Recently, I’ve been studying how Walrus and io.net work together, and it’s quite interesting. In simple terms, one handles computation, and the other handles privacy storage. This combination significantly reduces costs for AI startups.

Real-world examples include: DLP Labs using this solution to process electric vehicle log data, directly reducing costs by 22%; CudisWellness uses it to manage health data, ensuring privacy while also allowing profit sharing. This actually hits the pain point of the entire "data economy"—data needs to be active rather than locked in one place, and users can access it flexibly like with tools such as Seal.

From a broader perspective, this approach is somewhat like building an "AI full stack"—Sui as the foundation, with over 190 projects ranging from concept to product operation. On the anniversary of the mainnet launch, Seal’s daily request volume exceeds 80,000+, indicating that users truly need this. Even more interesting is the shift in logic: privacy protection is no longer about hiding things but about transparency under compliance—125 nodes with a utilization rate of 28%, these numbers are enough to show that the entire system has entered a verifiable operational stage.