Hello everyone, today I want to talk about a project that has been quietly working within the Sui ecosystem—Walrus.

To be honest, when seeing figures like a $2 billion valuation and $140 million in funding, many might think it's a bit exaggerated. But if you truly understand what the storage sector is doing, you'll realize that this project's approach is actually quite restrained. It’s not here to compete with all-in-one players like Filecoin or Arweave, but rather to identify two real pain points within the Sui ecosystem—expensive AI data and the difficulty of on-chain RWA (Real World Assets)—and to root itself there.

Since the mainnet officially launched in March 2025, it has stored over 113TB of on-chain data, with node counts exceeding 100. It may seem low-key, but the logic behind this steady growth is worth pondering.

**What exactly is it storing?**

This is the most critical question. Storage projects are everywhere, but only a few are truly implemented. Most projects shout "decentralized cloud" loudly, only to end up as jokes. Walrus’s cleverness lies in not being greedy—since the testnet phase, it has clearly focused on two directions: data needed for AI training and inference, and various data requirements for on-chain RWA (Real World Assets).

This isn’t aimless spending; it’s about finding what the market truly lacks. Over the past two years, data costs for AI have remained high, and on-chain storage for RWA has always been a bottleneck. Walrus is deeply cultivating these two tracks, gradually building its ecosystem.

From another perspective, the project’s strategy can be summarized as: leveraging Sui’s infrastructure advantages to push into the market, continuously improving features to feed back into the ecosystem, and then reducing storage costs in specific scenarios. This pragmatic approach instead gives it an opportunity to carve out a path amid the giants in the storage sector.