Web3 is booming, but the question of where to store the data hasn't been fully figured out yet. The iteration of storage infrastructure is shifting from monopolization by big tech companies to decentralized solutions. However, relying solely on technology stacks isn't enough; the key is to solve the actual pain points of users.

The issues currently on the table are quite pressing. Traditional cloud storage giants control a trillion-dollar market, but the problems are clear—high costs, susceptibility to single points of failure, and inadequate user data privacy. These issues conflict with the demands of Web3 applications. Many project teams need a solution that is secure, controllable, and affordable all at once, but so far, none have fully achieved this.

Filecoin's static storage ensures data won't be lost, but what if you want to modify data? Sorry, the architecture doesn't support dynamic updates, and the costs are exorbitant—$200 per TB per year. Arweave offers permanent storage, writing data once and for all, suitable for long-term cold storage. However, for scenarios requiring frequent interactions, such as AI training data or blockchain game assets that are often modified, it becomes too rigid.

The Walrus protocol in the Sui ecosystem seems to be trying to break through from a different angle. It's not just about increasing storage capacity but aims to connect the storage layer with the computation layer—allowing data to be stored and directly accessed by smart contracts on-chain, reducing the costs of cross-layer data transfer. This approach is intriguing; although it still needs time to validate market response, it at least addresses the real needs of current Web3 applications.

In short, the competition for decentralized storage has only just begun. Centralized cloud services still hold an absolute advantage, but the demand for data sovereignty in the Web3 ecosystem is growing stronger. The future will likely feature a coexistence of multiple chains and protocols.