Have you ever wondered why so many NFT projects, AI applications, and on-chain games are struggling with data storage? The simplest and most straightforward approach is to use centralized cloud services like AWS or Alibaba Cloud, but the problems that come with that are obvious—costs are exorbitant, security is hard to guarantee, and services can be ruthlessly shut down. In the Sui ecosystem, there is a project called Walrus that does exactly this: building a decentralized file storage system to completely change the game.

From a technical perspective, Walrus employs a scheme called erasure coding. Simply put, it breaks your files into small pieces and disperses them across multiple nodes in the network. Even if some nodes go offline or encounter issues, the system can fully recover your data. With just 4 to 5 times redundancy, it ensures foolproof data integrity. Compared to other decentralized storage projects, this scheme can reduce costs by over 80% while maintaining speed.

What’s even more interesting is that Walrus supports programmable storage. Data can directly interact with on-chain smart contracts—you can set who can access it, when to update versions, or even let data generate revenue through contract logic. This opens up a lot of possibilities for applications like NFTs, AI models, identity systems, and more.

Since its mainnet launch in March this year, Walrus’s growth has been visibly rapid. Major projects like Pudgy Penguins use it to store high-resolution NFT images, AI training teams manage massive datasets with it, and identity verification tools have also integrated it. It has already attracted deployment from over a hundred projects.

Behind this is the team at Mysten Labs, the founders of Sui, operating the project, with a top-tier investment lineup including firms like a16z. According to the roadmap, by 2026 they plan to support mainstream chains like Ethereum and Solana, and delve deeper into AI, aiming to make Walrus a true Web3 data infrastructure.