DATA Network joins Sui and Walrus to scale the supply chain for AI training data

‍Main Takeaways‍

• The DATA Network and Trace, the onchain registry for rights-cleared AI training data, are joining the Sui stack, with Walrus as the default data layer.
  
  
• AI labs get one place to source real-world training data they're cleared to use, with the provenance regulators increasingly require.
  
  
• Sui and Walrus operate at machine speed and scale, turning frontier AI training data into programmable, access-controlled, and verifiable assets.
  

Sui Foundation and The DATA Foundation today announced that the DATA Network and its Trace registry are joining the Sui stack, bringing a live supply of rights-cleared AI training data onto the infrastructure Sui has built for AI workloads. Walrus, the data platform with integrity proofs anchored on Sui, becomes the default storage layer for applications operated by DATA Foundation. Separately, the DATA Network will be upgraded to run as a protocol natively on Sui.

The upgrade gives AI labs a single place to source real-world training data they are allowed to use, with its provenance recorded by Trace and the original file stored and access-controlled on Walrus.

The data AI needs next

The next generation of AI models is constrained by access to real-world data. The open internet has been scraped, and the data that comes next must be sourced from people, with consent and with provenance that regulators and legal teams will accept. The EU AI Act and ongoing litigation over training data have made provenance a procurement requirement for labs.

The data those models need sits with people rather than on the open web: a nurse describing a patient’s night in her own words, a dashcam recording of a rainy commute in Jakarta, a factory technician’s hands as they diagnose a fault, or a customer call conducted in Bengali. It is personal and often regulated, and it belongs to the people who generated it.

Trace, launched by DATA Foundation in June as an onchain registry for AI training data, was built for exactly this kind of data. It records the origin of each contribution and the rights attached to it, and settles payment to the person who contributed it. It has registered more than 250 million contributions since launch and now processes over 3M receipts per day on average. Applications writing to Trace include Kled, the largest opt-in human data marketplace with over 1.5 billion records being onboarded, along with Numo and Oto, applications that already license data to frontier AI labs and robotics companies.

Why Sui and Walrus

Sui represents every asset as a programmable object, allowing datasets and their licenses to be managed onchain like tokens. Independent transactions execute in parallel and settle in under a second, making the network capable of handling millions of daily registrations. Walrus extends this capability to large files: it stores massive AI datasets efficiently and ensures data is only readable by licensed buyers. No other network offers storage, access control, and ownership tracking built natively into a single ecosystem, together with verifiable compute.

Trace now adds a flagship supply of rights-cleared data, with existing demand from the AI labs that license it today. The result of the integration will be stronger provenance guarantees and tamper evidence. Today a Trace record proves that a file was registered and when, and with Walrus, Trace will be able to hold an encrypted, immutable copy of the contributed file itself as a reference guarantee. The hash of the record can be regenerated from the data itself rather than from metadata an application supplied, and any later processing of the file can be checked against the original.

The upgrade is backed by a long-term strategic investment. Sui Foundation is investing $30 million in the DATA Network, establishing the largest institutional stake in the network, alongside the continuing holdings of PIP Labs and the DATA Foundation. Furthering the alignment, SY Lee, Chairman of Poseidon and CEO of PIP Labs, joins Sui Foundation as Strategic Advisor.

The protocol upgrade

DATA Network will be upgraded to run as a protocol natively on Sui, and its independent chain will be retired. $DATA continues as the token of the DATA protocol on Sui, carried over one-to-one, with holder economics and governance rights at least equivalent to today. Locked and vesting positions carry over unchanged. In parallel, Walrus becomes the default storage layer for DATA Foundation applications.

Adeniyi Abiodun, Co-Founder and Chief Product Officer at Mysten Labs, the original contributor to Sui and Walrus, said: “The AI industry has reached the point where the open internet has been read, and the next frontier is defined by access to datasets with the consent and provenance that regulators require. Sui and Walrus were built to be the home where AI data is stored, licensed, and paid for, capable of operating at the scale these systems demand. The migration of DATA Network and Trace to our stack now brings massive volumes of high-quality, rights-cleared data that is already serving the world’s frontier AI labs.”

SY Lee, Chairman of Poseidon and CEO of PIP Labs, said: “The next generation of AI will be trained on data that lives with people rather than on the open web, and that data needs consent and provenance built in, along with a way to pay the people it comes from, and Trace was built to be that layer. It now becomes part of the Sui stack alongside Walrus, backed by one of the strongest engineering organizations in the industry and with Sui Foundation as the network’s largest strategic holder. This is the right home for the network we built, and I am looking forward to helping Sui make it the data layer for AI.”

Andrea Muttoni, CEO of DATA Foundation, said: “Training data is critical for continuously expanding the frontier, and making sure it is traceable, auditable and cleared to use is just as important for AI safety as it is a precursor for data quality. With Trace, we have been tracking hundreds of millions of data assets. Joining forces with Sui now allows us to build out more of the stack, offering more than just an audit trail, like encrypted reference file storage and potentially fully verifiable TEEs for evals and fine tunings in the future.”

The existing DATA Network will continue to operate until the upgrade is complete, and a detailed timeline and holder transition plan will be published prior to the upgrade.‍

FAQ

What is Trace?
Trace is an onchain registry for AI training data, launched by DATA Foundation in June 2026. It records where each data contribution came from and the rights attached to it, and pays the person who contributed it. Trace has registered more than 250 million contributions and processes over 3 million receipts a day on average. Applications writing to Trace include Kled, Numo, and Oto, which license data to frontier AI labs and robotics companies.

Why does AI training data need provenance?
The next generation of AI models needs real-world data that comes from people: medical notes, dashcam footage, factory work, customer calls. That data is personal and often regulated. The EU AI Act and ongoing litigation over training data have made provenance a procurement requirement for AI labs, which need proof of where data came from and that they are cleared to use it.

Why are the DATA Network and Trace moving to Sui and Walrus?
On Sui, each dataset and its license become programmable objects, managed onchain like tokens, with transactions settling in under a second. Walrus strengthens provenance by storing an encrypted, unchangeable copy of each original file, which only licensed buyers can read and anyone can independently verify.