Key Takeaways
- • Multicoin Capital deployed capital from both its hedge fund and venture fund into Grass on September 30, 2026, a decentralized physical infrastructure (DePIN) network that monetizes idle internet bandwidth for AI applications, driving the GRASS token up 25% intraday to above $0.80.
- • Grass generated $17 million in revenue during 2025 and $17 million in the first half of 2026—representing near sevenfold year-over-year growth—with full-year 2026 guidance of $75 million and confirmed monthly profitability at $2–3 million in operating costs.
- • The investment thesis centers on Grass's shift from training-data supply (a capped market) to real-time search infrastructure for AI agents and models, enabling live web scraping without rate-limiting or stale data—a substantially larger addressable market in the AI inference era.
- • Grass operates a network of over 6 million participants who earn compensation by sharing unused bandwidth, allowing AI labs to route web requests through decentralized connections for live data retrieval without triggering blocking or rate restrictions.
- • Multicoin's backing signals institutional validation of DePIN infrastructure as foundational "read layer" architecture for production AI systems, aligning with the venture firm's track record of early-stage bets on ecosystem-level infrastructure (Solana precedent).
Summary
Multicoin Capital announced investments from both its hedge fund and venture fund in Grass on September 30, 2026, validating the Austin-based crypto fund's thesis that decentralized bandwidth networks will become critical infrastructure for AI inference workloads. The announcement triggered immediate market momentum, with the GRASS token rallying 25% intraday and breaking above $0.80 per token.
Grass operates as a DePIN (decentralized physical infrastructure) network compensating over 6 million participants for sharing unused internet bandwidth. The core mechanism routes web requests through distributed residential connections, enabling AI companies to scrape live web content without encountering rate-limiting, IP blocking, or serving stale data—a technical requirement for AI agents that must verify real-time information before executing tasks, such as checking current stock prices or retrieving fresh news.
Financially, Grass has demonstrated substantial revenue acceleration and path to profitability. The network generated $17 million in revenue during 2025 and matched that figure in just the first half of 2026, representing near sevenfold year-over-year growth. Management is guiding to $75 million in full-year 2026 revenue while claiming profitability, with monthly operating costs for compute and storage running between $2 million and $3 million. According to Multicoin's announcement, one source attributed $14.53 million in verified revenue to the first half of 2026, though the primary figures cited remain $17 million for the half-year.
Multicoin's investment rationale, articulated by general partner Shayon Sengupta, hinges on a strategic market transition. Grass initially built its business supplying training data to foundation model labs—a finite market that taps out once model training concludes. Multicoin's thesis argues the substantially larger opportunity lies in real-time data infrastructure for AI inference: agents and models deployed at production scale require fresh, live web data to execute reliably. This shift from training-data supply to inference-era "read layer" infrastructure positions Grass in a market with far greater revenue potential and stickiness, justifying the venture firm's allocation at this inflection point rather than earlier.
Grass operates as a DePIN (decentralized physical infrastructure) network compensating over 6 million participants for sharing unused internet bandwidth. The core mechanism routes web requests through distributed residential connections, enabling AI companies to scrape live web content without encountering rate-limiting, IP blocking, or serving stale data—a technical requirement for AI agents that must verify real-time information before executing tasks, such as checking current stock prices or retrieving fresh news.
Financially, Grass has demonstrated substantial revenue acceleration and path to profitability. The network generated $17 million in revenue during 2025 and matched that figure in just the first half of 2026, representing near sevenfold year-over-year growth. Management is guiding to $75 million in full-year 2026 revenue while claiming profitability, with monthly operating costs for compute and storage running between $2 million and $3 million. According to Multicoin's announcement, one source attributed $14.53 million in verified revenue to the first half of 2026, though the primary figures cited remain $17 million for the half-year.
Multicoin's investment rationale, articulated by general partner Shayon Sengupta, hinges on a strategic market transition. Grass initially built its business supplying training data to foundation model labs—a finite market that taps out once model training concludes. Multicoin's thesis argues the substantially larger opportunity lies in real-time data infrastructure for AI inference: agents and models deployed at production scale require fresh, live web data to execute reliably. This shift from training-data supply to inference-era "read layer" infrastructure positions Grass in a market with far greater revenue potential and stickiness, justifying the venture firm's allocation at this inflection point rather than earlier.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.