Nebius AI Cloud
Product · SaaS / cloud software by Nebius
The ultimate AI cloud engineered from silicon to API.
Nebius AI Cloud is an integrated cloud platform engineered specifically for the complete AI lifecycle, from data preparation and model training to production deployment. It provides businesses and AI developers with high-performance infrastructure featuring NVIDIA GPUs, InfiniBand networking, and specialized MLOps tooling. The platform offers a range of services including managed Kubernetes, serverless AI endpoints, and managed MLflow for experiment tracking. By combining raw hardware power with managed services, it enables teams to scale AI workloads from small experiments to global-scale env
- Languages
- English
- GPU Platforms
- NVIDIA H100, H200, B200, B300, RTX 6000, L40S
- Networking
- InfiniBand support
- CPU Platforms
- AMD EPYC Genoa, Intel Ice Lake
- Storage Protocols
- S3-compatible, POSIX-compliant shared filesystems
- Management Interfaces
- Web Console, CLI, Terraform, gRPC API, REST API
- SDK Support
- Go, Python, JavaScript
Overview
Nebius AI Cloud is an integrated cloud platform engineered specifically for the complete AI lifecycle, from data preparation and model training to production deployment. It provides businesses and AI developers with high-performance infrastructure featuring NVIDIA GPUs, InfiniBand networking, and specialized MLOps tooling. The platform offers a range of services including managed Kubernetes, serverless AI endpoints, and managed MLflow for experiment tracking. By combining raw hardware power with managed services, it enables teams to scale AI workloads from small experiments to global-scale environments.
Compute Infrastructure
- GPU-Accelerated VMs: Deploy and run AI models on high-performance virtual machines with dedicated NVIDIA GPU support.
- InfiniBand Networking: Utilize InfiniBand networks for high-speed distributed computing across GPU clusters.
- Preemptible VMs: Run interruptible workloads at market-driven prices with the ability to set maximum price limits.
Platform Services
- Object Storage: Store ML/AI datasets and model artifacts in AWS S3-compatible object storage buckets.
- Disks and Shared Filesystems gateways: Access high-performance block and file storage for Compute VMs and Kubernetes nodes.
- Managed Kubernetes: Manage containerized application deployment with native support for GPUs and InfiniBand.
- Soperator Clusters: Deploy Slurm workload manager clusters for large-scale ML/AI experiments.
AI Development Tools
- Serverless AI: Run containerized AI workloads as interactive endpoints or background jobs without managing infrastructure.
- Managed Service for MLflow: Track machine learning experiments, organize model versions, and compare metrics in a managed environment.
- Cloud Shell: Access a browser-based terminal with preinstalled CLI, Terraform, and Python for resource management.
Pricing
- NVIDIA HGX H100 On-demand
- $4.50 per GPU-hour
- NVIDIA HGX H200 On-demand
- $5.40 per GPU-hour
- NVIDIA HGX B200 On-demand
- $8.50 per GPU-hour
- AMD EPYC Genoa CPU-only
- From $0.13 per CPU-hour
- Object Storage Standard Volume
- $0.0147 per GiB/month
- Shared Filesystem (by Nebius)
- $0.0800 per GiB/month
Specifications
- GPU Platforms
- NVIDIA H100, H200, B200, B300, RTX 6000, L40S
- Networking
- InfiniBand support
- CPU Platforms
- AMD EPYC Genoa, Intel Ice Lake
- Storage Protocols
- S3-compatible, POSIX-compliant shared filesystems
- Management Interfaces
- Web Console, CLI, Terraform, gRPC API, REST API
- SDK Support
- Go, Python, JavaScript
Added by AntSpace from public sources (nebius.com).