Pump.fun vs PumpSwap: two completely different games

People talk about running a volume bot on Pump.fun as though it were one activity. It is two, and they have almost nothing in common beyond the brand. On a bonding curve you are buying progress toward graduation in a contest where most attempts lose. On a PumpSwap pool you are buying visible, sustained activity in a market where almost every attempt lands. Different objective, different economics, different definition of success. Deciding which game you are playing is the first thing worth getting right, because the settings and the budget follow from it.

Volion Research Updated Jul 30, 2026 5 sections

The two campaigns want different things

A curve campaign has one meaningful objective: move the token closer to completing its bonding curve. A pool campaign has a different one: produce sustained, visible activity that data aggregators and human observers register. Volume on a curve barely serves the second goal, and volume in a pool cannot serve the first at all, because there is no longer a curve to advance.

The confusion arises because both are described as buying volume, and in a narrow accounting sense both do. But the thing that volume purchases differs completely. Curve buys move a deterministic counter toward a threshold. Pool buys create trades on a chart that other systems observe.

Being clear about which you want changes obvious things. If the objective is graduation, a curve campaign is the only tool, and its cost is what it is. If the objective is visibility, spending on a curve is an expensive way to get a small amount of what you wanted, and waiting is materially cheaper.

What actually gets picked up by third-party trackers is covered in how DEXScreener trending works.

The economics are not comparable

A bonding curve is a queue where many buyers compete for the same next fill and only one wins, so failures are the normal case and every loser pays a fee. A pool absorbs concurrent trades against depth, so two buyers in the same block both succeed at slightly different prices. The failure rate difference is the entire cost story.
ProgramSampled txFailure rateMedian feeCost per landed swap
Pump.fun bonding curve low sample2373.9%5,70021,839
Pump.fun AMM (post-graduation)17332.4%6,6559,845
The two halves of a Pump.fun token life, measured Sep 15, 2026 at 4:02 AM UTC across 1 finalized mainnet blocks. Fees in lamports.

In the current window 73.9% of curve transactions failed against 32.4% on the pool. Both figures still pay their fee.

The consequence is that a budget converts into landed swaps at very different rates on each side. The same SOL that produces a modest number of curve fills produces a considerably larger number of pool trades, and it is the count of landed trades, not the amount spent, that a chart reflects.

The mechanical detail of what changes at the boundary between the two is in what changes the day your token migrates, and the wider venue comparison is on which Solana DEX is cheapest for a volume bot.

Playing the curve game well

On a curve, you are competing for position in a queue. Speed matters, retry policy matters, and trade size matters mainly because smaller trades mean more attempts for the same budget. What does not help is patience, because the contest resolves per slot and being late is simply losing.

Three things determine whether a curve campaign is worth running at all:

  • How close the curve already is. A token near completion needs a modest push. A token at a fraction of the way needs a budget most people would not accept if it were stated plainly, and stalling curves stall for reasons a campaign cannot fix.
  • How contested it is. The failure rate on a curve nobody is watching is far lower than on one attracting attention. The cost of a curve campaign is set by competition, not by the program.
  • Whether graduation is actually the goal. If what you want is a chart that looks alive, this is the expensive road to it.

The mechanics of curve pricing, and why position in the buy order determines your effective slippage, are in how the Pump.fun bonding curve works.

Playing the pool game well

In a pool, you are not competing with anyone for a fill. You are trying to produce a pattern of activity that reads as genuine while spending as little as possible per landed trade. That makes pacing, wallet diversity and trade size distribution the levers that matter, rather than speed.

Everything that mattered on the curve stops mattering here. There is no race, so bidding priority fees up for position is pure overpayment. Failures are no longer routine, so aggressive retrying is wasteful rather than rational. And since concurrent trades all land, the constraint on throughput is your wallet fleet rather than the venue.

What replaces them is a set of concerns that did not exist on the curve. Activity that arrives in perfectly even intervals from wallets funded identically minutes earlier is legible as automation to anyone looking, and the people most likely to look are exactly the ones a campaign is meant to attract. Fleet preparation is covered in how many wallets you need and the patterns that give a fleet away are in wallet clustering risk.

Trade size deserves its own note. Because a chart reflects trade count and distinct participants more than notional size, many small trades usually outperform a few large ones at identical spend, and they also move the price less. The distinction is set out in makers versus volume.

Deciding which game you are in

If the token is still on its curve and completing it is the objective, run the curve campaign and budget for the failure rate. If the token has graduated, or if the objective is visibility rather than graduation, the pool is where the budget belongs. The expensive mistake is running a visibility campaign on a curve because that is where the token happens to be today.

A sequenced approach often makes sense when a token is close to graduating: a contained curve campaign to help it complete, then the main budget deployed once liquidity sits in a pool. That spends the expensive SOL only on the thing the cheap SOL cannot buy, which is progress toward a threshold that stops existing afterwards.

What neither game does is create demand. A campaign produces activity, and activity is a prerequisite for attention rather than a substitute for it. A token with nothing behind it ends up with an expensive chart and no holders, which is the outcome most disappointed campaigns describe. The honest version of what a campaign can and cannot deliver is on is a volume bot safe and in how volume bot scams work.

Both games are supported without changing anything on your side: entering a token address in the dashboard resolves which one applies from on-chain state, and a token that graduates mid-campaign is followed across. Costs for either can be modelled first with the calculator, and the measured basis behind those figures is explained on the page for our volume bot for Pump.fun tokens.

Questions

What is the difference between running a bot on Pump.fun and on PumpSwap?
On a bonding curve you are competing for a single next fill, so most attempts fail and each failure still pays a fee. In a PumpSwap pool trades are absorbed by depth, so concurrent swaps all land at slightly different prices. The objectives differ too: curve volume advances graduation, pool volume produces visible activity.
Can volume on a curve make a token trend?
Barely, and expensively. Curve activity does contribute to visibility, but the high failure rate means most of the budget produces nothing, and aggregator coverage of pool trading is more reliable. If visibility is the goal, waiting for graduation is materially cheaper.
Should I wait until my token graduates before running a campaign?
If the goal is visibility, usually yes. If the goal is helping the curve complete and the token is close, a contained campaign before graduation is the only tool that serves it. Sequencing the two is often better than choosing one.
Why do my settings need to change after graduation?
Because the constraints change entirely. Speed and retry aggression matter on a curve where you are racing; in a pool nobody is racing you, so bidding up priority fees is overpayment and hard retrying wastes fees. Pacing and fleet diversity take over as the levers that matter.
Does a volume bot create demand for a token?
No. It produces activity, which is a prerequisite for attention rather than a replacement for it. A token with nothing behind it ends up with an expensive chart and no holders, which is the most common disappointed outcome.