The Web3-native artificial intelligence startup Bluwhale AI Inc. said today it has raised $100 million in funding as it looks to pivot from its original mission and build a kind of “intelligence layer” that spans all Layer-1 and Layer-2 blockchain networks.

More specifically, the company explained, it’s partnering with another AI startup called Caldera Network Inc. to create a Layer-3 zero-knowledge rollup network called Oceanum, which will support the scaling of AI applications across L1 and L2 networks.

Today’s funding comes from multiple sources, including a $75 million token purchase commitment from investors, its seed and Series A funding rounds, and proceeds from its node sale. It detailed a long list of investors, including Cointelegraph, SwissBorg, Awaken Finance, CoinBureau, Altcoin Daily, Master Ventures, DWF Labs, Baboon Ventures, Cryptonite Capital, Hub71, NewTribe Capital, Vitality Ventures, Citizen Capital, OffChain Labs and many others.

The startup has built what it describes as a “decentralized AI personalization protocol,” which leverages the data from Web3 user’s digital interactions to create a kind of “shared brain” that any application can tap into, in order to learn from user’s combined data. The protocol works by transforming data from user’s crypto wallets into a format that AI applications can easily use and learn from, while ensuring that information remains secure. Because the protocol is built on the blockchain, it also has mechanisms for compensating users for sharing their data, with its value determined by market demand.

Now, Bluwhale is looking to expand on that mission with the upcoming launch of its AI-specific Layer-3 network Oceanum, which is a computing environment for AI agents and models that employs a technology called zero-knowledge proofs to provide aggregation, contextualization and privacy.

ZK proofs, as they’re known, use cryptographic verification techniques, where one party can demonstrate to a second party that it has knowledge of secret data that needs to be verified, without actually revealing that data. This technology is uniquely able to secure and verify data and demonstrate the accuracy of computations without disclosing any information.

With the Oceanum layer, Bluwhale says, it will be uniquely positioned to decentralize the AI industry, as it’s powered by thousands of individual nodes that can run on user’s mobile devices. The network is designed so that these nodes collaborate with one another to contribute and verify data for the network, while also providing storage and computing resources for the AI applications hosted on it.

Bluwhale Chief Executive Han Jin said the problem with earlier attempts to decentralize AI is that the individual building blocks available tend to be scattered across fragmented blockchains and networks. Being fragmented makes them almost useless to AI developers, he explained. Rather, they need to exist on a single aggregation layer.

“Our L3, built in collaboration with Caldera, will deliver data contextualization and orchestration for these scattered blocks to create the most optimal and secure aggregation layer for AI agents, models and large language models to plug into multi-functional RAG systems,” Jin said.

Bluwhale intends to build its customized zk-layer atop the Arbitrum Layer-2, which is a Layer-2 network designed to scale Ethereum applications by processing transactions off of the main blockchain. Sitting above Arbitrum, it will provide a dedicated environment for AI applications to process and share user’s data, optimized for privacy, speed and cost-efficient storage and compute.

Kiet Fong of OffChain Labs said Bluwhale’s Oceanum will simplify the complexity of blockchain-based AI applications. “It’s making it as easy as plug-and-play, or as simple as having a friendly chat with retail users,” he added.

Image: SiliconANGLE/Freepik AI