How much should a dev buy at launch on Pump.fun?

Your dev buy is a number that stays public for as long as the coin exists. Size, timing, and everything you do with it afterwards.

Most creators decide it in 30 seconds and then live with it. Worth spending longer.

What the decision actually trades off

Two failure modes pull in opposite directions.

Buy too little and the curve sits flat in the opening minutes. A coin at 0% with no trades reads as abandoned before anyone has had a chance to look properly, and that first impression is doing real work.

Buy too much and every holder who checks distribution sees one address holding a large share. That reads as a distribution waiting to happen, and the people most likely to buy early are exactly the people most likely to check.

So you're choosing between looking dead and looking dangerous, and both cost you buyers.

What holders actually look at

Be specific, because "looks bad" is uselessly vague.

Somebody evaluating your coin opens the holder list and scans the top 10. What share does the biggest wallet hold? Is the creator wallet among them? How steep is the drop from first to tenth?

They're pattern-matching against every coin they've seen, and the pattern that worries them is concentration. A creator holding a small single-digit percentage among a spread of other holders barely registers. A creator sitting in the top 3 with a double-digit share registers immediately. How to use Solscan covers where to check your own numbers before somebody else does.

Timing matters as much as size

Early on the curve, a given amount of SOL moves price far more than the same amount later, because price comes from cumulative purchases and the pile is tiny at the start.

Which means a dev buy in the first seconds does two things at once: it takes tokens cheaply, and it moves the price for everybody arriving after. Both are visible. Both get interpreted. Why early buys move the price covers the arithmetic.

Buying a minute or two in, after some organic activity, produces a less conspicuous entry at a slightly worse price. Whether that trade is worth making depends on how much you care about how the holder list reads.

The thing that actually destroys a launch

Not the size of the buy. What you do with it.

A creator wallet that sells into its own launch is the single most damaging visible action available to you. It shows up in the transfer list with a timestamp, it moves curve progress backwards, and it tells every holder that the person who made this thing is taking money off the table while asking them to stay.

Coins recover from small mistakes. They don't recover from that.

A workable approach

Nobody can give you a universal number, so here's the reasoning instead:

1. Buy enough that the curve isn't flat when the first stranger arrives
2. Keep the creator wallet out of the top few holders if you can
3. Buy once, at the start, rather than adding through the day where it looks like accumulation
4. Decide before minting whether you'll ever sell, and if the answer is yes, decide when
5. Assume every part of this is being read, because it is

The fourth point matters more than it sounds. Creators who never decided end up selling reactively during a dip, which is the worst possible moment and the most damaging possible signal.

What the dev buy can't do

It can't make your coin discovered.

Feeds rank on volume, trade count, unique wallets and recency. One wallet buying once moves the first two barely and the third not at all. A coin with a healthy dev buy and no other activity is a coin nobody has seen, and the dev buy just makes the emptiness better funded.

That's the distribution half of a launch, and it's covered in the launch checklist. If it's the gap you're facing, the console prices sustained activity against a specific window and wallet count, with the full cost shown before anything is signed.