Physics / mechanism
A Data Union or Data DAO is a smart-contract-governed collective through which individuals pool personal or sensor-generated data, negotiate access terms collectively, and receive tokenised compensation. Mechanically: members deposit data credentials (often via ERC-721/ERC-1155 NFTs or verifiable credentials) into a shared protocol layer; governance tokens determine voting weight on licensing terms; revenue splits execute automatically on-chain. Current deployments (Ocean Protocol, Streamr, Vana) handle O(10k–100k) member pools. Throughput is gated by oracle latency and L2 finality—typically 2–15 second settlement on Polygon/Arbitrum. Privacy-preserving variants layer federated learning or ZK-proofs to allow model training without raw data egress.
Competitive landscape
Centralised data brokers (Nielsen, LiveRamp, Acxiom) dominate enterprise pipelines with lower integration friction but zero user revenue share. Federated learning frameworks (PySyft, NVIDIA FLARE) address privacy without token incentives. Confidential computing enclaves (Intel TDX, AMD SEV) compete on compliance posture.
Companies using
Connected ideas
Sources
Frontier (open questions)
- To be added.