Decentralized data network Grass is facing a significant supply crunch as demand from AI laboratories for web data outpaces its current infrastructure. Founder Andrej Radonjic states that the network requires 30 to 50 times more capacity to satisfy existing requests. The platform currently utilizes over 8 million active nodes to route data collection through residential IP addresses.
The project reported revenue of $17 million in 2025 and an additional $17 million during the first half of 2026. Projections for the full year 2026 estimate revenue between $65 million and $75 million from training data alone. Operating expenses are currently reported at $2 million to $3 million per month following infrastructure upgrades.
Grass has delivered over 1 billion GB of public web data to date. Major AI labs, including OpenAI and Anthropic, are listed as recipients of this data. The network is currently scaling its own data centers and GPU processing capacity to improve data filtering and control, while also focusing on developing tools for real-time context retrieval for AI models.
