0G Labs has introduced ERC-7857, a new NFT standard created specifically for AI agents. They call these agents Intelligent NFTs (iNFTs), each carrying its own encrypted metadata. The group hopes this format will make AI agents easier to buy, sell, and use on decentralized platforms.
They also see it as a big step toward AI governed by its owners rather than a central team. This means everyone, from businesses to individual developers, could potentially trade AI agents as if they were any other asset on the blockchain.
Why Traditional NFT Standards Don’t Work
According to 0G Labs, existing standards like ERC-721 and ERC-1155 fall short when it comes to AI. These older models are good at handling collectibles such as digital art or game items, but AI agents are more complicated. For instance, they must hide specific data because that information might be valuable or private.
Traditional NFTs store an asset’s metadata in the open, which doesn’t protect confidential AI code. Furthermore, standard NFTs don’t allow for secure updates when the AI learns new things over time. As a result, someone who buys an AI agent using an old NFT standard might end up with an incomplete product and no way to protect it from unauthorized access.
0G and its ERC-7857 Approach
The team behind 0G Labs designed ERC-7857 to handle AI-specific challenges. At its core, this standard secures an agent’s “intelligence” with encrypted metadata and offers a proof mechanism so everyone can confirm it’s genuine.
When a person sells or transfers an AI agent, an oracle—or some type of trusted system—re-encrypts sensitive data for the new owner. This process ensures only the owner can unlock the AI’s core functions. For example, if a developer has fine-tuned a neural network for advanced tasks, they can attach that model to an iNFT.
Source 0G Labs
Whenever it changes hands, the buyer receives full access, with no risk of the data leaking during the transaction. Buyers also benefit from predictable updates, since ERC-7857 allows an AI to evolve without compromising encryption.
In addition, the 0G Labs team emphasizes that their standard slots right into the rest of their ecosystem, which includes services like 0G Storage and 0G Data Availability. With these in place, AI agents can store their data to keep it accessible to the rightful owner and safe from everyone else.
0G Labs believe this opens the door to new AI marketplaces where iNFTs can be traded easily. Creators also get more control over their earnings since they can set royalties on every resale.
Potential Applications and Outlook
Several use cases have sparked interest among developers. For one, businesses can turn their in-house AI services—like chatbots or data analytics tools—into iNFTs. Doing so could simplify handing these tools to different departments or outside partners without worrying about losing private information.
Individual developers who specialize in AI might also benefit. Instead of selling their services one contract at a time, they could build specialized iNFTs and offer them on a marketplace, collecting royalties whenever those AI agents change hands.
This could inspire a surge in AI-driven applications. People who own an iNFT with unique capabilities might join forces with others in a shared AI marketplace, creating hybrid solutions. A data-focused AI could combine with a creative AI, generating new possibilities that no one could achieve alone.
Additionally, AI-as-a-Service (AIaaS) becomes simpler under ERC-7857. Agents can be tokenized and leased on subscription models, giving businesses or individuals a straightforward way to incorporate specialized AI tools without directly building them.
Conclusion
By rolling out ERC-7857, 0G Labs, which recently raised $325 million, hopes to shift how the world uses and circulates artificial intelligence. Developers, collectors, and businesses can all participate, whether they want to acquire high-powered AI models or build them from scratch and sell them to others.
If this new standard sees widespread adoption, it may mark a turning point in the relationship between AI, ownership, and decentralized technology—where AI agents operate securely and independently, all while keeping control in the hands of their actual owners.
Editor’s note: Written with the assistance of AI – Edited and fact-checked by Jason Newey.
