Pump Fun Crypto Explained: What It Is, Fees & the PUMP Token (2026)

Search "pump fun crypto" and Google hands you two completely different things on the same page: the pump.fun launchpad, where anyone can mint a Solana memecoin in seconds, and the PUMP coin, the platform's own token with a live price chart next to it. Most explainers pick one and ignore the other. This guide covers both, because the fees you pay, the token you hold, and the odds you face depend entirely on which one you actually mean.

Direct answer: pump.fun is a Solana memecoin launchpad where tokens are created for free, trade against a bonding curve with a 1.25% fee, and automatically migrate to PumpSwap once the curve fills; PUMP is the platform's native token, backed by a policy of spending half of protocol fees on buybacks and burns. Everything below is the September 2026 version of that answer, with the platform's own fee schedule, the changes that landed this year (Mayhem Mode, USDC curves, creator fee sharing), and on-chain data on who actually profits.

If you have seen the screenshots (0.1 SOL to 50 SOL in twenty minutes) you have also seen the other screenshots. Both are real. The point of this article is to make sure you can tell which one you are more likely to produce, and what the tools that serious traders use, from pump fun bots to copy trading platforms, actually change about those odds.

The keyword itself is ambiguous, and the ambiguity costs people money. When someone says "pump fun crypto" they mean one of these:

• Pump.fun, the platform. A launchpad on Solana (pump.fun) where tokens are created, priced on a bonding curve, and graduated to a DEX. You never "buy pump.fun" here; you buy the individual memecoins listed on it, and every one of them is a separate, mostly worthless, occasionally life-changing asset.
• PUMP, the coin. The platform's native token, launched in July 2025 through a public sale that pump.fun says sold out in about 12 minutes. It trades on centralized exchanges and Solana DEXs like any other token. It is not on a bonding curve, and buying it is a bet on the platform's fee revenue and buyback policy, not on any single memecoin.

The two are related the way a casino stock is related to a slot machine: the coin earns from the platform's activity, the memecoins are the activity. Confusing them leads to the classic mistake of buying a random token called "PUMP" on the bonding curve because it looked like the official one. The official coin is the one linked from pump.fun/pump-token; anything else with that ticker is a memecoin that borrowed the name.

We cover the platform first (sections 2 to 6), then the coin (section 7), then the question everyone actually has: whether people make money here (section 8, with real numbers).

Pump.fun is a Solana-based memecoin launchpad that lets anyone create and trade a new token in under a minute. There is no coding, no audit, no liquidity to seed and no presale: a name, a ticker, an image, and the token is live and buyable. The platform handles token creation, the pricing mechanism, the trading interface and, once a token gets big enough, the migration to its own DEX.

The site launched in January 2024 and became one of the most-used applications in all of crypto. It drove a large share of Solana's on-chain activity during the 2024-2025 memecoin cycle, spawned an entire category of competing Solana launchpads, and by September 2026 is still generating protocol fees at a rate DefiLlama measured at roughly $46 million over a trailing 30 days.

Three design choices make pump.fun different from every launchpad that came before it:

• Fair launch by construction. No presale, no team allocation, no vesting. Every coin starts with its full tradable supply on the bonding curve, and the only way to get any is to buy it there. That does not stop a creator from buying first with a dozen wallets (bundling), but it does mean the contract itself gives nobody a head start.
• Bonding-curve pricing. Instead of a liquidity pool that someone has to fund, pump.fun uses a mathematical curve that raises the price with every buy and lowers it with every sell. The next section explains it properly.
• Automatic graduation. When a coin's market cap on the curve hits the graduation threshold, the curve closes and the entire pool migrates atomically to PumpSwap, pump.fun's own DEX. No human step, no liquidity to add, no way to reverse it.

Since the original version of this guide, pump.fun has also added a mobile app, a trading terminal, a creator dashboard for claiming fees, and, as of 2026, coins priced in USDC and other assets instead of SOL. We go through the changes that matter for traders in the "what changed in 2026" section below.

This is the part that confuses new traders, and getting it wrong costs money. The simple version: the bonding curve is a smart contract that is both the buyer and the seller of every token, with a price that rises mathematically as tokens leave the curve and falls as they come back. Developers read and trade that contract directly through the pump fun API and SDKs.

Think of it as a vending machine that reprices itself after every transaction:

• The first SOL in buys a lot of tokens (the price is tiny).
• The next SOL buys fewer tokens (the price has moved up).
• The hundredth SOL buys very few tokens (the price has moved a lot).

Under the hood, pump.fun's own documentation describes the curve as a constant-product market maker "similar to the math used by Uniswap", with virtual reserves of SOL and coin supply so that the curve has a defined starting price before anyone has deposited anything. Two consequences follow directly from that design, and the docs state both plainly: every buy moves the price up and every sell moves it down, and price impact scales with trade size, so a large order gets a worse fill than a small one.

Why this matters for traders:

• Being early matters more here than anywhere else in crypto. The price difference between the first block and the tenth block of a hot launch can be several multiples. That is the entire reason pump fun sniper bots exist.
• Selling hurts everyone. The curve is symmetric. When a large holder dumps, the curve pays them progressively less SOL per token, and the price for everyone still holding collapses.
• "Slippage" is built into the price, not added to it. On a bonding curve, most of what you experience as slippage is the curve repricing as your own order executes. A 0.5 SOL buy on a fresh launch literally creates a higher price for the next buyer.
• Bots now buy by budget, not by token amount. Integrator documentation notes that most curve swaps on mainnet now use a newer instruction (buy_exact_sol_in, "spend exactly this much SOL and take whatever tokens that buys"), alongside a unified set of v2 trade instructions pump.fun introduced so that SOL-quoted and USDC-quoted coins share one interface. Practically: the SOL you type is the SOL you spend, and your fill in tokens is whatever the curve gives you at that instant.

One more 2026 wrinkle: since May 2026 a creator can launch a coin whose curve is priced in USDC rather than SOL, and pump.fun has since extended that to "custom pairs" such as tokenized stocks, wrapped BTC and metals. The curve math is identical; what changes is that the coin's market cap no longer moves with the SOL price. Trading fees are the same 1.25% either way.

A pump.fun coin does not live on the bonding curve forever. Once its market cap on the curve reaches the graduation threshold (historically around $69,000, or roughly 85 SOL deposited, though the exact SOL figure moves with the SOL price), the contract automatically:

1. Closes the curve, so no further curve trades are possible.
2. Migrates the SOL (or USDC) and the remaining tokens atomically into a liquidity pool on PumpSwap, pump.fun's own automated market maker.
3. Charges a 0.015 SOL graduation fee, per pump.fun's fee schedule.
4. Leaves the pool owned by the protocol. Pump.fun's docs state it does not seed or remove that liquidity after graduation, and that graduation is "automatic and irreversible".

This is called migration or graduation, and it is the single most important event in a token's life. Before PumpSwap launched in March 2025, graduates went to Raydium and pump.fun paid a migration fee of several SOL per token; that path is gone, and the migration is now instant with no window where the token is untradeable. If you want the full picture of what happens after the curve, our PumpSwap deep-dive covers the AMM, its fees and how it compares with Raydium.

Graduation is rare. Pump.fun's own "under 2%" framing matches what public Dune dashboards have shown for years: well over 98% of coins die on the curve without ever reaching a DEX. The ones that do graduate often see a burst of buying at the moment of migration, because a whole class of traders and bots refuses to touch bonding-curve tokens and only enters once a coin is on a regular DEX. Tools like Axiom and GMGN have "about to migrate" filters for exactly this reason, and on uwuu.ai you will find wallets whose whole strategy is migration plays.

One thing the on-chain data says clearly, though: the edge is on the curve, not after it. In our study of 1,710 real copied positions, 77% of all copy trades executed on the pump.fun bonding curve, and closed bonding-curve positions had an 84.9% win rate, against 22.7% on Raydium and 33.3% on PumpSwap. By the time a coin has graduated, the asymmetric entry is usually gone.

People search "pump fun fees" expecting one number, and there is no one number. Here is the schedule pump.fun publishes on its own fee page (last updated May 20, 2026), plus the costs that sit on top of it:

Action	Fee (Sept 2026)	Who receives it
Create a coin	0 SOL / 0 USDC (network fee only; ~0.025 SOL if you buy at creation)	Solana validators
Bonding-curve trade (buy or sell)	1.25% total	0.95% protocol + 0.30% coin creator
Graduation to PumpSwap	0.015 SOL, once	Protocol
PumpSwap trade, canonical pool (a graduated coin)	1.25% at low market cap, stepping down to 0.30% at high market cap	0.20% LPs + 0.05% protocol + creator share (0.30% falling to 0.05% as market cap grows)
PumpSwap trade, non-canonical pool	0.30%	0.25% LPs + 0.05% protocol
Solana network + priority fee	Fractions of a cent base; priority tips can spike to 0.01-0.1+ SOL on hot launches	Validators / Jito

Some things worth noticing in that table:

• The "1% fee" everyone quotes is out of date. Bonding-curve trades cost 1.25%, and a coin's creator earns 0.30 points of it on every trade for as long as the coin trades. That creator cut is the incentive engine behind most of 2026's launch volume, and it is why a deployer can profit even if their coin never goes anywhere.
• Graduated coins are not automatically cheaper. A freshly graduated canonical pool still charges 1.25% until the market cap climbs into the higher tiers. The 0.30% floor applies to large coins and to non-canonical pools.
• Third-party tools stack on top. A browser terminal such as BullX or Axiom, or a Telegram sniper, charges its own 0.75-1% per trade in addition to pump.fun's 1.25%. Round trip through a bot on the curve, you are paying roughly 4-4.5% before priority fees, failed transactions and curve impact. Our Solana transaction fees guide shows how that arithmetic wrecks small accounts.
• Fees are paid win or lose. That is the real reason fee structure matters more than any feature list: an active trader doing 50 round trips a day at 4% is paying 200% of their position size in fees every day, and the curve does not care whether those trades were good. It is also why uwuu charges a performance-based fee: you pay only when a copied trade closes in profit.

If your mental model of pump.fun is from 2024 ("1% fee, 0.02 SOL to create, graduates to Raydium at $69k"), most of it is now wrong. Here is what changed, in the order that matters for a trader:

1. Creator fees became configurable and shareable (January 2026). Pump.fun's co-founder publicly conceded in January that the original creator-fee experiment had "failed to produce durable trading incentives", rewarding people for issuing coins rather than trading them. The overhaul that followed lets a creator (or the admins of a community takeover) route creator fees to up to 10 wallets with set percentages, transfer coin ownership and revoke update authority, and, on PumpSwap, ties the creator's share to the coin's market cap, stepping down from 0.30% to 0.05% as the coin grows. A later update capped fee redirects at one change per coin, so a deployer cannot keep re-pointing the revenue after launch. The stated direction is a "market-led" model where traders, not deployers, decide which coins deserve fee rewards.

2. Mayhem Mode and Token-2022 coins (November 2025 onward). A creator can now switch on Mayhem Mode at launch (never after). The coin is minted with 2 billion tokens instead of 1 billion, and an autonomous pump.fun trading agent uses the extra billion to buy and sell the coin in a random walk for its first 24 hours, then burns whatever it did not sell. Technically, these coins are created through the newer create_v2 instruction under Solana's Token-2022 program rather than the classic SPL token program, which is why older bots, trackers and some wallets either do not see them or fail to trade them. If a tool you rely on has not been updated for Token-2022, a Mayhem coin is invisible to it.

3. USDC curves and custom pairs (May 2026 onward). Coins can be launched with a USDC-priced bonding curve instead of SOL, and more recently against tokenized stocks, wrapped BTC/ETH and metals. Same 1.25% fee, same graduation logic, but the coin's market cap stops tracking SOL. Pump.fun rolled out a unified set of v2 trade instructions so integrations can trade every kind of coin with one interface.

4. Holder Rewards replaced Cashback. The short-lived cashback mode (refunds for churning volume) was retired in favor of "holder rewards" coins, where a creator can opt to distribute rewards to people who hold rather than people who trade. It is a creation-time flag like Mayhem, and it changes the incentive for the coin's community, not the fee you pay.

5. Creation is free, graduation is not. The token creation fee is now 0 SOL (you only pay the network fee, and ~0.025 SOL if you buy at creation). Graduation carries a 0.015 SOL charge and goes to PumpSwap, never Raydium.

6. PUMP buybacks became policy. Half of the protocol's net fee revenue is now committed to buying PUMP on the open market and burning it. That is the platform side of the coin story, and it is the next section.

What this adds up to for a trader: the platform now has explicit incentives for deployers (creator fees) and explicit tools for manipulating early price action (Mayhem's random-walk agent, bundled launches). Both make the first minutes of a coin noisier than they were in 2024, and both make the case for judging a wallet's track record over hundreds of trades rather than judging a single chart.

Many "pump fun crypto" and "pump fun token" searches are really about investing in the platform itself, not the coins listed on it. Here is what pump.fun states on its own token page and what the public record adds:

• Supply: 1 trillion PUMP total, with roughly 399 billion in circulation at the time of writing (pump.fun's own figure).
• Launch: a public sale in July 2025 that pump.fun says sold out in about 12 minutes, alongside allocations to the team, ecosystem and community, and an airdrop to active users.
• Buyback and burn: "Half of every dollar Pump.fun earns buys $PUMP on the open market, then burns it forever." The platform's live counter showed about $455 million deployed and about 166 billion PUMP removed from circulation at the time of writing. Reports describe an April 29, 2026 event in which roughly $370 million of accumulated PUMP was burned in two on-chain transactions, cutting circulating supply by about a third, after which the 50%-of-net-revenue commitment was locked into a smart contract for the following twelve months.
• No revenue rights. Pump.fun's page states in so many words that "the $PUMP token does not represent a right to revenues or any other distribution", and that future purchases may not continue. The buyback is a policy, not a legal claim.
• Where it trades: as a normal SPL token on Solana DEXs and on major centralized exchanges. It is not on a bonding curve.

The two questions Google's "People also ask" box keeps surfacing deserve straight answers. "Can PUMP reach $1?" At the ~$0.0037 price and ~$1.5 billion market cap shown on pump.fun's page at the time of writing, $1 would be a roughly 270x move, implying about $400 billion of market cap on the circulating supply and $1 trillion fully diluted. That is arithmetic, not a prediction; draw your own conclusion about the probability. "Is pump fun a good investment?" PUMP is a leveraged bet on Solana memecoin volume: the buyback scales with fees, fees scale with volume, and volume is the most cyclical number in crypto. We do not give investment advice; we do note that the wallets that consistently profit from pump.fun in our data do it by trading the coins on the platform, not by holding the platform's coin.

The airdrop question ("is there a pump fun airdrop?") is a topic of its own; our pump.fun airdrop guide covers the history, eligibility and how to position for future rounds.

Yes, a small number of wallets make a lot, and the on-chain record lets you check exactly which ones. That is the honest version of the answer to Google's most-asked question about the platform, and it is more useful than "most people lose" (true, but not actionable).

Two sets of numbers. First, aggregate data from our own production database, published in full in Solana copy trading statistics (1,710 copied positions, September 2026 snapshot):

• 77% of copied positions execute on the pump.fun bonding curve. PumpSwap and Raydium together account for less than a quarter.
• Closed bonding-curve positions won 84.9% of the time, versus 22.7% on Raydium, 33.3% on PumpSwap and 0% on Jupiter-routed trades. Closed-position win rates overstate reality (losers get abandoned as open dust more often than winners), but the gap between venues is the point.
• The median copied position is held 24 seconds. Pump.fun trading, done by people who are good at it, is latency-sensitive scalping, not investing.

Second, the individual wallets. These are the six most-searched Solana memecoin traders we profile, with the 30-day on-chain snapshot ending September 8, 2026 that each profile is built on. Every address is public, every figure is auditable on the linked tracker page, and none of it is a screenshot:

Trader	30d realized PnL	Win rate	Trades	Live tracker
rayan	$670,067	45.71%	5,153	tracked in real time
Cupsey	$447,372	94.25%	2,346	tracked in real time
Dior	$366,009	53.95%	1,655	tracked in real time
Cented	$224,081	77.42%	2,428	tracked in real time
Gake	$184,788	100%	80	tracked in real time
West	$135,299	83.87%	1,145	tracked in real time

Read that table as data, not as a leaderboard of heroes. The highest PnL in it comes with a sub-46% win rate (rayan: many small losses, a few enormous wins), the perfect win rate comes with only 80 trades in a month (Gake: a selective sniper who may not trade for hours), and the two wallets in the middle traded almost the same number of times with very different results per trade. Win rate alone tells you nothing; PnL alone tells you nothing; the combination, over hundreds of trades, tells you what a wallet's engine is. That is also why the full list of tracked traders shows all three numbers side by side rather than ranking on one.

The honest caveat, which every one of those profiles repeats: on-chain data proves what a wallet did, never why. Some of the profit on pump.fun comes from same-block entries next to a deployer, from bundled launches, or from information a copier will never have. An edge you cannot reproduce is worthless to you. The four checks worth running before you copy anyone are entry timing versus the launch block, token quality of what they buy, distribution of profit across trades (one outlier or many), and repeatability week over week. Tools like KolScan and the best KOL trackers exist for exactly that vetting.

If you have never traded a pump.fun coin, here is the end-to-end process, including the parts that go wrong for most beginners:

Set up a Solana wallet. Phantom, Solflare, Backpack or any Solana-compatible wallet. Fund it with at least 1-2 SOL; below that, priority fees and the 1.25% curve fee eat too much of every trade. Use a separate "burner" wallet for memecoin trading and keep your real holdings elsewhere; our best Solana wallet guide explains the vault / daily driver / burner split.

Decide where you will trade. Pump.fun's own site and app work, but the interface is built for browsing, not for speed. Most active traders use a browser terminal (Axiom, Photon, BullX, GMGN) for charts and one-click execution, or skip stock-picking altogether with a pump fun bot or copy trading. Whatever you use, check that it supports Token-2022 coins, or Mayhem launches will simply not appear.

Find a coin. This is where most people lose. The "trending" tab is full of coins that already ran and are about to unwind. Better signals: the deployer's history (have their previous coins graduated or rugged?), holder distribution (no single wallet above a few percent of supply), whether the top holders bought in the same block as the launch (a bundled launch), and where the coin is on the curve. Run every candidate through a rug check first; it takes ten seconds.

Buy carefully. Set a slippage cap (10-15% on the curve, lower on graduated coins) and size the position as money you can lose in full, because on the curve that is not hypothetical. Prefer MEV-protected or Jito-bundled submission on hot launches, where sandwich attacks and failed transactions are the norm.

Have the exit planned before the entry. The biggest source of pump.fun losses is not bad entries, it is not selling. Ladder take-profits (a quarter at 2x, a quarter at 5x, and so on), set a stop, and let the tool execute them. The good sniper bots and copy trading platforms automate exactly this.

Track and review. Log every trade with the reason you took it. Almost nobody does, which is why almost nobody improves. Our broader guide on how to trade memecoins covers sizing, timing and the psychology in detail.

Pump.fun is genuinely profitable for some traders and a meat grinder for most. The data picture, without the marketing:

• Under 2% of coins ever graduate. The rest die on the curve, usually within hours.
• Even graduates mostly fade. The migration pump is real, and so is what follows it once the migration buyers are the exit liquidity.
• Profit is concentrated in a small number of wallets that have at least one edge: infrastructure (private RPC, gRPC streaming, Jito bundles), information (deployer or community signal before it is public), or pure speed (sub-block execution on the first buy).
• Everyone else pays the fees. At 1.25% per side on the curve plus tooling, a retail trader with no edge is structurally negative before slippage and rugs are counted.

This is not unique to pump.fun; our analysis of whether copy trading is profitable covers the broader retail statistics, where the large majority of self-directed traders end up net negative. On pump.fun the velocity is higher and the scams are more aggressive, so the base rate is worse. The pragmatic conclusion: if you do not have an information, infrastructure or speed edge, your best expected value on pump.fun is to copy the wallets that do, after vetting that their edge is reproducible.

The tooling landscape is a topic of its own (see our tested rankings of pump fun bots and the best pump.fun bots by use case), so here is only the map:

For chart analysis and fast manual buys: Axiom (strong UI, lower fees), Photon (fast execution), GMGN (scanners and wallet tracking), BullX (multichain, pricier on pump.fun).

For sniping launches: Trojan, BonkBot, Maestro and Banana Gun on Telegram (convenient, but the bot holds or generates your key), or browser snipers like Photon and Axiom for non-custodial sniping. The whole category is covered in our memecoin trading bot comparison.

For wallet tracking and copy trading: uwuu.ai for automated, non-custodial copy trading on Solana with sub-400ms execution and a verified on-chain leaderboard; wallet trackers and KolScan for finding and vetting wallets before you copy them.

For research: Solscan for the raw on-chain record, Birdeye and DexScreener for token analytics, and Crypto Twitter with heavy skepticism applied.

The unifying pattern: profitable traders research with one set of tools, find proven wallets with another, and execute with either custom infrastructure or copy trading. The least profitable rely on the trending tab and influencer calls.

Connect the dots from the sections above:

1. The pump.fun edge is concentrated in a small number of wallets.
2. Those wallets' addresses and every trade they make are public on Solana, in real time.
3. If you mirror their trades fast enough, you get close to the same entry they do; 77% of the time that entry is on the bonding curve, where the measured win rate is highest.
4. You do not need to build the infrastructure or do the analysis yourself.

This is the core thesis of crypto copy trading, and pump.fun is the market where it works best, because the combination of extreme volatility, fully public data, a handful of consistent winners and sub-second Solana execution is unique. It is also why uwuu.ai is built the way it is:

• A verified leaderboard of top-performing wallets: real PnL, trade counts and win rates, all auditable on-chain.
• Pick two to four wallets with different engines (a high-volume scalper, a selective sniper, a migration specialist).
• Set your trade size and filters (slippage caps, token blacklists, stop-losses) once.
• uwuu mirrors every trade in roughly 400ms, non-custodially, through a copy key that never gives anyone your private key.
• You pay only when you profit. No per-trade charge on losing trades.

For the full setup, see how to copy trade on Solana; for the wider tool landscape, the best Solana trading bot guide compares copy trading with snipers, terminals and Telegram bots.

None of these are theoretical. All of them happen every day:

• Rug pulls and bundled launches. A deployer who bought most of the early curve with a cluster of wallets can dump it all at once. Most pump.fun rugs happen within the first hour.
• Creator-fee farming. Because a creator earns 0.30% of every trade, some launches exist only to generate volume for the deployer. The coin does not need to succeed for its creator to profit.
• Mayhem-mode noise. For 24 hours after a Mayhem launch, an automated agent is buying and selling the coin at random. The chart you are reading is partly synthetic by design.
• Tooling gaps on Token-2022 coins. If your bot, wallet or tracker has not been updated for Token-2022, it may fail to display or trade newer coins, or trade them with the wrong accounts and fail.
• Curve impact and slippage. On a thin, fresh curve your own buy can cost you 20-40% in immediate price impact. Set realistic caps.
• Failed transactions during congestion. Hot launches saturate Solana. Failed transactions still consume priority fees; budget for 20-40% failure rates at peak.
• MEV and sandwiches. Use MEV-protected submission or Jito bundles on anything hot.
• Custodial bots. A Telegram bot that holds your key puts the whole wallet at risk if the bot is compromised. Prefer non-custodial tools (uwuu, Axiom, Photon) and keep your real funds in a separate wallet.
• Emotional ruin. The biggest one. Pump.fun produces euphoric wins and devastating losses in the same afternoon. Most traders revenge-trade their way to zero; disciplined automated exits, or copying wallets that have them, protect you from yourself.

Pump.fun is not the only Solana launchpad any more, but it remains the dominant one. For the full head-to-head across LetsBonk, Raydium LaunchLab, Moonshot, Boop.fun and the rest, see our Solana launchpad comparison. Here is how the closest alternatives line up on mechanism and fee:

Platform	Chain	Mechanism	Trading fee
pump.fun	Solana	Bonding curve (SOL, USDC or custom pair) + auto-graduation to PumpSwap	1.25% on the curve (0.95% protocol + 0.30% creator)
PumpSwap	Solana	AMM for graduated pump.fun coins	0.30% to 1.25%, tiered by market cap on canonical pools
Raydium LaunchLab / LetsBonk	Solana	Bonding curve, graduates to Raydium pools	~1% on the curve (verify current)
Moonshot	Solana	Bonding curve variant	~1% per trade (verify current)
Four.meme	BSC	Bonding curve	~1% per trade (verify current)

For Solana memecoin trading specifically, pump.fun and its graduated counterpart PumpSwap dominate the volume, with Raydium and LaunchLab the main alternatives. Copy trading works across all of them: uwuu detects and mirrors trades on the pump.fun curve, PumpSwap, Raydium and LaunchLab from one dashboard, routing each copy through the same venue as the original trade.

Every coin on pump.fun follows the same arc. Knowing which phase you are entering removes most "surprises":

Phase 1: Creation (T+0 to T+30 seconds). The deployer uploads metadata, the curve is deployed, and the coin is buyable. The very first buyers, almost always bots, set the initial price action. If the deployer bundled buys into the creation transaction, the top of the holder list is already decided.

Phase 2: Initial pump (30 seconds to ~5 minutes). If the coin catches attention (name, deployer history, social signal, or a Mayhem agent randomly buying), buying spikes the price up the curve. Early snipers exit here for their multiples; late buyers in this phase are usually their exit liquidity.

Phase 3: First retracement (5-30 minutes). A 30-70% pullback as early buyers take profit. Most "I should have sold" moments happen here. Bots with exit logic do well; manual traders tend to hold through it.

Phase 4: Consolidation or death (30 minutes to hours). The overwhelming majority of coins die here, drifting to near zero. A few find a base and start building toward graduation.

Phase 5: Graduation push. As the curve nears the threshold, buying often accelerates because traders front-run the migration.

Phase 6: Migration (instant). The curve closes, the 0.015 SOL fee is taken, and liquidity migrates atomically to PumpSwap. Migration itself frequently produces a spike as DEX-only buyers finally get access.

Phase 7: Post-migration (ongoing). The coin now lives on PumpSwap at 1.25% fees that step down as it grows. Most fade over days; a small fraction become real memecoins with sustained communities.

The whole lifecycle compresses into one to three hours for a typical coin. The traders who win consistently know which phase they are entering and have a thesis for which phase they are exiting in.

Some scale, with sources, because "billions" gets thrown around loosely:

• Protocol fees: about $46 million over a trailing 30 days per DefiLlama at the time of writing, an annualized run rate of roughly $400 million; pump.fun's own dashboard was reported at a ~$468 million annualized rate (about $1.3 million per day) in September 2026.
• Launches per day: tens of thousands at peaks; Solana Compass reported a single day above 42,000 new coins, and pump.fun's own docs cite a daily average around 17,000-18,000 when Mayhem Mode launched.
• Graduation rate: under 2% of coins reach a DEX.
• PUMP buybacks: about $455 million deployed and about 166 billion PUMP burned, per pump.fun's live counter at the time of writing.
• Where copy trades happen: 77% on the bonding curve, per our own 1,710-position dataset.
• How long they last: a 24-second median hold.
• Top tracked wallets: six-figure realized PnL over 30 days for each of the six wallets in the table above, publicly verifiable.

The takeaway is not "pump.fun is bad", it is "pump.fun is skewed". A small number of coins and wallets capture an outsized share of the upside. Your job is either to be in that group (hard) or to copy it (much easier; the uwuu leaderboard shows who is in it right now).

Traders go through a predictable arc too. Recognizing your stage can save you a lot of money:

Stage 1, the lurker. You have seen the screenshots, you are watching, you have not traded. Healthy; people who skip it lose faster.

Stage 2, the first-timer. Your first 5-20 trades. Beginner's luck on memecoins is real, and you conclude this is easy.

Stage 3, the aggressive phase. You size up after early wins, enter trending coins late, and start losing. You blame the tools, the timing, the rugs.

Stage 4, the bot phase. You set up a Telegram sniper, configure filters, and auto-snipe. It feels like leveling up. Net PnL is mixed once fees and failed transactions are counted.

Stage 5, the filter-optimization phase. Hours a day tuning configurations for marginal returns. Some weeks are great; most are flat or negative after fees.

Stage 6, the realization. The wallets actually winning are not using your tools. They have private infrastructure, custom filters, sometimes teams. You cannot compete on their terms.

Stage 7a, the quit. Many stop here, bankroll spent on tuition.

Stage 7b, the pivot. A growing number stop trying to be the sniper and copy the snipers instead, using a platform like uwuu.ai to mirror the wallets that are winning. Hours per day become minutes per week, and PnL stops depending on their own reaction time.

If you are at stage 4-6, the rational move is to skip to 7b, and to spend the saved hours vetting which wallets deserve to be copied.

What is pump fun crypto?

"Pump fun crypto" refers to two things. Pump.fun is a Solana memecoin launchpad where anyone can create a token for free and trade it against a bonding curve with a 1.25% fee until it graduates to the PumpSwap DEX. PUMP is the platform's native token, a 1-trillion-supply coin backed by a policy of spending half of protocol fees on open-market buybacks and burns. Buying a coin on pump.fun and buying PUMP are completely different bets.

Is pump.fun legit and safe to use?

The platform itself is non-custodial: your funds stay in your wallet, and graduation is automatic and irreversible with liquidity owned by the protocol. The coins listed on it are a different matter. Most are worthless within hours, many are bundled or rug-pulled by their deployers, and the platform's safety does not extend to them. Treat pump.fun as safe infrastructure for trading extremely unsafe assets, and use a separate wallet for it.

Do people actually make money on pump.fun?

A small number of wallets do, verifiably. In our dataset of 1,710 copied positions, 84.9% of closed bonding-curve positions were profitable, and the six most-searched tracked traders each show six-figure realized PnL over 30 days on-chain. The majority of retail wallets are net negative after the 1.25% curve fee, tooling fees, failed transactions and rugs. The realistic paths to profit are having a real edge (speed, infrastructure, information) or copying wallets that have one.

How much does pump.fun make a day?

Roughly $1.3-1.5 million a day in protocol fees at the time of writing: DefiLlama showed about $46 million over a trailing 30 days, and pump.fun's own dashboard was reported at a ~$468 million annualized rate in September 2026. The number moves with memecoin volume, which is the most cyclical figure in crypto, so treat any daily number as a snapshot.

Can the pump fun token (PUMP) reach $1?

At the ~$0.0037 price and ~$1.5 billion market cap shown on pump.fun's token page at the time of writing, $1 would be about a 270x move, implying roughly $400 billion of market cap on the circulating supply and $1 trillion fully diluted. Pump.fun also states that PUMP carries no right to revenue; the buyback is a policy, not a claim. That is the arithmetic; the probability is your call, not a prediction we make.

What are pump.fun fees and how do creators claim them?

Creating a coin is free apart from the Solana network fee. Every bonding-curve trade costs 1.25%, split 0.95% to the protocol and 0.30% to the coin's creator; graduation costs 0.015 SOL; PumpSwap trades cost 0.30% to 1.25% depending on the coin's market cap. Creators claim their accumulated fees from the creator dashboard on pump.fun, and since January 2026 can route them to up to 10 wallets, though only one reassignment per coin is allowed.

Pump.fun is not going anywhere. It is the dominant launchpad in the fastest market on the highest-throughput chain, it is generating fees at a rate most protocols would envy, and its own token now recycles half of that back into buybacks. The volume is real, the wins are real, and the losses are more real still.

If you participate, do it with the right framing. This is the casino floor of crypto: some people own the casino (PUMP holders, deployers earning creator fees), some people count cards (the wallets in the table above), and most people lose. The difference is not luck; it is edge and a system.

If you have neither, copy the people who have both. Browse the uwuu leaderboard, check a wallet's 30-day PnL, win rate and trade count together, verify that its edge is one you can reproduce, and let the bot handle the 24-second trades. It is a cheaper education than the one most pump.fun traders pay for.