Written by Helena Markou
The DATA Foundation will retire DATA Network’s independent chain and run its protocol natively on Sui, with Walrus as the default storage layer, under a partnership in which the Sui Foundation invests $30 million and becomes the protocol’s largest institutional holder.
One of the more consequential moves in blockchain infrastructure this week is a chain that is being switched off. The DATA Foundation, which operates an AI training-data protocol, will migrate DATA Network to run natively on Sui and retire its independent chain entirely. Under the same arrangement, Walrus becomes the default storage layer for the protocol’s data, and the Sui Foundation will invest $30 million, becoming the largest institutional holder of the network’s token. The project’s Trace component migrates alongside it.
The background explains the shape of the deal. The project rebranded from Story Protocol to the DATA Foundation on 25 June 2026, and holders of the earlier IP token were migrated to the DATA token on a one-to-one basis. What is being announced now is not a new token but a change of venue: the protocol keeps its identity and its token while giving up its own consensus and block-production layer.
That is a recognisable trade. Running an independent chain means paying for validators, security, tooling and developer attention in exchange for sovereignty. For a protocol whose product is a data marketplace rather than a settlement network, those costs are hard to justify against the distribution and liquidity of an established layer-one. Retiring the chain is the logical conclusion, and it is a pattern likely to repeat as application-specific chains discover that they were never really in the chain business.
Storage is the part worth watching. AI training data is large, persistent and needs to be retrievable at predictable cost, which is precisely the gap that decentralised storage networks have struggled to fill against centralised object storage on both price and latency. Making Walrus the default is a test of whether a decentralised storage layer can be the economical choice for a demanding workload rather than a compliance checkbox. If it cannot compete on cost, the default will quietly become optional.
The $30 million commitment also deserves a clear-eyed reading. A foundation investing in a protocol that runs on its own network aligns incentives and deepens the ecosystem, but it also concentrates exposure. When the largest institutional holder of a token is the entity promoting the network it settles on, the arrangement is closer to a treasury allocation than to independent third-party demand. That does not make it improper — foundations routinely deploy treasury assets to support builders — but it means the investment is not evidence of external market demand, and should not be cited as such.
Three things can be checked against primary records rather than press language. First, the migration itself: whether the independent chain’s state is retired cleanly and whether holders retain balances through the transition. Second, the token mechanics: the one-to-one migration from the earlier token already establishes a precedent, and any deviation from it would be material. Third, and most important, whether the protocol attracts paying users. Network activity generated by incentives and activity generated by customers who need training data are different quantities, and only the second one compounds.
The strategic logic here is sound, and consolidation onto fewer, better-capitalised layers is a healthy development for an industry that overproduced chains. The burden of proof is not the migration. It is whether anyone pays for the data.
