ToolsCompaniesAbout

Nebius AI Cloud

by Nebius GroupAmsterdamNo ratings yet

GPU cloud for training and inference on NVIDIA hardware

Visit nebius.com

Linked from nebius.com · checked 29 Oct

Tool type
Cloud infrastructure for AI workloads
Made for
AI engineers and infrastructure teams
Runs on
Cloud console, CLI and APIs
Last checked
29 Oct 2026
4 product sources

What it does

Nebius AI Cloud is an AI infrastructure platform that provides GPU virtual machines, GPU clusters, storage and managed services for training and inference. Users provision NVIDIA GPU instances (such as HGX H100, H200, B200 and B300 classes), attach shared or object storage, and manage workloads through the Nebius console, CLI and SDKs for Go, Python and JavaScript. Spot-style Preemptible VMs are offered at market-driven prices for interruptible workloads.

People use it to set up ML infrastructure, for example by creating a multi-GPU training VM with a shared filesystem alongside a single-GPU inference VM, then connecting over SSH. The page highlights non-virtualized GPUs with InfiniBand networking, built-in MLOps tooling, serverless and managed inference, and a live spot price view by GPU platform and region that can be checked via API.

It suits AI developers and teams running training or inference at scale who are comfortable with SSH, CLI or SDK setup. Access requires a Nebius account; payments are card or company bank transfer in USD with a $25 minimum first payment. Large GPU clusters and some GPU types, such as GB300 NVL72, are handled through sales contact, and spot capacity is interruptible.

What it can do

  • Provision NVIDIA GPU virtual machines for training and inference
  • Create GPU clusters with InfiniBand networking and non-virtualized GPUs
  • Live spot pricing for interruptible GPU workloads by platform and region
  • Managed Kubernetes and Slurm-on-Kubernetes (Managed Soperator)
  • Object, shared and block storage options
  • Managed services including PostgreSQL and MLflow

Works with

  • Managed Service for Kubernetes
  • Managed Service for MLflow
  • Managed Service for PostgreSQL
  • Nebius CLI
  • Nebius SDK for Go, Python and JavaScript

Example workflows

Training and inference VM setup
Illustrative example. Input: a region (eu-north1) and VM specs. Steps: create an 8-GPU training VM with a shared filesystem, create a 1-GPU inference VM, generate an SSH key pair, and connect to each VM via its public IP. Output: two reachable GPU VMs ready for ML workloads.
Price-aware spot training
Illustrative example. Input: a GPU platform and acceptable price ceiling. Steps: check live spot prices per region on the spot prices page or via the API, pick a region with lower current rates, and launch a Preemptible VM. Output: interruptible GPU capacity at market-driven rates.

Before you choose

  • Check current GPU, region and storage rates for the planned workload.
  • Preemptible workloads can be interrupted; distinguish dynamic spot from fixed-price preemptible options.
  • Storage and network charges vary by service; do not apply one egress rule to all storage classes.

Listed under

Get started

Create an account, then start building in the web console.

  1. Generate an SSH key pair on your local machine.
  2. Install and configure the Nebius AI Cloud CLI, or install and initialize the Nebius SDK for your language.
  3. Create the VMs in the console or via CLI/SDK, then retrieve each VM's public IP and connect over SSH.
Create a cloud account

Plans and prices

Prices in USD

GPU and CPU compute, storage and applicable network services have separate meters. Dynamic spot prices can change every 15 minutes; fixed-price preemptible L40S follows different terms. A first payment of at least $25 is required. Token Factory model API pricing is separate.

On-demand GPU instancesUsage-basedNamed meter: on-demand GPU-hour; see pricing page for per-platform rates.
  • Scalable on-demand rates
  • Commitment discounts for reserved clusters
Preemptible VM (spot)Usage-basedNamed meter: preemptible GPU-hour; rates vary by platform and region.
  • Market-driven interruptible rates
  • Live spot price history
StorageUsage-basedNamed meter: storage volume; egress billed separately.
  • Shared filesystem, object and block volumes

Source: nebius.com/prices, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

Ratings and reviews

No ratings for Nebius AI Cloud yet.

Be the first to rate it

What’s changed

Free account

See what’s changed at Nebius AI Cloud

Launches, price moves and pivots, each with a date and a source. Free with an account.

Signals

Free account

See Nebius AI Cloud’s signals

Launch traction, hiring and release activity, tracked over time. Free with an account.

Data and security

Free account

See how Nebius AI Cloud handles your data

Training on your files, compliance, hosting and the models behind it. Free with an account.

Compare with similar tools

Free account

Compare Nebius AI Cloud side by side

Prices, platforms and features against similar tools, with sources. Free with an account.

Where this comes from

How we research
Company website
nebius.com: product, pricing and policy pages
YC directory mirror
Batch and company name
Social accounts
Accounts the company links, and its posts on X
Human review
Checked by a person
knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.