Story becomes DATA Foundation, shifts focus to AI training data

Story has transitioned to DATA Foundation, shifting its focus from general intellectual property infrastructure to AI training data, citing growing demand for rights-cleared datasets to train AI models.

The transition also includes the launch of Trace, an onchain registry and public audit platform for AI training data, alongside a "flagship" integration with Kled that brings more than 1.5 billion user-contributed records onto the DATA network. Story co-founder Andrea Muttoni has become CEO of DATA Foundation, while Kled founder Avi Patel has joined as chief data officer.

Story originally set out to build an IP layer where digital assets could be registered and licensed onchain. Its ecosystem included projects such as Magma, which brought more than 3 million artists onto Story, and Aria, which put music rights from artists including Justin Bieber and BTS onchain, Muttoni told The Block.

When asked why the focus changed, Muttoni said the strongest demand increasingly came from AI companies looking for high-quality, rights-cleared training data. He said the clearest signal came from Poseidon, an AI data-processing project incubated by Story that later raised a $15 million seed round led by a16z crypto and gained traction with frontier AI labs.

"Our core mission remains the same: intellectual property should be programmable, licensable and attributable," Muttoni said. "What we've learned is that AI training data has emerged as one of the most valuable and yet least solved forms of IP."

According to DATA Foundation, AI developers are increasingly running into shortages of high-quality training data as much of the publicly available internet has already been scraped. At the same time, AI companies face growing legal and regulatory scrutiny over whether their training data was collected with proper consent and licensing.

To address that problem, DATA Foundation is building infrastructure that records where training data came from, how it was licensed, whether contributors gave consent, and how contributors were paid.

Its new product, Trace, generates an onchain receipt for every data contribution, allowing AI companies to verify the origin, licensing terms, and provenance of datasets before using them for model training.

The launch includes an integration with Kled, an opt-in marketplace for AI training data, that brings more than 1.5 billion user-contributed records onto the DATA network. Those records include voice, video, images, and other forms of real-world data used to train and evaluate AI models, Muttoni said.

"Kled continues to operate independently as its own marketplace and company. What's changing is that Kled's licensing rails and contributor receipts now run on DATA Network, and its data is auditable through Trace," Muttoni said.

When asked who contributes data to the network, Muttoni said anyone can participate through contributor applications, including Kled, Numo, and Oto.

Numo, another project incubated by Story and now operating on DATA, focuses on collecting human-generated data for AI training. Muttoni said Numo collected nearly 800,000 contributor files during its first month, primarily voice recordings in languages including Bengali, Tamil, Hindi, and Telugu. Contributors complete tasks and are paid in stablecoins or fiat, with payments available through bank accounts or crypto wallets.

Kled contributors are also paid in stablecoins or fiat. Muttoni said contributors can receive USDC on the DATA Chain alongside other stablecoin options, while every contribution is permanently recorded on Trace, along with consent records and licensing information.

When asked who the intended customers are, Muttoni said the primary buyers are AI labs. He said AI companies consistently ask data suppliers whether they can source data at scale, prove where it came from and guarantee its quality. Beyond AI developers, he expects more companies building custom AI models to need specialized training data over the next five to ten years.

DATA Foundation said Poseidon validates and scores contributed datasets before they reach buyers, helping verify authenticity and quality. The project now runs entirely on DATA Protocol.

The foundation also announced that Story's native IP token will migrate to the new DATA token on a one-to-one basis. Existing holders will not need to take any action, with migration guidance and exchange timing to be provided closer to launch. Kled and its token remain separate from the foundation, Muttoni noted.