An AI training data provenance registry helps buyers verify source, rights, and consent behind datasets used to train models. DATA, backed by a16z, is centering its Trace system on this need and moving to Sui after a $30 million deal, cutting blockchain costs and sharpening focus. DATA, the startup once called Story Protocol, narrowed its plan after a tough lesson: moving the entire world of IP onchain is slow and crowded. The team raised big money and built its own chain, but many rights holders did not want open licensing. So the company shifted to one clear pain point in AI: show where training data comes from, prove consent, and track who can use it. CEO Andrea Muttoni says the change is about focus. The goal is simple and bold. Make datasets easier to trust. Make deals easier to check. Spend resources on product, not on running a blockchain. A new $30 million investment from Sui Foundation supports this move. It also opens new tools for storage and access control.

Why an AI training data provenance registry matters

AI models learn from massive datasets. Those datasets often come from many people, places, and platforms. Buyers need to know what they are getting. Sellers want credit and payment. Creators want control. Without clear records, people get burned. Duplicates pass for “exclusive” sets. Rights are unclear. Payouts are messy. A clear registry solves these issues in three ways: – It shows origin. Each record links a file to a person or source and a time. – It shows consent. Each record lists the terms the contributor agreed to. – It shows use. Each record can list licensing events and payments over time. This turns a gray box into a paper trail. It helps buyers avoid risk and waste. It helps sellers prove value. It helps creators keep a voice in how their works train models. In short, trust grows when everyone can check the same source of truth.

Inside DATA’s pivot from on-chain IP to AI data

DATA started in 2022 with a wide goal: put music, patents, and more onchain for open licensing. The team soon found heavy friction. Music and patent systems already exist. Big rights owners did not rush to open access. The team spread itself thin across many types of IP. In June, Story Protocol rebranded to DATA Foundation and focused on AI data. The team also incubated Poseidon to collect and process training data. That work showed a larger problem. Vendors pitched “exclusive” datasets that others sold too. In one case, someone even tried to sell Poseidon a set the team had created. This was a wake-up call. The market needs basic verification before scale. So DATA chose to build that layer.

How Trace works and what Kled brings

Receipt-based records

Trace creates a cryptographic receipt when a person or partner uploads a file to a supported marketplace. The receipt logs who submitted the file, when, and under what terms. It also can append future events like sales and payouts. Today, Trace records the facts about a file, not the file itself.

Partner-driven supply

Kled, a marketplace that pays people to contribute data, drives much of the current activity. According to DATA, Kled adds 3 million to 4 million contributions per day, including older items being imported. This shows strong supply-side momentum. It does not show sales yet. DATA has not shared how many records have led to licensing revenue.

What Trace means for each side

– Buyers can compare claims with proof. They can see where a set came from and what rights attach to it. – Sellers can point to a neutral record of their offers and deals. – Creators can hold platforms accountable for the terms they accepted. Trace serves as an AI training data provenance registry that makes auditability practical. It gives the market a standard way to confirm “who, when, and under what rights.”

Why Sui? Architecture gains and cost cuts

Running a blockchain is hard and expensive. It needs protocol engineers, validators, audits, and infrastructure. The DATA team asked a simple question: Should we run an L1, or ship product faster? The $30 million investment from Sui Foundation made the choice clear. DATA will shut down its own chain and build on Sui. Two Sui-native tools fit DATA’s roadmap: – Walrus: decentralized storage that can hold copies of original files. This would let buyers compare a delivered file to the registered original. – Seal: encryption and permission controls. This would unlock gated access for buyers who hold the right license. Today, Trace stores proofs and metadata. With Walrus and Seal, DATA can add storage and access control without rebuilding the wheel. Sui Foundation will become DATA Network’s largest institutional holder. DATA will migrate the $DATA token to Sui on a 1:1 basis. The current chain will remain live until the move is complete. Choosing Sui to power its AI training data provenance registry should let DATA focus on features and partner growth, not on infrastructure.

What this means for builders, buyers, and creators

Builders

– Faster integrations: Open APIs and Sui tooling can reduce build time for marketplaces and apps. – Clearer compliance: Standard receipts make it easier to meet internal audit rules.

Buyers

– Better due diligence: Verify origin, consent terms, and past licensing in minutes. – Less waste: Avoid paying twice for the same “exclusive” data.

Creators

– Proof of contribution: Get a durable record of what you shared and when. – Leverage for payouts: Track how works move through the market and press for fair share.

Risks, open questions, and what to watch

No shift is risk-free. Several open items deserve attention: – Market adoption: Registries work when many players use them. Watch for new marketplaces and model labs integrating Trace. – Economic incentives: Contributors need fair pay. Buyers need clear value. The model must reward verified supply without bloating costs. – Legal standing: Receipts help, but courts and regulators decide what counts. Expect tests on consent, exclusivity, and data rights. – Privacy and safety: Storing or linking to originals can raise privacy and security issues. Seal and access controls must be strong. – Duplicate suppression: A clean graph of contributions should reduce repeats. The team must show real-world drops in duplicate sales. – Revenue proof: Registration volume is not sales. The key metric is licensing revenue tied to Trace records over time.

Roadmap signals and metrics to track

If you follow this space, a few signals will show real progress: – Total verified buyers using Trace for procurement checks. – Number of active data marketplaces writing receipts to Trace. – Share of sales with sealed delivery and on-chain receipts. – Time-to-audit for a sample dataset before and after Trace integration. – Cases where Trace data helps resolve disputes or block duplicate “exclusive” sets. – Growth in Sui-native tools used by DATA, including Walrus and Seal. Together, these metrics tell us if the system moves from promise to habit. The day this registry becomes part of everyday procurement is the day trust costs go down across the industry.

DATA’s focus is a bet on trust, not hype

The company tried to change IP at large and found the task too broad. Now it has a narrow, high-need job. Trace documents origin, consent, and licensing. Sui adds storage and permissions to back that up. Partners like Kled bring volume. Model builders bring demand. If these pieces click, the market gets a shared ledger of facts instead of sales pitches. That could save buyers money, send fair pay to contributors, and reduce legal risk for everyone. The path ahead is still hard. Adoption must spread. Legal clarity must grow. But the direction makes sense. In a world racing to feed models with more data, an AI training data provenance registry can turn noise into signal and turn claims into proof. (p) (Source: https://www.coindesk.com/business/2026/10/21/how-a16z-backed-data-is-rethinking-its-blockchain-strategy-after-usd30-million-sui-deal)

For more news: Click Here

FAQ