An Axie Infinity clone copies the design of Axie Infinity, the first play-to-earn game with mainstream reach: creature NFTs, a breeding system, two tokens and a dedicated chain. This guide explains how each part of Axie Infinity worked, what failed in 2022, and which parts a new NFT game marketplace should copy.
The Axie marketplace was the center of the game economy. Axie Infinity is also the best-known case of a reward token that lost almost all its value, and its bridge lost hundreds of millions of dollars to attackers.
How Axie Infinity's economy worked
Sky Mavis, a Vietnamese studio, released Axie Infinity in March 2018. A player needed three Axies, the game's creature NFTs, to start playing. Each Axie had a class, body parts and stats, which set its abilities in turn-based card battles.
- Axies were ERC-721 tokens, and each body part had a visible trait and hidden genes that could pass to offspring.
- Smooth Love Potion (SLP) was the reward token. Players earned SLP in battles and spent it on breeding.
- AXS was the governance and staking token, and breeding also required an AXS fee.
- Breeding combined two Axies to mint a new one. Each Axie could breed a limited number of times, and the cost rose with each breed.
- Land plots in the game world of Lunacia were sold as separate NFTs.
The marketplace connected these parts. New players bought starter teams from existing players, breeders sold offspring, and players sold the SLP they earned. Every seller on the marketplace depended on new players arriving to buy.
Ronin: a chain built for one game
Ethereum gas fees made small in-game transactions too expensive, so Sky Mavis built Ronin, an EVM-compatible sidechain for Axie, and launched its mainnet in February 2021. Trading and breeding on Ronin became cheap and fast, and Sky Mavis controlled the whole user experience.
Sky Mavis later opened Ronin to other games, and the farming game Pixels moved from Polygon to Ronin in late 2023. Ronin's documentation says Ronin returned to Ethereum as a Layer 2 network in 2026.
Scholarships
Starter teams became expensive in 2021, and owners responded with "scholarship" programs. An owner lent Axies to a player, often in Southeast Asia, and the two split the SLP that the player earned. Guilds formed to run scholarships with thousands of players. Scholarships lowered the entry cost, but they also showed that many players played for income, not for fun.
What broke in 2022
- Token inflation. Battles created SLP faster than breeding used it up. When new-player growth slowed, fewer people bred Axies, so demand for SLP fell. SLP lost more than 99% of its peak value, and scholars' earnings fell with it.
- The Ronin bridge hack. On March 23, 2022, attackers took control of enough validator keys on the Ronin bridge to withdraw about $620 million in ETH and USDC. The FBI named North Korea's Lazarus Group as the attacker on April 14, 2022. The bridge had a small validator set, and a few parties held most of the keys.
- Players came to earn. When rewards fell, many players left, and average daily players fell from about 2.7 million at the peak to about 250,000, according to a Reuters report from April 2023.
Sky Mavis cut SLP emissions, reworked rewards, added validators to the bridge and repaid users who lost funds in the hack. It also released Axie Infinity: Origins, a version of the card game with more focus on gameplay.
What an Axie Infinity clone should copy
- Creature NFTs with game traits. Traits that change gameplay give an NFT a use beyond collecting.
- Breeding limits. A breed count per creature and a rising cost per breed both slow supply, so model the limits to hold under slow player growth.
- A low-fee chain from the first day. Use a Layer 2 network, an app-chain or an existing gaming chain, so players do not pay Ethereum mainnet gas to breed or trade.
- An in-game marketplace. Trading inside the game lets you show filters for body parts, class and breed count, which general marketplaces do not offer.
What an Axie Infinity clone should avoid
- Rewards with no outside revenue. If rewards come only from new players' purchases, the economy shrinks when growth slows.
- A weak bridge. A bridge usually holds more value than any other contract in a game. Use a large, independent validator set or a well-audited canonical bridge, and set withdrawal limits.
- Earnings as the main pitch. Marketing should present the game first. Earnings should come from skill or ownership and should not be the reason to play.
| Design choice | Axie's version | Safer option |
|---|---|---|
| Reward token | SLP created for play with no cap | Capped or season-based rewards, or no reward token |
| Breeding cost | Fees in SLP and AXS | Fees partly burned, plus items that breeding uses up |
| Chain | Own sidechain with a small validator set | An established Layer 2 or app-chain with strong bridge security |
| Entry cost | Buy three NFTs | Free starter creatures, with NFTs for upgrades and cosmetics |
Games like Axie Infinity covered on this site
Several other NFT games share parts of Axie's design, and each has its own guide here:
- ZED RUN used horse NFTs with bloodlines, breeding and paid races, and its successor, ZED Champions, runs on Base.
- Gods Unchained is a trading card game where players own their cards as NFTs on Immutable.
- Decentraland sells land parcels and wearables in a virtual world run by a DAO.
- Polkacity sold city assets such as taxis and gas stations that paid rewards to their owners.
The guides to metaverse NFT game development and Web3 game development cover the game build itself, and the IGO launchpad guide covers token sales that fund a game.
Frequently asked questions
Can I copy Axie's two-token model?
You can. Axie's reward token failed because breeding used up far less SLP than battles created. If you use two tokens, model supply and demand under slow player growth, not only under launch demand.
Does an Axie Infinity clone need its own chain like Ronin?
Almost never at launch. Existing Layer 2 networks and gaming chains give low fees without the need to run validators or a bridge. An app-chain makes sense only for a game with a very large number of transactions.
Where should game NFT data be stored?
Keep fixed traits on-chain or in content-addressed storage, and keep fast-changing game data, such as levels and experience, on the game server. Write that data on-chain only when it matters for trading.
Is play-to-earn still viable?
Games where income is the main reason to play have mostly failed. Studios still build games that use NFTs for item ownership, usually with small token rewards or none.