Nebius is a European cloud providing training-optimized infrastructure for AI practitioners.

We offer a variety of NVIDIA GPUs for training and inference (including H100, A100, L40s and V100) and a set of tools for efficient multi-node training (Infiniband, Managed Kubernetes, Slurm solution library for Terraform, etc).

We provide services to customers from all over the world – Europe, the USA, Middle East, Africa and Asia. Majority of our customers run multi-node training scenarios, including LLMs.

GPU resources are available both on-demand and through reserved instances. You can choose the most cost-effective option based on your computational needs and budget.

We focus a lot on UX: there is a detailed documentation with how-to guides, user-friendly cloud console, infrastructure management via CLI or Terraform, VM access via SSH and a Marketplace with more than 50 3rd party tools

At the same time we offer onboarding support to our users if they need it. A team of dedicated cloud solutions architects is always ready to help with your use-case and get the most of our platform.

About the team

Nebius is built around the talents of more than 500 highly skilled engineers based in Amsterdam (HQ), Finland, Israel, and Serbia.

Our in-house R&D team designs our datacenter and hardware, making it optimised for ML/AI tasks and extremely energy efficient - e.g. we use outdoor air to cool down our data center and created #16 supercomputer in the world.

Nebius is NVIDIA’s preferred partner which allows us to get access to the latest GPUs – we are among first companies who will get Blackwell GPUs.