Perle Gets $9M Seed Funding for Web3 Reward Data Platform - InforCapital

Perle, an AI training data platform that utilizes Web3 to bring human insights to AI models, has successfully closed a $9 million seed round led by Framework Ventures. The funding round takes the total investment in Perle to $17.5 million across seed and pre-seed rounds, signifying a significant milestone in the company’s mission to enhance the training and evaluation of AI systems. Perle is based in San Francisco.

The fresh capital injection will allow Perle to launch Perle Labs, a crypto-native ecosystem designed to revolutionize human input in AI. By incorporating blockchain infrastructure and incentive design, Perle Labs intends to offer transparent payments, on-chain attribution, and confirmable work histories, thereby encouraging worldwide participation and enhancing data quality on a mass scale.

Ahmed Rashad, CEO of Perle, explained that the success of increasingly advanced AI models depends on their ability to handle the long tail of data inputs, which includes rare, ambiguous, or context-specific scenarios. He noted that high-quality human-in-the-loop annotation has been shown to outperform Amazon Rekognition by more than 70%, demonstrating the vital role of human input in filling crucial data gaps. By decentralizing this process, Perle hopes to encourage global participation, reduce bias, and significantly enhance model performance.

Despite the fact that the quality of AI models today is determined by the data they are trained on, even the most advanced systems struggle with complex tasks. Perle Labs plans to tackle these issues by broadening access to diverse, high-quality, and verified data sets, rewarding contributors, establishing on-chain proof of work, and enabling a global community to actively participate in shaping AI.

Vance Spencer, Co-Founder of Framework Ventures, suggested that the progress of AI will be driven more by better data than by simply scaling models. He sees a natural synergy between Perle's robust training platform and the transparent incentive structures enabled by crypto networks, suggesting that they could together unlock a wave of valuable, practical data.

Perle, founded by AI experts from Scale AI, Meta, MIT, and Berkeley, blends human expertise with adaptive workflows to help teams collect, annotate, and evaluate specialized training data quickly and accurately. Its self-serve platform supports the entire AI development lifecycle, including multimodal data collection, reinforcement learning from human feedback, and assistant fine-tuning. Perle is currently aiding leading companies in scaling agentic training processes for complex use cases.

Framework Ventures, a prominent crypto venture capital firm known for its early investments in crucial AI, DeFi, and blockchain infrastructure projects, led Perle’s latest funding round. Perle's previous $8.5M seed round in October 2024 was led by CoinFund and saw participation from Protagonist, HashKey, and Peer VC. With fresh capital and an expanding network of partners, Perle is hastening its vision to make high-quality human feedback the basis of next-generation AI. As the boundaries of AI continue to expand, Perle is dedicated to creating the infrastructure that ensures progress is not merely quicker, but smarter, safer, and more inclusive.