DATA Network Moves to Sui, Making Walrus Its Default Storage Layer

The DATA Network is moving onto Sui, pairing its Trace provenance registry with Walrus storage in a bid to connect AI developers with data contributors and the rights attached to their files. The change shifts the network from an independent chain to a protocol running natively on Sui; Walrus will become the default storage layer for applications operated by DATA Foundation.

The proposed setup links three functions: Trace records where a contribution came from and what rights apply, Walrus holds an encrypted reference copy of the underlying file, and the DATA Network supports registration and related protocol activity. For AI labs, the intended result is a way to source real-world data with provenance records and access controls, rather than relying only on information collected from the open web.

Why provenance matters for training data

DATA Foundation frames the move around a constraint facing AI development: useful training material increasingly comes from people and workplaces, not just publicly available websites. Examples include a nurse describing a patient’s night, dashcam footage of a commute in Jakarta, a technician diagnosing equipment, or a customer call in Bengali. Such material may be personal or subject to restrictions, and contributors retain an interest in how it is used.

Trace is designed to record each contribution’s origin and attached rights, while settling payment to the person who supplied it. DATA Foundation launched the registry in June. Since then, it has registered more than 250 million contributions and processes an average of over 3 million receipts per day, according to the announcement.

Applications that write to Trace include Kled, described as the largest opt-in human data marketplace, which has over 1.5 billion records being onboarded. Numo and Oto are also named as applications that already license data to frontier AI labs and robotics companies. Those figures and descriptions indicate the scale the network is targeting, though the announcement doesn’t provide details on how many records are fully onboarded or how licensing volume is distributed across applications.

What changes with Sui and Walrus

Sui represents assets as programmable objects, a model that DATA Foundation says can support on-chain management of datasets and their licenses. The network’s parallel transaction execution and settlement in under a second are intended to accommodate high-volume registrations. The announcement says the system can handle millions of daily registrations, but does not specify performance measurements or conditions for that figure.

Walrus adds storage for large files, alongside controls intended to restrict readability to licensed buyers. The integration also changes what a Trace record can verify. Today, a record shows that a file was registered and when. With Walrus, Trace is expected to retain an encrypted, immutable copy of the contributed file as a reference. The file’s hash can then be regenerated from the data itself, rather than relying on metadata supplied by an application. Later processing can be checked against that reference.

That distinction matters for auditing: a provenance entry tied to the file itself can provide a stronger basis for checking whether a stored or processed dataset matches the contribution originally registered. It doesn’t, by itself, establish that a particular use is permitted; the rights recorded for each contribution and the access controls around the data remain central to the system’s stated purpose.

Network migration and token transition

The DATA Network’s independent chain will be retired once the upgrade is complete. The $DATA token will continue as the DATA protocol token on Sui, with holders’ positions carried over one-to-one. The announcement says holder economics and governance rights will be at least as strong as they are today, and that locked and vesting positions will transfer unchanged.

The existing network will keep operating until the migration is complete. A detailed timeline and holder transition plan are due before the upgrade. Until those details are published, users and application teams will need to wait for specifics on the transition process and any operational steps required of them.

Sui Foundation is investing $30 million in the DATA Network, which the announcement describes as the network’s largest institutional stake. PIP Labs and DATA Foundation will continue to hold positions. SY Lee, Chairman of Poseidon and CEO of PIP Labs, is also joining Sui Foundation as a Strategic Advisor.

What the teams say comes next

Adeniyi Abiodun, Co-Founder and Chief Product Officer at Mysten Labs, said the migration brings large volumes of rights-cleared data already serving frontier AI labs into the Sui and Walrus stack. SY Lee said Trace is intended to combine consent and provenance with a way to pay contributors. Andrea Muttoni, CEO of DATA Foundation, pointed to encrypted reference storage as an extension beyond an audit trail, and said the teams may explore verifiable trusted execution environments for evaluations and fine-tuning in the future.

Those possible additions are not described as part of the current upgrade. For now, the concrete changes are the move to Sui, Walrus storage for DATA Foundation applications, and a tighter link between provenance records and the files they describe. The migration plan will determine how those functions carry across for holders and applications.

About DATA Foundation

DATA Foundation supports the DATA Network, a protocol for tracking the provenance of real-world data used by AI models. Its Trace system registers contributions with provenance receipts and audit trails. Applications building on the network include Kled, Numo and Oto. Formerly Story Foundation, the organization refocused on AI training data in June 2026.