a real-time cultural event, the moment has passed
Both platforms curate which markets get listed specifically because they know markets without market maker support are dead on arrival
Over $100B in DeFi capital sits in lending pools and yield vaults with no architectural pathway to prediction markets, while prediction market TVL is only $550M
The industry consolidating into a duopoly where professional intermediaries decide which questions get liquid markets contradicts the original promise of democratizing forecasting
The article argues for AMM-style infrastructure where the first participant bootstraps liquidity for the second, removing the professional market maker gatekeeper

Decoder

CLOB (Central Limit Order Book): Traditional exchange architecture where buyers post bids and sellers post asks, with trades executing when prices match
AMM (Automated Market Maker): Protocol that uses liquidity pools and algorithms to enable passive users to provide liquidity without active management
Market maker: Professional firm that continuously quotes buy and sell prices to provide liquidity, profiting from the spread
TVL (Total Value Locked): The total amount of capital deposited in a protocol or platform
Liquidity: The availability of assets to trade without significantly moving the price; markets with good liquidity have tight spreads and deep order books

Original article

Prediction markets reached $6.5B in combined weekly volume across Polymarket and Kalshi in April, but the CLOB architecture concentrating that growth also limits it: Kalshi's top 3 market makers supply 70% of election contract liquidity, leaving thousands of long-tail markets in entertainment, science, and culture without support. The result is 85-90% of prediction market volume locked to politics and sports, while ~$550M in total TVL sits disconnected from the $100B in DeFi capital deployed in lending and yield protocols. AMM-style permissionless infrastructure where the first participant bootstraps liquidity for the second removes the professional market maker requirement that currently gates new market creation.