When centralized storage hits a ceiling, distributed solutions are emerging. Walrus is such an option—built on Sui, using distributed Blob storage to host videos, images, and large-scale datasets.

Rather than being a new project, it’s more of a new approach. Files originally stored in one place are broken apart and distributed across a global network of nodes. The benefits are clear: erasure coding technology allows data to be fully recovered despite network fluctuations, and high availability no longer depends on a single central node. More importantly, the cost—roughly one-tenth of centralized solutions.

The testnet launched in early 2025 makes this approach more concrete. Developers can upload and retrieve data via API without being bogged down by underlying complexities. From node validation to governance design, Walrus seems to be paving a long-term path for data.

Ecological collaboration is also gradually expanding. The Walrus Developer Platform has become a hub—AI researchers upload training data, game developers store high-definition assets, and different roles collaborate and circulate on the same platform. While there are ecological incentives, they are not loud; rather, they serve as necessary fuel to keep the system running.

The next step is WalrusTrade, which aims to bring compliant trading and sharing of distributed datasets into practical application. The waiting list opened in January is already attracting the first explorers.

This path may not be very noisy, but it is sufficiently broad.