10,000 spheres. Merge them and they grow heavier - and harder to move. Drag the dial and watch the math.
Mass is what a sphere looks like, it grows every time you merge. Weight is what a sphere earns, and it climbs in discrete steps, deliberately slower than mass. One giant sphere is not worth the same as its equivalent in small ones. That gap is the entire point.
Most merge collectibles use a single number for everything: how rare a piece looks, how much it's worth, how much it earns. That works fine until yield enters the picture. The moment a merged asset pays out proportionally to its full size, the optimal strategy becomes obvious: buy everything, merge it into one object, and collect a share of the pool that no individually-held sphere could ever match. The collection collapses into one wallet.
Atlas breaks that link on purpose. Mass still grows linearly, merge two mass-4 spheres and you get roughly one mass-7.36 sphere (before the 8% burn even factors in, 8 × 0.92). But weight is read from the tier table, not calculated from mass directly, and each tier band is deliberately wide. A mass-257 sphere earns 18x. Fifty-seven separate mass-1 spheres earn 57x combined. Spreading is mathematically rewarded; consolidating is not.
Take ten mass-1 spheres. Held separately, they sit at Tier 1 each: 10 tokens × 1x weight = 10x total, for 10 × 500 = 5,000 $ATLAS in activation cost.
Merge all ten into one sphere instead (nine merges, each losing 8% to the burn) and you land at roughly mass 8.07, still Tier 2. That's 3x weight for 2,000 $ATLAS in activation cost, a third of the yield for less than half the cost. You saved on activation. You lost far more on weight. Neither choice is "wrong", they serve different goals, visual rarity versus reward efficiency, and the Merge Lab below lets you test the trade-off with your own numbers before committing anything on-chain.
| Tier | Mass range | Weight | Activation cost | Burned / Staked |
|---|
Weight multipliers were chosen so that going up one tier roughly triples or quadruples the mass required, while only doubling the weight earned. That widening gap is what makes spreading the mathematically stronger play at every tier, not just the extremes.
Merging isn't free, and it isn't meant to be. Every time two spheres combine, a slice of their combined mass is destroyed on the spot, converted into $ATLAS, and dropped straight into the reward pool for everyone still holding smaller, separate spheres. It's the one place in the protocol where a holder's own choice directly funds everyone else's yield.
When two spheres merge, their combined mass isn't preserved, 8% of it is burned on contact, permanently, with no monetary transfer attached to the mass itself. The flat 0.001 ETH protocol fee charged on top is the actual revenue: it joins the vault's ETH balance and, once a buyback vote executes, becomes part of the reward pool, see Revenue and Governance below.
Every merge makes the collection smaller and the pool a little heavier. It's a cost the mergers pay, and a dividend everyone else collects, whether or not they ever merge anything themselves.
mass 1 + mass 1 → mass 1.92, not mass 2
The burn rate is fixed at 8% no matter how big either sphere already is, there's no volume discount for consolidating late. Ten mass-1 spheres, merged down to one, one at a time, land here:
| Merges | Resulting mass | Tier |
|---|---|---|
| 1 sphere, no merges | 1.00 | 1 |
| after 3 merges | 3.65 | 1 |
| after 6 merges | 6.16 | 2 |
| after 9 merges | 8.07 | 2 |
Each row assumes every prior merge already happened. The mass climbs quickly at first and flattens fast, the 8% burn compounds with every step.
Set a starting mass for each sphere, then drag them together. This is the exact formula the contract runs, nothing here is simulated separately from what's described above, it's the same math, just played out with your own numbers instead of ours. Feed the same sphere over and over and watch how quickly the tier gains slow down relative to the mass you're pouring in.
Want to try this on spheres you actually own? The live merge tool moved to the App, along with activation and your full sphere list.
Owning a sphere earns nothing by itself. It has to be activated, a deposit of $ATLAS, sized to its tier, that turns a passive mass value into a claim on the reward pool. Every buyback, that pool splits by weight share among everyone currently activated.
Activating a sphere costs the $ATLAS listed for its tier, 500 for Tier 1, up to 100,000 for Tier 5. Of that, 8% is burned permanently and 92% is held as a stake you can withdraw any time by deactivating. There's no lockup period and no penalty for changing your mind, deactivating simply stops that sphere from earning until you activate it again.
An inactive sphere costs nothing to hold. It doesn't earn anything either. The whole system is opt-in: you decide, per sphere, whether the yield is worth the $ATLAS you'd have tied up.
No single wallet can claim more than 5% of a given buyback's holder pool, regardless of how much mass or how many spheres it holds. If your share would exceed that, you're paid the capped 5%; the rest isn't redistributed to others in that round, it simply stays in the vault's balance as a buffer.
In practice this matters most for anyone chasing Tier 5 through aggressive merging. A single mass-2000 sphere earns the same fixed 18x weight as a much smaller Tier 5 sphere would, weight doesn't keep climbing past the tier threshold, and if that 18x still happens to exceed 5% of total active weight, the cap binds before the tier math even becomes the limiting factor.
Want to activate a sphere you actually own? That tool moved to the App, along with deactivation and your full sphere list.
Try your own numbers, including what happens if you push past the whale cap. Nothing here touches the chain, it's the same formulas run locally.
The buyback pool figure above is adjustable so you can test different scenarios, it isn't a live number. The actual pool size each buyback depends entirely on how much merge and royalty activity accumulated since the last one, see Revenue below for how that's calculated.
Minting a sphere is 0.005 ETH, paid once, and it goes entirely to the team, it never touches the reward pool. Holding a sphere costs nothing. The pool is fed only by usage: merge fees and secondary royalties, split every buyback. The two revenue streams are kept structurally separate on purpose, so that what funds the team's launch capital and what funds ongoing holder yield never compete with each other.
10,000 spheres at 0.005 ETH is 50 ETH in total mint revenue, paid entirely to the team treasury as one-time launch capital. It doesn't get split, doesn't get vested to holders, and doesn't recur. A holder who mints and never touches the protocol again has paid the team once and owes nothing further, and is owed nothing further either.
Everything after mint is usage-based. Every merge charges a flat 0.001 ETH protocol fee on top of the 8% mass burn. Every secondary sale pays a 5% royalty. Both accumulate in ETH, they aren't converted to $ATLAS automatically or on a schedule, that conversion only happens through the buyback vote described below.
| Source | Rate | Who pays it |
|---|---|---|
| Mint | 0.005 ETH / sphere | Minter, once, 100% to team |
| Merge protocol fee | 0.001 ETH | Whoever initiates the merge |
| Merge mass burn | 8% of combined mass | Whoever initiates the merge |
| Secondary royalty | 5% | Buyer or seller, per marketplace |
This split is applied every time a buyback executes, to whatever ETH the protocol has collected since the last one. No wallet can claim more than 5% of the active pool from a single buyback, no matter how much mass it holds. If a buyback hasn't triggered in a while, the ETH doesn't disappear or expire, it just sits in the vault waiting for the next one.
The ETH sitting in the vault doesn't convert to $ATLAS on its own, or on a timer, or at the team's discretion. It converts when activated holders vote to make it happen. This is the one lever in the protocol that's entirely in holders' hands.
Once a day, at a fixed time, a one-hour voting window opens. Anyone with an activated sphere can cast a yes vote, weighted by their active weight at the moment they vote, one vote per wallet per accumulating cycle. If enough weight votes yes to cross 33% of total active weight, the buyback executes immediately: the vault's ETH swaps for $ATLAS on the open market and splits 70/30 as described above.
A sphere only counts toward voting weight if it was activated before the window opened, activating in the same block as a vote doesn't work. This closes off flash-loan style voting, where someone could otherwise borrow $ATLAS, activate a pile of spheres, vote, and unwind it all in a single transaction.
Nothing is lost. The yes-votes already cast don't reset, they carry into the next day's window and keep accumulating until 33% is crossed. A quiet week doesn't erase a quiet week's votes, it just means the buyback happens a few windows later than it might have. The ETH keeps accumulating in the vault the entire time.
Drag to see what accumulated yes-weight looks like against the 33% line, and what would happen at that level.
Both extend the core loop above rather than starting a new one, in active development alongside everything else, full detail is in the whitepaper.
A native marketplace where the 5% royalty isn't optional, unlike most external marketplaces. Trading fees feed the same reward pool as everything else. The interesting part: a "buy-and-merge" action, purchase someone's sphere and merge it into yours in one step, built entirely at the marketplace layer so the core merge mechanic stays exactly as simple as it is today.
A bonding-curve launchpad in the spirit of Atlas's own mint, no presale, no team carve-outs. Every graduating token's liquidity locks permanently through the same LPLock design protecting $ATLAS's own pool. Posting a bond to launch works close to any standard launchpad, that part isn't reinvented, activated sphere holders can then promote a launch with their weight (the same mechanism already used for buybacks), and $ATLAS-paired launches get priority placement by default. Holders boost visibility, they don't gatekeep who's allowed to launch.
Every sphere starts identical: mass 1, Tier 1, zero history. There's no presale tier that mints at a discount and no reserved allocation that starts heavier than everyone else's. What a sphere becomes after that, merged or kept small, activated or left dormant, is entirely a function of what its holder does with it.
Spheres go live first, on their own. Mint and merge work immediately. $ATLAS doesn't exist yet, so there's nothing to activate and no rewards to earn, this phase is purely about getting the collection out and letting people start merging.
Once mint proceeds are in, a quarter of that ETH funds a $ATLAS/WETH pool on Uniswap V4 directly, no launchpad cut taken off the top. The resulting LP position locks permanently in its own contract, LPLock, no function anywhere can ever withdraw it, only its trading fees stay collectible. A 5% fee applies to sells only, via a V4 hook, split 80% to holders and 20% to the team. This is what makes $ATLAS actually tradeable, and what the buyback votes will swap against later.
With $ATLAS live and liquid, activation, epochs, and the buyback vote turn on. This is when holding a sphere starts being able to earn something, not before.
500 spheres (5%) are reserved for the team, minted the same way as everyone else's, mass 1, Tier 1, no head start. No presale tiers and no discounted pricing for anyone. What a sphere becomes after mint, merged or kept small, activated or left dormant, is entirely up to whoever holds it.
Nothing happens, and nothing is owed. You hold a mass-1, Tier 1 sphere indefinitely, it earns no yield, but it also costs nothing beyond the original mint, and there's no expiry or decay on an inactive sphere.
Yes, 92% of it. Deactivating a sphere at any time returns the staked portion. The 8% burned at activation is permanent, the same way the 8% burned on a merge is permanent, it's the cost of participating in the reward pool at all, not a fee you pay repeatedly.
Burning mass is a permanent size reduction with no monetary transfer attached to it, it's what makes merging costly. The actual revenue from merging is the separate flat 0.001 ETH protocol fee, which joins the vault's ETH balance and reaches the reward pool once a buyback executes. Burned $ATLAS from activation is removed from circulating supply entirely, it isn't redistributed. The two burns serve different purposes: one is a cost of merging, the other tightens $ATLAS supply over time.
Every cap is a chosen number, but the reasoning isn't arbitrary: without a cap, the sub-linear weight tiers slow down consolidation, they don't stop it. A wallet with enough capital can still out-merge and out-activate everyone else combined. The cap is a hard backstop underneath the tier math, not a replacement for it.
That's the intent, and the Merge Lab above runs the identical formula client-side so you can check it against your own numbers before anything is final. The tier table, burn rate, and split percentages shown here are the current design; if any of them change before deployment, this page will be updated to match, not the other way around.
An automatic buyback on a timer would trade at whatever price happens to exist at that exact moment, good or bad, with no one deciding it's actually a good time. Gating it behind a 33% quorum vote means it only executes when a real, weighted majority of active holders choose to. The ETH doesn't expire while waiting, so there's no cost to being patient.
500 spheres (5% of supply) and 5,000,000 $ATLAS (5% of supply), both disclosed here rather than left out of the numbers. The team's spheres mint the same way as everyone else's, mass 1, Tier 1, no discount. The team's $ATLAS vests linearly over time rather than unlocking all at once. Beyond that, the team's ongoing income is the 30% side of every buyback split, same mechanism everyone else's 70% comes from.
Nothing on this page is financial advice, an offer, or a guarantee of return. Reading it here doesn't replace reading the contract once it's deployed and verified.