# Perceptron Plans Walrus Storage Shift for 800,000-Plus AI Network

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/38631
Published: 2026-10-23T03:19:50.000Z
Updated: 2026-10-23T03:19:50.000Z
Section: Technology

> The Thursday, Oct. 22 plan covers datasets from a network with more than 807,000 total nodes.

Perceptron plans to make Walrus the primary storage layer for datasets generated by its network of more than 800,000 nodes, a shift aimed at making stored AI data easier to check for changes.

The plan, announced Thursday, Oct. 22, concerns dataset storage rather than a migration of the network’s nodes. The network has more than 807,000 total nodes.

The network combines nodes that share unused bandwidth with contributors who provide data for artificial intelligence. Its data-questing platform supports structured contributions. The storage shift is intended to replace centralized storage, reduce friction from egress fees and make pricing more transparent.

Walrus assigns stored data an identifier derived from its contents and records that identifier on Sui. Comparing a dataset with its identifier can show whether the data changed after storage. Storage nodes also post proofs that they hold the data.

Those checks address whether a stored dataset matches its recorded version. They do not establish whether its information was accurate or collected appropriately before upload.

Walrus was created by Mysten Labs and is part of the Sui ecosystem. Another project, [DATA Foundation’s planned move to Sui](<https://www.tokenpost.com/news/technology/37431>), also names Walrus as the default storage layer for its applications.



## Links in this article

- [DATA Foundation’s planned move to Sui](https://www.tokenpost.com/news/technology/37431)
