Some time ago, a leading exchange's CreatorPad event offered a nice reward, and the task process was quite easy—posting and trading interactions. Many people were just trying to exploit the system for free rewards; I started with that mindset too. But the more I used it, the more hooked I became, and I really developed an interest in Walrus Protocol.

The most immediate feeling is the cost. I tried uploading a training file for an AI model over 2GB in size, and it cost less than $0.4, with a very fast speed. If you switch to AWS cloud storage, the same capacity would cost dozens of dollars per year. The key point is that you never know if your data is being used to train their models unless you have solid evidence, and without that, you can't prove that the data was originally created by you.

Walrus's core logic is actually quite clever—it splits files into pieces and distributes them across multiple nodes. This approach reduces costs and significantly enhances security. Not long ago, Tusky was about to shut down (data migration deadline: January 19), and Walrus stepped in directly to help the community receive NFT metadata. Major projects like Pudgy Penguins remained intact, with no data loss.

Even a16z mentioned Walrus in their 2026 outlook report, calling it a key infrastructure in the Sui ecosystem for privacy and storage. Such endorsement is not given lightly.

What's more interesting is that Sui plans to launch private transaction features in Q1 2024—transaction records can be hidden and do not need to be publicly on-chain. Paired with Walrus and Seal's privacy tools, sensitive data like medical records and travel trajectories can be safely uploaded to the chain, accessible only to authorized persons. DLP Labs has already implemented this solution in practice. From concept to deployment, the ecosystem's progress is indeed impressive.