How decentralized AI training will create a new asset class for digital intelligence
Decentralized AI training networks are emerging to democratize access and create tokenized AI models as a new, tradable asset class.Summary
Frontier AI development is currently inaccessible to most investors due to high training costs. However, new decentralized AI networks are connecting global GPUs—from high-end hardware to consumer devices—into unified training fabrics. These networks coordinate ownership by issuing tokens to contributors based on their compute and bandwidth contributions, giving them a direct stake in the resulting AI models. This tokenization provides an economic structure, allowing tokenized AI models to function like stocks, generating revenue via API access similar to centralized labs. Investors can gain direct exposure to these models, either through tokens conferring access rights or revenue shares, effectively creating a liquid, globally coordinated asset class for digital intelligence that fits within the growing trend of onchain asset tokenization.
(Source：CoinDesk)