DATA Leaves Its Own Blockchain for Sui to Focus on AI Data Provenance

Table of Contents

• You might want to know


• Main Topic


• Key Insights Table


• Afterwards...

You might want to know

• Why is DATA ending its own blockchain after raising $140 million and pursuing onchain intellectual property?


• How could moving to Sui help the company establish reliable records for AI training data?

Main Topic

DATA, the blockchain startup backed by Andreessen Horowitz’s a16z crypto and formerly known as Story Protocol, is changing both its technical infrastructure and its business focus. The company plans to shut down its own blockchain and migrate to Sui after receiving a $30 million investment from the Sui Foundation. Rather than maintaining a separate network, DATA intends to concentrate on its data verification tools, particularly Trace, a registry designed to document the origins and usage rights of data used to train artificial intelligence models.

The shift follows years of work to place intellectual property onchain. DATA was founded in 2022 in Bellevue, Washington, with an expansive concept: owners could register intellectual property on a blockchain and set conditions for others to use it. The intended scope included music royalties, patents and trademarks. In practice, however, many of these fields already had established licensing systems, and large rights holders were often reluctant to make their assets available for permissionless licensing.

CEO Andrea Muttoni described the original ambition as an attempt to “boil the ocean” in markets that were already saturated and entrenched. The company had raised $140 million and built its own blockchain, yet found that changing the way intellectual property is licensed worldwide was more difficult than expected. Muttoni also said the company had spread itself too thin across many different kinds of intellectual property, each with its own norms, intermediaries and commercial challenges.

The company’s rebrand to DATA Foundation in June marked a move toward AI training data. Four months later, its strategy narrowed further: the priority is now to track where training data comes from, who contributed it and what permissions apply to its use. Muttoni presents this refinement as a continuation of the earlier rebrand, rather than a wholly separate change in direction. The new focus reflects a market problem the team encountered while working with Poseidon, an AI data company incubated by the Story team that collects and processes data for AI training.

According to Muttoni, buyers of datasets can have limited ways to verify whether material is original, appropriately licensed or exclusive. Poseidon encountered vendors offering supposedly exclusive datasets that were already available from other sellers. In one instance, Muttoni said, the team was given its own dataset back. Such experiences highlighted a gap between the growing demand for training material and the ability of buyers to establish the history and rights attached to what they purchase.

DATA’s response is Trace. When a contributor uploads a video, voice recording or other file through a participating data marketplace, Trace creates a record associated with that submission. The record can include when the file was submitted, the terms accepted by the contributor, and later licensing or payment events. This information is intended to give buyers and sellers a way to inspect a data asset’s history, rather than relying only on claims made during a transaction.

One of Trace’s major sources of activity is Kled, a marketplace that pays people to contribute data. Muttoni said Kled is adding between 3 million and 4 million contributions a day, including older records being brought onto Trace. These figures indicate the volume of data being registered, but do not establish how much has been sold or licensed. DATA has not disclosed how many records have generated licensing revenue, an important distinction when assessing whether usage is translating into commercial demand.

Trace is designed to provide provenance and auditability, not to prove that every registered record has been sold or has generated income. For now, Muttoni says the company is not trying to compete directly with data marketplaces. Its stated goal is to build infrastructure that allows participants to check the history and rights information associated with data. He has not ruled out DATA operating a marketplace in the future, but has offered no further details about such a plan.

The strategic reset also raises a practical question: whether DATA needs to operate an entire layer-one blockchain to deliver a focused data registry. Running a network requires engineering resources, validators, audits and infrastructure providers. As Trace became the center of the company’s work, the team reconsidered whether it made sense to keep investing in a separate chain or redirect those resources toward the protocol and product.

The move to Sui is intended to address that issue. DATA will no longer need to maintain its own blockchain, while gaining access to Sui tools that Muttoni says fit the company’s future plans. Two products are particularly relevant. Walrus is a decentralized storage platform that can hold copies of original files registered through Trace. Seal can encrypt those files and limit access to buyers with suitable permissions.

At present, Trace records information about a file but does not store the file itself. Adding storage could allow a buyer to compare received material with the original submission and make it easier to identify whether a file has been changed. Encryption and permission controls could also support more restricted access to sensitive or commercially licensed data. These capabilities are potential components of the system; their value will depend on how they are implemented and adopted by contributors, marketplaces and data buyers.

The Sui Foundation’s $30 million investment will make it DATA Network’s largest institutional holder. Muttoni said Sui is taking a hands-off approach to DATA’s product plans, and the companies did not disclose financial details beyond the investment amount. Existing investors, including a16z crypto, remain involved. Muttoni did not disclose how much of Story’s original $140 million in funding remains.

DATA also plans to move its existing token to Sui on a one-to-one basis. The current blockchain is expected to remain operational until the migration is complete. According to a press release, existing $DATA tokens will migrate one-to-one, while holders’ rights and vesting schedules will remain unchanged. The migration is therefore intended to change the network supporting the token without changing those stated holder terms.

The company’s decision comes amid a broader interest in AI-related applications across the crypto sector. DATA is betting that demand for AI training material will increase alongside the need to establish where that material came from and what uses are permitted. Muttoni estimates the AI training data market could be worth hundreds of billions of dollars, although that estimate is not a measure of DATA’s current revenue or market share.

Ultimately, DATA is concentrating on the part of its original business that attracted the strongest interest: recording data history and usage permissions. Its earlier goal of making many types of intellectual property available for onchain licensing gave way to a narrower product thesis, shaped by the difficulties of changing entrenched markets and by direct experience with opaque dataset claims. The success of the new strategy will depend on whether Trace can become useful and trusted across the data supply chain, and whether buyers and contributors consistently rely on its records.

Key Insights Table

Afterwards...

DATA’s next challenge is to demonstrate that verifiable provenance can become a dependable part of AI data commerce. That will require more than registering large volumes of contributions. Buyers need records they can interpret and trust, contributors need clear explanations of consent and licensing terms, and marketplaces need practical ways to connect those records to the files they sell. Independent review and transparent reporting could help distinguish registered activity from actual commercial use.

Several areas merit further exploration. Technical standards could make provenance information easier to carry between marketplaces and AI systems. Privacy-preserving methods may help verify permissions without unnecessarily exposing personal information or sensitive source material. Secure storage and access controls, including tools such as Walrus and Seal, could be evaluated for reliability, cost and suitability across different data types.

More broadly, researchers, policymakers and industry participants will need to examine how consent, compensation and accountability should work as AI training data is collected and reused. Clearer documentation may help, but it cannot by itself resolve every legal or ethical disagreement about data rights. The long-term opportunity lies in combining verifiable technical records with understandable rules, meaningful contributor choices and responsible oversight. DATA’s move to Sui gives the company a narrower operational foundation; whether that foundation supports a widely adopted system will depend on continued technical development and real-world participation.