The meme coin market, decoded
We dug through every trade, transfer and launch on chain so you don't have to. Here's what's really going on behind the charts (spoiler: the crowd is often one guy).
The last 24 hours on pump.fun
2026-11-08 22:00 UTC to 2026-11-09 22:00 UTC, straight from the chainThe market, week by week
A day's average for every full week since April. The big number is the week of November 2. Each one opens its finding.
What the chain says
Every finding answers a question with a number, the window we measured it over, and the data behind it, free to download.
Launches and graduations
pump.fun by the numbers
trades a day on pump.fun and PumpSwap, 3.2 times as many as in May.
Measured 2026-11-09New pageHow many pump.fun coins graduate?
graduate through trading within a day. 38% of graduations are instant one-bundle migrations.
Measured 2026-11-09New pageMost coins are over in a minute
of launches see no trade of 0.05 SOL or more after their first minute.
Measured 2026-10-30New pageGraduating isn't the finish line
of graduates are below half their graduation price an hour later.
Measured 2026-10-30New pageA chunk is gone before you can click
of launches lose 10% or more of the supply in the coin's very first block.
Measured 2026-11-09New pageCheck the contract, not the ticker
of launches share their ticker with another coin launched in the same three weeks.
Measured 2026-10-30In briefThose big "winners" are usually staged
of coins that hit a big market cap after migrating were staged fake-cap launches.
Measured 2026-04-18 to 2026-09-24Who's really trading
Linked wallets do a lot of the early buying
of a hot coin's first-minute buying comes from groups of linked wallets (the median).
Measured 2026-11-09New pageYou were never going to be first
to a new coin's first outside buy, the median. 26% of launches never get one.
Measured 2026-10-30New pageDevs repeat themselves (a lot)
of all launches come from 6,961 wallets that launched 100+ coins each.
Measured 2026-11-09New pageThe dev is often first out
of launches have the dev's wallet selling within a minute.
Measured 2026-10-30New pageMost active wallets lose
of wallets with 10 or more closed positions ended down, before fees.
Measured 2026-10-30New pageThe crowd isn't doing the trading
of trades come from 1% of wallets. 1,539 wallets make 31%.
Measured 2026-10-30In briefThe launch crowd is often one guy
buys by the top sniper wallets are sybils: one dev funding the whole crowd.
Fresh numbers comingIn briefSkill is real, just rare (and it sticks)
A few wallets really are good, and it shows up again on trades we never measured them on.
Fresh numbers comingCopy-trading and bots
The findings so far
Each of these is getting its own page, with fresh numbers and the data behind it.
Most coins are run, not traded
Look at how wallets actually behave (who signs together, who funds who, who buys the same coin in the same millisecond) and the "crowd" around a typical launch shrinks to a handful of linked groups. Across the whole market, ~4.5M wallets boil down to a few hundred thousand of them. The score in the panel is simply how much of a coin insiders hold right now: the dev, its linked wallets, and wallets funded together.
The launch crowd is often one guy
About a quarter of the buys made by the top sniper wallets land in the coin's own launch block, from wallets that buy one dev's coins over and over: the dev's own, dressed up as real demand. That "organic" launch on the chart is often one person wearing dozens of masks, and the funding trail gives them away.
Those big "winners" are usually staged
We looked at every coin that hit a big market cap after migrating. About 99% were fake-cap launches built to look like a breakout, and then they rugged. Real breakouts do happen. They're just rare, and you can only tell them apart with the on-chain history, not the chart.
Copying the "smart money" loses
Following smart wallets sounds like a cheat code. Our data says nope. At the moment you'd buy, the coins that moon and the coins that tank look exactly the same, and once you account for the delay and cost of following, copy-trading comes out negative. So we don't sell it. Hidden Hand shows you the setup, and you make the call.
The leaderboard wallets are getting farmed
The wallets topping public "smart money" leaderboards are mostly the ones being harvested: bots front-run and copy them within seconds. A heavily copied wallet is a warning sign, not a tip, and the wallet card will tell you.
Skill is real, just rare (and it sticks)
A small set of wallets really are good, and it shows up again on trades we never measured them on. That's the signal behind the wallet records: real, just a lot rarer than the leaderboards make it look.
Devs repeat themselves
The same people launch coin after coin, and they run the same playbook every time. A dev's track record is one of the best tells for how their next launch will go, so it's right there on the coin card.
The numbers, and when we measured them
Where coins actually trade, what's really behind their liquidity, and what their tokens let the dev do. Every number comes with the window it was measured over.
Where the coins trade
Launchpad coins make up about half of all Solana meme coin volume, and they bounce around a dozen venues: a launch pool first, then a graduation pool, in SOL or some other token. We read all of them, not just pump.fun.
| Venue | Share of launchpad volume | What it is |
|---|---|---|
| PumpSwap | 74.9% | Where pump.fun coins trade after they migrate |
| Pump bonding curve | 13.8% | Where a pump.fun coin lives before it migrates |
| Raydium LaunchLab | 4.5% | The launch venue for Stonkfun, Bonk and others |
| Meteora DLMM | 3.0% | Concentrated-liquidity pools, and the SOL market for many quote tokens |
| Raydium CPMM | 1.6% | Where LaunchLab coins graduate |
| Meteora DBC | 1.1% | Dynamic bonding curves, where DBC-launched coins start |
| Meteora DAMM v2 | 1.0% | Where DBC-launched coins graduate |
| Meteora DAMM v1 | <0.1% | Older pools |
Measured across 28.8 million trades in 24 hours on 2026-09-25. That day saw 34,510 pump.fun launches (24,023 standard and 10,487 in Mayhem mode), plus 2,451 DBC→DAMM v2, 1,030 pump→PumpSwap and 84 LaunchLab→CPMM graduations.
What's really behind the liquidity
A market cap on a chart tells you nothing about whether the liquidity is real or who controls it. On-chain, it's crystal clear.
| Pool fact | What it means for you |
|---|---|
| Drainable liquidity | A lot of trading can sit in a pool the dev can still pull. The panel flags it, because a market cap you can't sell into isn't worth much. |
| Locked or burned at graduation | Legit migrations lock or burn the liquidity. The card shows you how much was locked. |
| Mayhem coins | About 30% of pump.fun launches run in "Mayhem" mode, where an agent program holds a big block of supply. Flagged on the card. |
| Instant fake-cap migrations | One bundle creates a coin, migrates it, and buys about 96% of the new pool in the same block, printing a huge market cap out of thin air. Of coins that cross a 50,000 SOL market cap in their first day, 99.6% are these, and they crash below half their price in about 31 minutes (the median). |
| Real breakouts are rare | Coins that organically cross 5k / 10k / 20k / 50k SOL in market cap in their first day show up about 5.4 / 2.1 / 0.86 / 0.25 times a day, across all of pump.fun's migrated coins. |
Migration study: 87,898 graduations from 2026-04-18 to 2026-09-24, each coin followed through its first day, with the same-block bundles confirmed on chain, block by block.
Not everything trades in SOL
More and more coins are priced in another token, and that token comes with its own risks. The card shows them before you trade.
| Priced in | Share of volume | Notes |
|---|---|---|
| SOL | 89.0% | The default |
| Other tokens | ~8.3% | PUMP, ZEC, CARDS, BP and hundreds more. We counted 450 different ones in a single day, and the top 20 carry about 78% of that volume. |
| Tokenized stocks | ~2.1% | xStocks (NVDAx, SPCXx, COINx), Backpack Securities (SPCX, HIMS) and PreStocks (OPENAI, ANTHROPIC) |
| Stablecoins | ~0.6% | USDC and USD1 |
Stock-priced coins are brand new: pump.fun Custom Pairs switched them on 2026-09-09 (93 quote assets) and Meteora StockLaunch on 2026-09-15. Heads up, though: the stock token's issuer can seize, freeze or pause it, and if they pause it, nobody can sell any coin priced in it. Stablecoin issuers can freeze theirs too.
Breakdown measured on 2026-09-25. Launch dates from each platform.
Sneaky token powers
Solana's Token-2022 standard lets a dev bake powers into a coin that most holders never check. The panel reads them straight off the coin and flags the scary ones.
| Power | What it lets the dev (or issuer) do |
|---|---|
| Mint authority | Print more of the coin whenever they like, diluting everyone |
| Freeze authority | Freeze any holder's tokens, so they can't sell |
| Transfer hook | Run custom code on every transfer, which can tax or block your sell |
| Transfer fee | Skim a built-in cut of every transfer, including your sells |
| Permanent delegate | Pull tokens out of anyone's wallet, any time |
| Pausable | Stop all transfers of the coin |
| Default-frozen accounts | New holders start out frozen until the issuer lets them in |
Read live from each coin. Tokenized stocks carry permanent delegate, freeze and pause on purpose, which is exactly why the card calls them out.
How we know
Everything here comes straight from the chain: every Pump trade, transfer and launch since May 2026. No screenshots, no self-reported numbers, and nothing a scammer can fake. The models get rebuilt every night and checked on trades they've never seen before.
See it on live coins
All of this runs live in the panel, on whatever coin or wallet you're looking at. Free on the Chrome Web Store, no wallet needed.