Grass does not yet have enough computing power to meet growing demand from artificial intelligence labs, according to founder Andrej Radonjic. The decentralized network uses residential internet connections to collect public web data that may be blocked by websites when it comes from corporate data centers. As of mid-2026, Grass reported more than 8 million active nodes, over 1 billion GB delivered and a peak of 4 PB collected in one day, with OpenAI and Anthropic identified as major customers. Radonjic estimates that Grass needs 30 to 50 times its current infrastructure to satisfy demand for training datasets and real-time inference data. The company generated about $17 million in revenue in 2025 and another $17 million in the first half of 2026, and projects full-year revenue of $65 million to $75 million from training data alone.
Source: Read the original article
