Volume bot scams and the cost floor that exposes them

You do not need to know whether an operator is honest. You need to know whether their quote is physically possible. Every campaign has a floor made of costs nobody controls: the network fee on every attempt including the failures, the venue swap fee on every trade, and the rent on every token account. A quote below that floor is not a bargain, it is arithmetic that does not close. Here is how to compute the floor from measured data and test any offer against it.

Volion Research Updated Jul 30, 2026 4 sections

Computing the floor for any volume target

The floor has three parts. Network fee is the number of swaps divided by the success rate, multiplied by the median fee. Venue swap fee is the volume target multiplied by whatever percentage the pool charges. Rent is roughly 0.00204 SOL per token account. Add them and you have the minimum a campaign can cost before anyone takes a margin.
Volume target (SOL)Swaps at 0.2 SOLNetwork floor, curve (SOL)Network floor, AMM (SOL)Rent held (SOL)Venue fee at 1% (SOL)
1005000.0410.0051.021.00
3001,5000.1220.0163.063.00
1,0005,0000.4070.05410.2010.00
3,00015,0001.2220.16130.6030.00
Network floor uses a measured 92.4% failure rate on the Pump.fun bonding curve and 23.3% on the Pump.fun AMM, from our latest sample of finalized blocks. The venue fee column is illustrative at one percent; substitute the actual fee of the pool your token trades in, because it varies by venue and by fee tier. Jito tips are excluded because they are auction-priced.

Two things fall out of that table immediately. First, the venue swap fee is usually the largest line by a wide margin, and it is charged by the pool rather than by any operator, so no vendor can discount it. Second, the network floor is not the same number on every venue: routing the identical campaign to a contested bonding curve rather than a settled pool multiplies it, because failures are charged in full.

Now use it. Take any quote you have been given, subtract the floor for the volume it promises, and look at what is left. If the remainder is negative, the offer cannot be delivered as described regardless of who is offering it. If the remainder is small but positive, you are looking at a thin, plausible margin. If the quote is enormous relative to the floor, you are paying for margin rather than for execution.

The patterns worth walking away from

Requests for a private key or seed phrase, quotes that omit network cost entirely, guaranteed trending or price outcomes, anonymous Telegram-only contact with no site of record, and pricing that does not change with routing target. Each one is independently sufficient reason to stop.
SignalWhat it usually means
Asks for a private key or seed phraseRequests control of your treasury. No campaign mechanism requires it.
Quote omits network costEither it is hidden and you pay it later, or the quote was never costed.
Guarantees trending or a price levelPromises an outcome that depends on a third party algorithm and on strangers trading.
Telegram account with no site of recordNothing to hold accountable and nothing to check afterwards.
Same price for every token and venueThe pricing was not derived from cost, so it cannot reflect what your campaign will actually spend.
Pressure to send within minutesCompressed decision time is a technique, not a business constraint.
The first row is the only one that can cost you everything at once. The others cost you the campaign.

The custody point deserves its own sentence because it is the one that is non-negotiable. A volume campaign works by funding wallets it derives itself. At no point does the mechanism require access to yours. Anyone telling you otherwise is describing a different transaction from the one you think you are entering. What non-custodial actually protects.

Six questions that separate operators quickly

Ask what failure rate they assume, which venue they route to and why, what the rent comes to at your wallet count, whether tips are included, what happens to leftover SOL, and what they do at graduation. An operator who has costed their own product answers all six without hesitation.
  1. What failure rate does your estimate assume?

    There is only one wrong answer, which is not knowing. Failed transactions are charged in full and are the single largest source of variance between a quote and a bill.

  2. Which venue will my token route to, and why that one?

    The answer should come from the mint rather than from a dropdown you filled in. Cost per landed swap differs by venue by an order of magnitude.

  3. What is the rent at my wallet count, and do I get it back?

    Roughly 0.00204 SOL per token account. At scale this is a real number, and it is held rather than spent, so the honest answer includes how it is recovered.

  4. Are Jito tips included in the quote?

    They cannot be, precisely. Tips are auction bids. An operator quoting a fixed tip is quoting a guess, and the correct answer is that tips sit on top and vary with demand.

  5. What happens to leftover SOL?

    It should return to the address it came from. Vagueness here is the single most common way small amounts disappear across many customers.

  6. What happens when my token graduates mid-campaign?

    The routing target changes. An engine that cannot follow it goes silent exactly when new attention arrives.

Volion answers all six on this site rather than in a private chat: the failure rates are published with sample sizes on the measurement page, the routing target is resolved from the mint in the console, rent and tips are separate lines in the pricing, and graduation handoff is a documented behaviour rather than a support request.

Why the impossible offers persist

Because most buyers compare quoted prices rather than computing floors, and because the costs a quote omits arrive later, separately, and feel like bad luck rather than a missing line in the original arithmetic. The remedy is to price the floor yourself before reading anyone quote.

The typical shape is a quote for a large volume figure at a price that sounds like a discount. It works because volume in SOL and cost in SOL are the same unit, so a reader who has not done the arithmetic has no intuition for the ratio between them. Once the floor is on paper, the same offer reads very differently.

The second shape is subtler and more common among otherwise legitimate operators: a quote that is real but computed against a network average failure rate. It is not dishonest, it is just wrong for your token, and it is wrong in the direction that favours the seller. This is exactly why Solana volume bot estimates on this site are computed per venue from sampled blocks rather than from a single global figure, and why a Pump.fun volume bot campaign against a live bonding curve is priced differently from the same campaign after graduation.

Nothing here requires trusting us either. The sampler reads public blocks with a public RPC method. Anyone can run the same measurement, and the method page describes exactly how, including the slot range of the sample the figures above came from.

One last test worth applying: ask which venue the quote assumes. A price built on a PumpSwap, Raydium, Meteora DLMM or Orca Whirlpool failure rate does not survive contact with a live Pump.fun bonding curve, and a quote assuming Jupiter routing is assuming the worst failure rate we measure. An operator who cannot name the venue has not costed your campaign.

Questions

How do I know if a volume bot quote is a scam?
Compute the cost floor for the volume it promises: swaps divided by the success rate multiplied by the median network fee, plus the venue swap fee on the volume, plus rent per token account. If the quote sits below that floor, it cannot be delivered as described.
Should I ever give a volume bot my private key?
No, under any circumstances. Campaign wallets are derived by the engine and funded from an address you send to. There is no mechanism in a volume campaign that requires access to your wallet.
Why do quotes vary so much between vendors?
Mostly because of what they omit. A quote covering only the service fee looks far cheaper than one that shows network cost, rent and tips separately, even when the total is identical or worse.
What is the biggest cost in a campaign?
Usually the venue swap fee charged on trade value, then the network fee including failed attempts, then rent. The service fee is often the smallest line, which is why comparing service fees alone is misleading.
Are Telegram-only operators always scams?
Not always, but there is nothing to check and nothing to hold accountable. The relevant test is whether the numbers they give you survive the floor calculation, and whether they answer the six questions above without hedging.
Can anyone verify your published failure rates?
Yes. They come from getBlock on finalized mainnet blocks, reading fee and error status per transaction and attributing each to a program. The method, the slot range and the sample size are published, so the measurement can be reproduced with any RPC endpoint.