The graduation numbers this week

By Siddharth Kale — 2026-07-04 — 8-hour-recap

The graduation numbers this week

We watched 121,768 pump.fun tokens launch across a 168-hour window ending July 4. This is market background noise, the baseline. Of that volume, 836 graduated to Raydium. That's 0.69 percent. The number matters because graduation is the only exit condition that resembles a real market.

Most tokens on pump.fun burn. They reach their bonding curve ceiling, fail to reach it, or simply bleed to zero as momentum collapses. Graduation is the rare survival case. It means enough volume moved through the bonding curve to hit the hard cap, and the smart contract executed the migration to Raydium—where actual liquidity pools form and price discovery happens on a real exchange.

That 0.69% figure is not a volatility measure or a success rate in any trader's sense. It's a structural fact. It tells you the shape of the market at the source.

Speed of graduation

The tokens that made it graduated fast. Average time from launch to graduation: 77 seconds. The fastest was 1 second. The slowest we saw in the graduation cohort was 15 minutes.

One second is not luck. One second means volume arrived instantly—either a pre-positioned buyer moving immediately after launch, or a coordinated group hitting their entries at the moment the contract went live. Fifteen minutes means sustained buying pressure across a longer window, less concentrated but still sufficient to breach the ceiling before momentum broke.

The median graduation we saw was probably closer to the fast end of that range. If you're watching for tokens that escape pump.fun alive, watch the first two minutes. If it hasn't graduated by then, the probability it will drops hard.

What this means for market structure

Pump.fun's bonding curve was designed to create urgency. The curve accelerates as more people buy. Early buyers face lower prices. Late buyers face exponential cost increases. That arithmetic creates a deadline. Hit the curve ceiling and you graduate. Miss the window and the token either stalls or fails.

The 0.69% graduation rate is not a design flaw. It's the intended outcome. Most tokens fail because the design is supposed to be zero-sum. Early buyers win by dumping on later buyers. Later buyers lose because they arrived too late to capture the curve advantage. Graduation happens only when the incentive structure aligns—when enough capital is committed at launch that the ceiling gets breached before the pyramid of buy-side interest collapses.

What's structural in this data is the speed bifurcation. Tokens that graduate do so in minutes, often seconds. Tokens that don't graduate tend to stall hours in. There's no middle ground. Either the initial capital wave hits hard enough to close the curve, or it doesn't. You either have volume arriving in the first moments or you don't have it at all.

Observational note on market behavior

Over 168 hours, 121,768 tokens generated enough transaction data to track. The volume per token was not uniform. Some attracted thousands of participants. Most attracted handfuls. Only 836 generated enough money flow to trigger the graduation rule.

We're not in a period of widespread graduation acceleration. The rate has been stable around this figure week to week. It suggests the bonding-curve model is working exactly as specified—most tokens fail as designed, a tiny fraction succeed and move to a real market.

For anyone watching these launches, the data is clear. Graduation is the only outcome that separates a token from immediate death. It's rare. When it happens, it happens in seconds to low double-digit minutes. If you're trading pump.fun tokens, you're timing the window before graduation or the window after. There is no profitable middle state. The tokens that don't graduate fast enough don't graduate at all.

The takeaway is mechanical: If your entry window is longer than two minutes, you're probably already late.

Tags: pump.fun, graduated, solana launch