🧠Introduction

The Data Anchoring Token (DAT) is a semi-fungible token (SFT) standard developed by LazAI to represent AI-native digital assets such as datasets, models, and inference outputs.

Unlike general-purpose NFT or token formats, DAT integrates three essential properties into a unified token structure:

• Ownership Certificate: Proof of contribution or claim over datasets, models, or computation results;

• Usage Right: Access quota to invoke AI services, such as agent execution or model calls;

• Value Share: Economic entitlement to future revenue, proportional to the token’s value and shareRatio.

With a Class-based architecture, value-based metering, and on-chain verifiability, DAT enables:

• Composable AI datasets and modular agents

• Tokenized inference and usage-based access

• Royalty-backed economic models for AI contributors

This standard serves as the core abstraction for AI assets in LazAI, supporting programmable licensing, fine-grained rights enforcement, and seamless integration with the broader AI data economy. It represents a next-generation framework that moves beyond static NFTs or ERC-20s - optimized for the dynamic, evolving world of decentralized AI.

DAT Use Case Examples

Use Case	How DAT is Used
User accesses an AI Agent	Consumes value from the associated model’s DAT as service credit
iDAO member uploads data & mints DAT	Receives a DAT representing ownership and future revenue rights
iDAO distributes inference access	Issues temporary Inference DATs for access control
User purchases high-quality data	Buys Dataset DATs, gaining both access rights and revenue share

Additional Technical Features

• Programmable Authorization (UsagePolicy): Define usage windows, agent whitelists, transferability

• Multi-version Proof Tracking: Support for multiple rounds of model training with updated proofs

• Chain Resource Awareness: Tracks storage or compute footprint of different AI asset types