How much initial liquidity does a token launch need?

“How much of the supply goes to liquidity?” is the wrong first question. Liquidity isn’t measured in percent of supply — it’s measured in dollars of depth against the trade sizes you expect. Get the depth right and a 2% liquidity allocation is plenty. Get it wrong and 10% still won’t stop the first $20k sell from tanking your chart.

Liquidity is depth, not a slice of the pie

A constant-product pool (Uniswap v2 and its forks) holds two reserves — your token and a paired asset — and prices trades against them. The price impact of a trade is, to a close approximation, the trade size divided by the reserve on the side being spent. That gives you a rule you can actually plan around:

This is why “% of supply” is a trap. 5% of a 1B supply at a $30M FDV is $1.5M of tokens — but if you only pair it with $150k of ETH, your real depth is $150k, and a single $10k order moves the price ~7%. The token side of the pool is almost irrelevant; the paired-asset side sets the depth.

Choosing the paired asset

The pairing decides who can trade you cleanly and how your chart behaves when the market moves:

The hidden cost of a volatile pair is impermanent loss: if your token 3×s against ETH, the pool rebalances and you end up holding more ETH and less token than if you’d just held. For a treasury seeding its own liquidity, that’s real money. A stablecoin pair removes IL on the quote side but forces you to part with actual dollars you might rather keep in the treasury.

How much to actually seed

Work backwards from the largest trade you want to absorb without a visible candle. For most launches that’s the first wave of airdrop and TGE sellers, plus a marquee buyer or two:

Notice the constraint is the dollars on the paired side, which come out of your raise or treasury — not the token percentage. You can express that as 2% or 8% of supply depending on FDV; the percentage is an output, not the input.

The token side of an LP is free — you minted it. The paired-asset side is real money, and it’s the only side that sets your depth.

Locking it — and why the lock is a trust signal

Seeding a pool and keeping the LP tokens in a team wallet means you can pull the liquidity at any moment — the literal definition of a rug. Locking the LP position (or burning it) is how you prove you can’t. The market reads an unlocked LP as a red flag within minutes of launch.

• Lock providers: Team Finance, UNCX, and Unicrypt are the standard third-party lockers; on Solana, Streamflow. They hold the LP tokens in a time-locked contract with a public, verifiable unlock date.
• Duration: 12 months is the floor that reads as serious; 6 months or less looks like a soft rug waiting to happen. Many credible launches lock for 24+ or burn outright.
• Burn vs lock: burning LP tokens is maximally trustless but permanent — you can never migrate the liquidity to a better venue or a v3 range. A long lock keeps that option open.

v2 full-range vs v3 concentrated liquidity

A Uniswap-v2-style pool spreads your liquidity across every price from zero to infinity. Most of it sits at prices that will never trade, so your effective depth at the current price is a fraction of the dollars you put in. A v3 concentrated position lets you place that same capital in a tight band around the launch price — often 10–50× more effective depth for the same dollars.

The catch is management: a concentrated range that the price exits stops earning and goes fully one-sided. For a hands-off treasury, a wide v3 range or a v2 pool is safer; if you have someone actively managing the position (or you’re on a chain like Base where Aerodrome handles ranges via gauges), concentrated liquidity stretches a small treasury much further.

Two real shapes

• GMX — ~15% to liquidity. As a DeFi protocol launching without a major CEX listing, it had to be its own market maker, so a large self-seeded position was the right call.
• Uniswap, Arbitrum — ~0% to liquidity. Both launched straight onto every major CEX with market-maker support, so seeding a DEX pool from the treasury would have been redundant. This is the exception that proves the rule: liquidity strategy is a function of distribution, not a fixed percentage.

The defensible answer

If you’re a normal launch without day-one CEX access: seed a pool deep enough that your expected first-week sell wave stays under ~3% slippage — for most teams that’s $0.5M–$2M of paired-asset depth — pair against your chain’s native asset (add a stablecoin pool once volume justifies it), and lock the LP for at least 12 months with a public unlock date. Express it as whatever % of supply that works out to; the dollars are the real decision.