Fraction AI Launches Mainnet on Ethereum Layer 2 Base, 320,000 Users Onboarded

Fraction AI, a decentralized AI agent auto-training platform, has announced the launch of its mainnet on the Ethereum Layer 2 (L2) network Base. This transition from the testnet phase to live, scalable deployment marks a significant milestone for the protocol, which enables users to create and train AI agents through open and decentralized reinforcement learning.

During its testnet phase, Fraction AI accumulated 320,000 users who created 1.1 million agents, generating over 30 million data sessions. The platform's smart contract processed more than 90% of the total Wrapped Ether (wETH) volume on the Sepolia testnet, highlighting its robust performance and user engagement.

With the Base launch, users can now deploy AI Agents in live competitions within “Spaces,” environments designed to mimic real-world tasks such as financial analysis, code generation, and copywriting. The platform employs a novel framework called Reinforcement Learning from Agent Feedback (RLAF), allowing independently created agents to interact, compete, and evolve by earning experience points. These points unlock perks such as persistent identity, token issuance, and other premium features, transforming reinforcement learning into an intuitive, permissionless, and user-driven feedback loop.

Fraction AI’s CEO, Shashank Yadav, emphasized the company’s mission to decentralize reinforcement learning and empower users to guide intelligent agents with their unique insights. He noted that the current AI landscape is dominated by centralization, where access to top-tier training methods is limited to a few corporations with substantial compute budgets. Fraction AI aims to challenge this paradigm by making reinforcement learning accessible to a broader audience.

The project has garnered attention from leading investors, including Spartan, Borderless, Anagram, and Symbolic Capital, as well as advisors from Polygon, Near, and 0G. Users of Fraction AI can earn Fractals, which are proofs of contribution that will influence future FRAC token allocations. Additionally, the platform offers staking mechanisms that support decentralization and its broader goal of creating an AI landscape based on accessibility and technological sovereignty.