Grass (Wynd Network)

Grass, developed by Wynd Network, is a decentralized physical infrastructure network (DePIN) for collecting public web data. Instead of relying on centrally owned proxy pools, it mobilizes unused residential internet bandwidth contributed by everyday users, who install a node or browser extension and earn token rewards in exchange for routing web-data requests through their connections.

How it works

Participants share bandwidth that the network uses to access and retrieve publicly available web pages from many geographic locations. The data is collected, verified, and structured into a layer intended to feed AI model training, where geographically diverse and representative data is valuable. Coordination and rewards are handled on-chain, giving contributors a share of the value that AI data collection generates.

• Segment: Decentralized web-data collection (DePIN)
• Model: Peer-contributed residential bandwidth with token incentives
• Focus: Public web data structured for AI training
• Founded: Wynd Network, established in 2023

Where it fits in the AI data supply chain

Grass sits at the sourcing layer as an alternative to traditional proxy and scraping vendors, distributing data collection across a large network of individual participants rather than concentrating it in a single provider. Its emphasis on verification aims to make the collected data auditable and attributable. For AI teams, it represents an emerging, crypto-native approach to acquiring web-scale training data, while also functioning as a way for individuals to monetize otherwise idle internet capacity.