Arc Launchpads Compared: ARCLaunch, Arch, Sharc, ubi.fun, 5042, AstraPump and Bullcheese

Last verified: 14 Sep 2026

New to Arc launchpads? Start with the two kinds and how to choose.

Read the seven front ends and you will find the same screens: a form, a chart, a feed, a connect button. Read the seven mechanisms and you will find at least four genuinely different answers to one question. The front end is not where these venues differ. The liquidity is.

This page sets them side by side, every cell attributed to the venue's own site or a dated third-party source, with unknowns written as "not published" rather than guessed. One of the seven is Bullcheese, which TrustSwap built and operates, and which launches on 16 September 2026, the day Arc's public mainnet opens. A comparison whose author has an interest is only worth reading if the interest is declared and the sourcing is checkable.

There are two models on Arc, and the second one splits three ways

The first model is the bonding curve, familiar from Solana. A token trades inside a platform-held contract along a fixed pricing function. Only if it clears a threshold does it "graduate" onto a public DEX, at which point the accumulated liquidity migrates into a real pool. A small single-digit percentage of curve tokens ever graduate, by most published estimates. The curve's virtue is that it needs no parameters from the creator; its cost is that the graduation threshold is a filter most tokens die against, and the migration itself can gap, stall or be front-run.

The second model is the single-sided concentrated pool, which Arc's venues reach for because Uniswap's concentrated-liquidity contracts are due on the chain from mainnet. A Uniswap v3 position is liquidity assigned to a price range rather than a curve. Set that range wholly above the current price and the position holds only the token, since a range the market has not yet reached contains no USDC by definition — so it can be funded with the token alone. Buyers walk price up through it. No counter-asset, no threshold, and a public pool from the first trade.

Where the single-sided venues diverge is the lock. "Burned" destroys the position and, with it, the LP fee stream, permanently. "Permanently locked" means the position cannot be withdrawn — whether its accrued fees remain claimable is a separate design decision a venue has to state in writing, and not all of them do. "Locked" through a third-party vault means the position is held for a defined term by a contract someone else audited, with fee collection separated from withdrawal. Three phrases that look interchangeable on a landing page and are not interchangeable in a contract.

The table

Every cell below reflects what the venue publishes about itself, read on the date shown. Blank claims are not filled in from inference.

More than half of that grid reads "not published," and that is the single most useful thing in it. A venue that has not published its fee share, its lock term or its audit before mainnet is not hiding a bad answer so much as telling you the answer is not yet a commitment.

For a creator, one column decides it and it is not the fee percentage

The headline number every venue eventually advertises is the creator's cut of trading fees. It is the wrong column to read first, because a percentage of a stream that has been destroyed is nothing.

Read "LP treatment" and "fees claimable during lock" together, in that order. If the position is burned at graduation, the LP fee stream ends there regardless of what share the creator was promised on the way in. If the position is permanently locked and the venue has not published how accrued fees are treated, the creator is accepting an unbounded term on an undefined stream — which may be exactly what a fair-launch purist wants, but it is a decision to make knowingly rather than discover in month four. If the position is locked for a defined term through a third-party vault with fee collection separated from withdrawal, the stream survives the lock and the percentage becomes a number that means something.

Only then does the percentage matter. The second-order question after that is what the creator holds when the term ends: a permanent lock returns nothing, ever, by design; a term lock returns a decision.

For a buyer, one column decides it and it is not the chart

A buyer's version of the same discipline is to read "public DEX from block one" before anything else. A token inside a bonding curve is not yet an asset with a market; it is an entry in a platform-held contract that may or may not become one. Until graduation there is no pool for an aggregator to route through and no chart that reflects anything but the curve's own function.

Once a token is in a public pool, the buyer's checks are on-chain rather than on-site. The pool address and the contract are readable on Arcscan. The lock, if there is one, is a transaction and a contract state with an expiry field, and how to verify a liquidity lock on Arc shows what a lock record looks like when it exists and what a venue's page looks like when it only claims one.

The venue's marketing is not evidence for any of this. The chain is.

How this table was compiled, and who compiled it

TrustSwap built Bullcheese, operates it, and earns from it. That interest does not disqualify this page; it obliges the page to be checkable, so here is the method in full.

Each competitor row was read off that venue's own live site on 14 September 2026 and quoted rather than summarised wherever the site states something directly, one quotation per source. Where a site did not render its body content — ARCLaunch, on the day — the row is sourced to dated third-party press material and labelled as such. Where a venue publishes nothing on a column, the cell reads "not published"; nothing is inferred from a competitor's silence, and no negative claim about any venue appears on this page that the venue's own source does not state. Status is reported as the venue's site presents it on the check date, not as a claim about mainnet activity we confirmed ourselves — which is why the live-looking venues are described as appearing live per their own sites.

Bullcheese's row is held to the same rule, and it is where the rule bites hardest. "Audited" attaches to Team Finance's locking contracts, which Certik, Hacken and BailSec have audited, and to MintPlus, audited by SolidProof. It does not attach to Bullcheese's own launch contracts, and the cell says so. Until Bullcheese's own documentation pages are live, the row is sourced to TrustSwap as operator, and the source cell will move to those pages the day they exist.

If a cell here is wrong, it is falsifiable in about ninety seconds against the linked source. That is the point of building it this way.

The strongest objection is that this table measures the wrong thing

Here is the case against everything above, made as well as we can make it.

Liquidity mechanics are a creator's concern and a lawyer's concern. They are not what determines whether a launch works. What determines that is attention — how many people see the token in its first hour, whether the venue's audience has money on the chain already. Measured that way, a venue with a curve, a burn and an enormous user base beats a venue with an immaculate lock and no traffic every time, and the creator who optimised for fee claimability on a token nobody bought has optimised a rounding error. Pump.fun did not win Solana on lock design. On that reading, a table of mechanism columns is a document written by infrastructure people for infrastructure people, and the market is somewhere else.

That argument is right about the first week and wrong about everything after it. Distribution decides which tokens get bought; mechanism decides what a buyer can check before buying and what a creator is left holding once the attention moves on. Those are sequential considerations, not competing ones — and on a chain whose gas, quote asset and pool contracts are all public and all denominated in the same unit, the cost of checking has collapsed to near zero. When verification is cheap, opacity stops being a neutral default and starts being a signal.

What happens when the table fills in

The interesting thing about this page is not its current state. It is that most of it is blank, and blanks are perishable. Every venue here will publish a fee share and a lock term eventually, because creators will ask and because the first venue to answer makes the silence of the others conspicuous.

So the question for the next quarter is not which of these seven wins. It is which of them is still describing its lock in a marketing sentence rather than a contract when the first Arc creator goes looking for their fees — and whether, by then, anyone is still willing to launch into a mechanism they cannot read.

Bullcheese is a permissionless launch venue. Tokens launched on it are created by anyone, carry no endorsement, and can go to zero. Nothing here is financial advice.