Avalon Labs has unveiled the world's first blockchain platform for on-chain AI assets, as detailed in its newly released whitepaper. The platform introduces AI-Model-as-a-Service (AI-MaaS), which tokenizes pre-trained machine learning models into tradable digital assets on a decentralized marketplace. This allows developers and institutions to purchase or lease access to optimized models without the need for GPU rental or infrastructure overhead, with inaugural models built on H200 GPU hardware for enterprise-grade performance.
A key innovation is the Commercial Rights Tokenization (CRT) framework, which focuses on tokenizing legal rights to use, rent, or access services rather than ownership stakes. This framework is grounded in U.S. commercial statutes, specifically the Uniform Commercial Code's Articles 7, 9, and 12, providing enforceable digital rights with real-world legal weight and bridging smart contracts with compliance.
The marketplace employs a two-tier financial structure: Senior Allocation (sAI) targets risk-averse participants with stable, predictable access to AI resources, aiming for a soft peg of $1 managed by a VWAP-based mechanism. Junior Allocation (jAI) offers flexible access to residual rental capacity, capturing upside during high demand but bearing initial downside risks, creating a DeFi-native approach with a venture-style risk profile.
The platform will launch exclusively on BNB Chain, serving as a decentralized clearinghouse that connects GPU hardware owners with developers. Additionally, Avalon's ecosystem includes BTCFi products, such as the USDa stablecoin backed by Bitcoin collateral, which can be minted with assets like FBTC or swapped from USDT, featuring fixed borrowing rates and staking for yield.
MIM
LTC
ETH
ZK