Greenhash runs hydro‑powered data centers in Sweden and New Zealand that provide on‑demand high‑performance compute for AI, machine‑learning and scientific workloads. Its real‑time grid‑balancing scheduler automatically shifts jobs to periods of excess renewable generation, delivering carbon‑neutral HPC capacity at usage‑based rates tied to electricity costs. Customers access the service via a self‑service portal or API and receive carbon‑accounting metrics for each compute hour.
Funding
Funding not disclosed
Founders
Product
Problem
Data centers consume large amounts of electricity, and reliance on conventional grid power often leads to high carbon emissions and volatile energy costs. When renewable generation exceeds demand, the excess energy is wasted, while during peak demand periods data centers may be forced to draw from non‑renewable sources or incur premium rates.
Solution
Greenhash operates purpose‑built data centers in Sweden and New Zealand that are supplied exclusively by hydroelectric power, guaranteeing a 100 % renewable energy mix. An energy‑aware workload scheduler continuously monitors real‑time grid output and shifts compute load to match periods of surplus renewable generation, effectively acting as a demand‑response resource. Surplus hydro power is converted into high‑performance computing (HPC) capacity, which is offered to customers at market‑competitive electricity prices. The platform delivers low‑latency, AI‑ and ML‑optimized compute while maintaining a carbon‑neutral footprint and stabilizing the local power grid. Clients access the compute via a self‑service portal or API, receiving usage‑based billing that reflects the underlying renewable energy cost.
Target Audience
The primary customers are enterprises, AI/ML developers, and research institutions that require large‑scale compute while meeting sustainability commitments, as well as cloud service providers seeking green‑backed HPC capacity.
Features
- Hydro‑powered server farms with full PUE optimization for maximal energy efficiency
- Real‑time grid‑balancing engine that auto‑scales compute workloads based on renewable supply signals
- Proprietary energy‑aware scheduler that prioritizes HPC jobs during periods of excess hydro generation
- Dedicated HPC clusters tuned for AI, machine‑learning, and scientific simulations, delivering up to 10 PFLOPS per site
- Low‑overhead, self‑service provisioning portal and RESTful API for on‑demand capacity allocation
- Integrated carbon‑accounting dashboard that reports avoided emissions per compute hour
- Compatibility layers for major cloud orchestration tools (Kubernetes, OpenStack) and standard VM/container images
- Tiered usage‑based pricing model that reflects real‑time renewable electricity costs, with volume discounts for long‑term contracts