Data Leaf provides modular, solar‑powered micro data centers that operate off‑grid with proprietary closed‑loop cooling, eliminating the need for grid electricity and water. These self‑contained units can be deployed in months, offering 99.9% uptime and scalable high‑performance compute for AI, cryptocurrency and other workloads, from a single site up to thousands of locations.
Funding
Funding not disclosed
Founders
Product
Problem
Enterprises and AI developers often face long lead times and limited availability when scaling compute resources, especially for large GPU workloads, due to reliance on centralized data centers and inflexible utility power supplies.
Solution
Data Leaf provides a vertically integrated, grid‑independent infrastructure platform that delivers reservable hyperscale GPU compute capacity on demand. By deploying modular data centers powered by on‑site renewable energy and sustainable cooling, the company creates “macrogrid” clusters that can be provisioned within months rather than years. Customers specify required compute, and Data Leaf allocates bare‑metal or colocation resources from a pre‑built portfolio, ensuring immediate access to high‑performance AI hardware while maintaining reliability through N+1 redundancy. The platform’s Baseload as a Service® model balances intermittent renewable generation, delivering consistent power to the compute nodes without dependence on traditional utilities.
Target Audience
Primary customers are AI research labs, enterprise machine‑learning teams, and cloud service providers that require rapid, large‑scale GPU compute with guaranteed power reliability and sustainability.
Features
- Reservable 40 MW of GPU compute capacity ready for deployment in Q1 2026
- Modular, vertically integrated data center sites powered by on‑site solar, wind, and battery systems
- Sustainable cooling practices that reduce energy consumption and carbon footprint
- Grid‑independent, N+1 hardened infrastructure ensuring high availability for AI training workloads
- Lateral execution of distributed macrogrid projects to avoid bottlenecks of linear data‑center builds
- Support for bare‑metal, colocation, or build‑to‑own hyperscale configurations