Your workload heats someone's shower
Our servers live inside Dutch apartment buildings (Leaf Sites). The heat your GPU generates goes straight into the hot water system. No carbon credits — just real impact and warm showers.
GPU training model
temp: 82°C 🥵
🔥waste heat → hot waterheat → water💧
tenant showering
water: 38°C 😌
Everything you need to compute
From machine learning training to web apps, pick the right instance for your workload.
GPU Instances
NVIDIA A100 and H100 GPUs for machine learning, AI inference, and rendering workloads.
Compute Instances
AMD EPYC-powered virtual machines for web servers, databases, and general workloads.
Green by Default
Every instance reuses waste heat to warm Dutch homes. Same performance, zero guilt.
Need more control? Get a full Leafcloud account
Multiple projects, teams, Managed Kubernetes, OpenStack, S3, and pay-after billing. €2 trial — 2 weeks free to try (see leaf.cloud/faq)
- Multiple projects, teams, and user access
- Managed Kubernetes
- Full OpenStack control (network, firewall)
- S3 object storage
- Pay-after billing and invoicing
€2 trial — 2 weeks free to try (see leaf.cloud/faq)
$ leafcloud instance create \
--flavor gpu.a100.1x \
--image ubuntu-22.04 \
--name my-ml-trainer
Creating instance... done
Instance my-ml-trainer is active
IP: 185.2.4.123
$ ssh ubuntu@185.2.4.123
Welcome to Ubuntu 22.04 LTS
ubuntu@my-ml-trainer:~$ _
Three steps. That's it.
No setup wizards. No waiting. Just pick, launch, and go.
Pick a flavor
Choose between GPU and compute instances. Scale from 1 vCPU to multi-GPU beasts.
Choose your OS
Ubuntu, Debian, or your favorite Linux distro. Add your SSH keys and you're set.
SSH in and build
Your instance is live in seconds. SSH in and start shipping. It's that simple.
Ready to launch?
Join the green cloud. Deploy instances that perform great and heat homes while they're at it.