What Is a Bonding Curve? How Pump.fun Token Pricing Works

Every token that launches on Pump.fun starts its life on a bonding curve, and not understanding that one mechanic is exactly why people overpay at the top and panic-sell at the bottom. A bonding curve is a piece of code that sets a token's price automatically based on how much of it has been bought — no order book, no market maker, no one on the other side of your trade. The price isn't negotiated; it's calculated. Once you can read the curve, a lot of memecoin behavior that looks random suddenly makes sense. Here is how it actually works.

What a bonding curve actually is

A bonding curve is a smart contract that defines price as a deterministic function of supply. Put plainly: the contract holds a formula that says "given how many tokens have already been sold, the next one costs this much." Buy more, and the price the contract quotes goes up. Sell back, and it goes down. The relationship is fixed in code before the token ever trades.

The important part is what this replaces. On a normal exchange, a price exists because a buyer and a seller agree on one — there's an order book matching the two sides, and someone has to seed liquidity before trading can start. A bonding curve needs none of that. There's no counterparty and no pre-funded liquidity pool: you trade directly against the contract, and the curve itself plays the role of market maker. That's why a brand-new Pump.fun token is tradable the instant it's created, with zero starting capital. The curve is the market. If you want the platform-level picture of where these tokens come from, start with what Pump.fun is and how it works.

The math, made simple

Pump.fun's curve uses a constant-product style of pricing — the same family of math behind Uniswap-style AMMs, expressed as x · y = k. Conceptually, the contract tracks two reserves: a supply of tokens and a balance of SOL. Their product is held at a constant. When you buy, SOL goes in and tokens come out, which shifts the ratio between the two reserves — and that ratio is the price. Because the product has to stay constant, every buy makes the remaining tokens scarcer and pushes the next quote higher.

You don't need the exact contract constants to grasp the consequence, and I won't pretend to quote them precisely — the behavior is what matters. Think of it as a one-way ratchet on the way up: each purchase moves you a step further along the curve, and each step is priced higher than the last. Selling does the reverse — tokens go back to the contract, SOL comes out, and the quoted price slides back down the same curve. There is no separate "sell side" to find; the math handles both directions automatically.

The single takeaway that pays for itself: early buyers pay less and later buyers pay more, by construction. Not because of hype or momentum, but because the curve mechanically charges more for each token as supply gets bought up. The person who buys at 10% of the curve filled is paying a fraction of what the person buying at 90% pays for the identical token.

Why early buyers pay less — and the snipe dynamic

A bonding curve is steepest in relative terms right at the start, where the smallest buys produce the largest percentage moves. That's the structural reason early entries are so coveted: get in near the bottom of the curve and every later buyer is, in effect, marking up your position for you. It's also the structural reason the start of a launch turns into a fight.

This is where sniping comes in. Bots compete to land their buy in the very first block — block 0 — so they sit at the absolute bottom of the curve before anyone else. When that happens, the curve can jump several steps in a single block, which means every human who buys a few seconds later is already paying a snipers-inflated price. The curve doesn't care who you are; it just quotes the next price. So the snipe dynamic isn't a side issue — it's a real cost borne by everyone who isn't first. If a chart shows a token that opened and immediately spiked before you could click, that's the curve being eaten from the bottom by faster participants. Knowing that is half of not getting wrecked by it.

Graduation — when the curve fills

A bonding curve has a finite amount of token to sell. When buyers have worked all the way up it and the curve is fully bought out, the token graduates: the liquidity that accumulated on the curve gets migrated into a real automated market maker (AMM) pool, and trading continues there instead.

The long-standing threshold is roughly a $69,000 market cap, which corresponds to approximately 85 SOL raised on the curve. Treat the dollar figure as approximate, not a hard constant — because the curve is denominated in SOL, the USD market cap at graduation drifts with SOL's price, which is why you'll see slightly different numbers quoted across sources. The SOL amount is the more stable anchor.

What's changed is where the token lands. In the early days, graduated Pump.fun tokens migrated to Raydium, and that migration was a clunky, paid event — a separate pool had to be created, it carried a fee of around 6 SOL, and trading paused during the hand-off. Since March 2025, Pump.fun runs its own DEX, and graduated tokens now migrate to PumpSwap instead — instantly, with the bonded SOL seeding the new pool directly and no separate migration fee. After graduation the token behaves like any normal Solana token: it can be quoted and routed by aggregators, picked up by scanners, and traded through bots. The deterministic curve is gone; from here, price is set by ordinary pool supply and demand.

Bonding curves beyond Pump.fun

Pump.fun popularized the model, but it isn't the only launchpad built on a bonding curve. The clearest contrast is LetsBonk, which runs on Raydium's LaunchLab. The shape of the experience is the same — buyers mint on a bonding curve, and once it fills (LaunchLab's default threshold is also around 85 SOL) the liquidity auto-migrates to an AMM and the LP tokens are burned. The difference is the destination and the tuning: LaunchLab graduates tokens into Raydium's own deep CPMM pools rather than a Pump.fun-native venue, and it exposes more knobs — curve shape, supply, vesting, fee structure — for creators who want to customize the launch instead of accepting one fixed default.

The lesson is that "bonding curve" describes a pricing mechanism, not a single product. Different launchpads pick different curve parameters and different graduation destinations, and those choices change the trading dynamics. If you're weighing where a token came from, the launchpads-compared breakdown lines the major ones up side by side.

The risks the curve hides

The same mechanic that lifts a token on the way up drops it just as fast on the way out. A bonding curve is symmetric: if buys ratchet the price up step by step, sells slide it down the exact same steps. When sentiment turns and holders start exiting, the curve doesn't cushion the fall — it accelerates it, because every sell makes the next quote lower. The wall of green you watched on the way up becomes a wall of red in reverse.

Two things people routinely get wrong are worth stating bluntly. Graduation is not safety. Hitting the threshold and migrating to PumpSwap or Raydium means the token cleared a liquidity bar — nothing more. It is not a quality check, an endorsement, or proof the team is honest. Plenty of tokens graduate and then dump immediately, because the very people who pushed it to graduation are the ones with the largest, cheapest positions to unload. And the curve does not stop rugs. Post-migration, the usual risks are all still on the table — malicious mint or freeze authority, insider supply, liquidity that can be pulled, coordinated dumps. The curve governs price discovery; it says nothing about whether the token is a scam. Run the due-diligence checklist before you buy, on the curve or after it — and learn to read the pool itself with DexScreener so a thin or fake market doesn't catch you out.

How MoonHydra fits

MoonHydra doesn't change how the curve works — it just gives you a clean way to act on it. You can buy and sell a token whether it's still on its bonding curve or already graduated to PumpSwap or Raydium: you paste a contract address and the bot routes the order through Jupiter, which finds the liquidity for you regardless of which stage the token is in. There's no venue to pick and no migration step to babysit. The only fee MoonHydra adds is a flat 1% per trade — buy and sell — on top of the network and DEX fees, with no subscription. Keys are encrypted with AES-256-GCM and the bot is non-custodial, so you hold your own funds, and you can run separate Hydra Head wallets to keep curve-stage entries apart from longer-term holds. None of that is a substitute for judgment: the curve tells you what a token costs, not whether it's worth buying. Pair the tooling with real due diligence before you ape.

Bottom line

A bonding curve is a contract that prices a token by formula instead of by an order book — supply in, price out, deterministically. On Pump.fun it uses constant-product math, so early buyers pay less and later buyers pay more by construction, which is exactly why launches turn into a sniping race for the bottom of the curve. When the curve fills near ~$69k market cap (~85 SOL), the token graduates: liquidity migrates to an AMM — PumpSwap on Pump.fun since March 2025, Raydium in the old days and still the destination for launchpads like LetsBonk. The catch is that the curve runs in both directions, graduation is a liquidity event and not a safety stamp, and rugs still happen after migration. Read the curve, then verify the token.

Next: read what Pump.fun is for the full launchpad picture, what PumpSwap is for where graduated tokens trade, and run the due diligence checklist before any buy. Start trading at t.me/moonhydrabot.