Autonomous Worlds: The Engine Trying to Put Whole Games on Ethereum

Every online game you've played has an off switch. Somewhere a company runs the servers, and when it stops paying for them, the world ends. A small but stubborn group of developers has spent years asking: what if it didn't? What if the rules of a game world lived on a public blockchain, so the world kept running no matter what happened to the studio that made it?

They call these games autonomous worlds. And the most widely used toolkit for building them on Ethereum is a framework called MUD, made by a studio called Lattice.

Where the idea came from

The movement traces back to Dark Forest, the 2020 space strategy game that ran entirely on Ethereum and used zero-knowledge proofs to hide players' positions. (Our Dark Forest feature explains how.) Dark Forest proved a whole game could live on-chain, but it also showed how painful that was to build. Every developer was writing the same plumbing from scratch.

Around the same time, writers in this community began describing an "autonomous world": a world whose rules are enforced by a blockchain, that anyone can read, extend or build on, and that no single company can turn off. In the sense that matters, it's closer to a physical place than to a product.

What MUD actually does

Building a game on Ethereum is awkward because smart contracts weren't designed for game state. Storing thousands of units, tiles and inventories, and keeping a game client in sync with them, takes a lot of custom code.

MUD tackles that with a few ideas game developers will recognise:

• A shared database on-chain. Game data is stored in tables, like a spreadsheet, so developers don't have to design storage from scratch.
• Automatic syncing. The game client can mirror the on-chain state, so the screen updates when the world changes.
• Modular systems. Game logic is split into "systems" that act on data, similar to the entity-component-system (ECS) pattern used in many modern game engines.
• Open extension. Because the world's data and rules are public, other developers can add new systems on top, with permission rules the world defines.

In plain terms, MUD is to on-chain games what a game engine is to regular games: the boring foundations, so creators can focus on the fun.

What's been built

Lattice used its own tools to make games. OPCraft, an early experiment, was a voxel world on an Ethereum layer 2 with a familiar block-building feel. Sky Strife, Lattice's best-known game, is a real-time strategy game of short matches across floating islands, where every move is a blockchain transaction. Lattice also launched Redstone, its own layer 2 network built for on-chain games, to make those transactions cheap enough to play.

Other teams have used MUD for strategy games, social experiments and digital pets. Most are small, and many are closer to playable prototypes than finished products.

Why it's still hard

The honest list of problems:

• Speed. Every action waits for a block. Even on fast networks, that's a long way from the instant feel of a normal multiplayer game.
• Cost. Someone has to pay for every transaction, whether the player or the studio.
• Privacy. Everything is public by default, which makes hidden information, like fog of war, very hard to do without the kind of cryptography Dark Forest used.
• Players. Most people don't care where a game's state is stored. They care whether it's fun.

Why it matters

Autonomous worlds are the most idealistic corner of blockchain gaming, and also the one least obsessed with tokens. The goal isn't "earn money playing" but "build worlds that outlive their creators". Most experiments are small, but the ideas, open rules, player-made extensions and games that can't be switched off, address real frustrations every gamer has felt when a favourite game shut down.

For a parallel approach on another network, read our explainer on Dojo and on-chain games on Starknet.