GPU Rental Plans
On-demand access to the world's most powerful AI GPUs — powered entirely by renewable solar energy. No contracts, no minimums, no carbon guilt.
H100 SXM5
NVIDIA H100
The flagship AI training GPU. Purpose-built for large language models, diffusion models, and multi-node distributed training.
A100 SXM4
NVIDIA A100
The proven workhorse for AI/ML. Exceptional price-to-performance for training, fine-tuning, and large-scale inference.
RTX 4090
NVIDIA RTX 4090
High-performance consumer GPU ideal for inference, rendering, and smaller training runs. Outstanding cost efficiency.
Included with every plan
Everything you need, out of the box
Need dedicated clusters or custom capacity?
For large-scale AI training runs, enterprise inference infrastructure, or research institutions requiring dedicated GPU capacity — we build custom solar-powered clusters sized to your workload.
Talk to Enterprise SalesCommon questions
How is billing calculated?
Hourly plans are billed per second of actual GPU usage, rounded to the nearest minute. Monthly plans reserve a dedicated GPU for 30 days at a fixed rate.
Is there a minimum commitment?
No. On-demand hourly plans have zero minimum commitment — spin up and tear down at any time. Monthly plans require a 30-day reservation.
How do I verify the renewable energy claim?
We provide real-time energy dashboards and monthly Renewable Energy Certificates (RECs) for every GPU-hour consumed. Enterprise customers receive third-party audited reports.
What software comes pre-installed?
All nodes ship with CUDA 12.x, cuDNN, NCCL, PyTorch, TensorFlow, and Docker. Custom images and Kubernetes orchestration are available on request.
Can I run multi-node training jobs?
Yes. H100 and A100 clusters support NVLink and InfiniBand interconnects for multi-node distributed training. Contact enterprise sales for dedicated cluster provisioning.