Scaleway GPU Instances

Cloud compute instances (Linodes) with shared, dedicated and premium CPU, high-memory and GPU plans and predictable pricing, deployed in Akamai's core compute regions

Secure, elastic, high-performance virtual machine instances across general-purpose, compute-optimised, memory-optimised, GPU, and bare metal families with per-second billing

OpenStack-based virtual servers in Aruba's own Italian data centres, offered in General Purpose and RAM-oriented flavours (1–32 vCPU) and billed hourly, monthly or yearly; part of the Virtual Data Center ecosystem with Block Storage and VPC networking.

GPU resources for Kubernetes workloads with NVIDIA GeForce RTX 4080 Super (16 GB) or L40S (48 GB) cards, billed at a fixed monthly fee with free set-up on six-month commitments; data kept in Aruba's European data centres.

Secure, resizable virtual servers in the cloud with on-demand, reserved, spot, and savings-plan pricing, over 600 instance types, and Nitro-based hardware isolation

Simplified VPS-style compute — flat-fee instances bundled with data transfer, DNS, storage, and load balancers, targeted at developers who want a Heroku-like experience on AWS without operating the full VPC stack

On-demand scalable virtual machines with D, E, F, L, M, and N series sizes, spot pricing, reserved instances, availability sets, and availability zones

Azure Virtual Machines operated by Bleu from data centres located exclusively in France, across two regions more than 300 km apart, run by Bleu's French operations teams. Linux and Windows workloads using Azure tooling.

KVM-based cloud servers billed hourly (usage metered each minute) from Brightbox's Manchester data centres. Local-SSD types from 512 MB RAM / 1 vCPU and network block-storage types from 2 GB / 2 vCPU, with live resize and automated snapshots.

OpenStack Nova virtual servers billed hourly in the Porirua (nz-por-1) and Hamilton (nz-hlz-1) regions. Shared-CPU c1, burstable c2-burst, dedicated-CPU c3 and NVIDIA A100/L40S GPU flavours; Linux and Windows images and cloud-init.

Virtual machines billed hourly (rounded up to the next full hour) with Linux distributions, custom images and GPU instances (GPU in LON1 only), managed through the same API, CLI and dashboard as Civo Kubernetes

NVIDIA GPU instances and GPU-backed Kubernetes nodes (A100, L40S, H100, H200, B200, B300) in the UK, billed hourly on demand or with 6–36 month commitments

OpenStack Nova virtual machines in Cleura's Swedish and German regions. Flavour tiers: general-purpose b (HA network storage), s (high-performance local storage) and c (dedicated physical cores); GPU cores in beta. Usage measured per second, billed hourly.

Shared-infrastructure virtual machines on open-source hypervisors and Cisco UCS blades, billed on actual hourly usage across four tiers (Dev to Enterprise); HDS and ISO 27001 certified, with SecNumCloud qualification stated as in progress for this offer

NVIDIA GPU capacity delivered on dedicated Bare Metal servers (L40S and H100 configurations) with MIG partitioning, usable from IaaS, OpenShift and Managed Kubernetes, for sovereign-FR AI training and inference

Scalable virtual machines (Linux-based) with predictable pricing and one-click apps

Virtual machines with single or 8-GPU NVIDIA and AMD configurations for AI training and inference, available on-demand, as spot capacity or on contract

Dedicated single-tenant 8-GPU servers for large-scale AI training and inference, provisioned on contract

KVM-based VMs in Standard, CPU-Optimised, Memory-Optimised and Storage-Optimised families, billed per second, with Anti-Affinity Groups for HA; GPU instances are listed separately

NVIDIA GPU instances with 1 to 8 cards (A30, V100, A40, A5000, RTX 3080 Ti, RTX Pro 6000 and B300) for AI/ML training, inference and HPC, billed per second in selected zones

Virtual machine instances with Linux and Windows images, shared-CPU Basic and dedicated-vCPU plans (Standard, CPU-optimised, memory-optimised, high-frequency), attached block storage, and pay-as-you-go per-minute billing

GPU compute instances on NVIDIA H100, H200, and A100 hardware for AI training and inference workloads, available as bare metal or virtualised, with hourly billing and reserved capacity options

Scalable high-performance virtual machines with custom machine types, live migration, per-second billing, and sustained-use and committed-use discounts

Google-designed Tensor Processing Unit accelerators for training and serving large ML models, available as TPU VMs and slices, usable from GKE, Compute Engine and Agent Platform.

Shared-vCPU KVM cloud VMs (CX/CPX x86 and CAX Ampere ARM lines) with hourly billing capped at a monthly maximum and free traffic allowance

CCX line of dedicated-core cloud VMs for steady production workloads where shared-vCPU CPU steal is unacceptable

On-demand virtual cloud servers across Huawei Cloud's 34 regions, with general-purpose, memory-optimised, compute-optimised, GPU and Kunpeng Arm instance families, per-second billing, and integration with Huawei VPC and EVS block storage

End-to-end AI development platform with AutoML, data labelling, distributed training on Ascend and GPU clusters, and one-click deployment to cloud or edge

Virtual server instances on IBM Cloud VPC, on shared multi-tenant hosts, discounted spot instances or single-tenant dedicated hosts, with x86 profiles including GPU options and hourly billing; IBM Power workloads run on the separate Power Virtual Server offering

IBM Power-based virtual servers running AIX, IBM i or Linux in IBM Cloud data centres, for lifting and shifting Power workloads

OpenStack-based virtual machine instances hosted in Swiss data centres with general-purpose, compute-optimised, memory-optimised, and GPU flavours, per-hour billing, and integration with private networks and Block Storage

Customisable virtual server instances with per-minute billing in IONOS EU datacentres — choose CPU type, cores, RAM, and storage independently rather than picking from a fixed shape list, with GDPR-resident defaults and BSI C5-certified infrastructure

Pre-sized shared-vCPU virtual servers (Basic and Memory Cubes) with one directly attached NVMe volume, billed per minute on a pay-as-you-go basis; a lower-cost alternative to dedicated-core Compute Engine servers, without live vertical scaling of CPU or RAM.

Virtual machines with dedicated NVIDIA H200 GPUs (PCIe passthrough) in fixed S, M, L and XL templates for AI training, inference and HPC, available in Frankfurt (de/fra/2) and Worcester (gb/bhx).

Standard / CPU-optimised / memory-optimised / GPU virtual servers in Kakao Cloud's kr-central-2 region (Seoul metropolitan area), per-second billing after the first minute

Deploys built-in catalogue or custom AI models to dedicated GPU-backed inference endpoints called via API key; available to pre-approved customers

KVM-based virtual machines on OpenStack Nova with hourly billing, general-purpose / CPU-optimised / memory-optimised / GPU shapes, and Amsterdam-resident hosting. The IaaS-compute foundation under everything else in the Leaf catalogue.

NVIDIA A30, A100 (80 GB), H100 (80 GB) and RTX 6000 Blackwell (96 GB) GPU VMs in Amsterdam for AI, ML and rendering, billed hourly with the VM included and discounts for commitments; deployable via OpenStack or Gardener-managed Kubernetes.

KVM-based x86 virtual servers with SSD and HDD storage, custom ISO uploads and a provisioning API. Prepaid monthly/quarterly/yearly or on-demand per-second billing; available in London (SOV, CLL, THS), Cambridge, Amsterdam and Fremont, California.

Virtual servers (SSD and GPU) and Bare Metal Servers in Korea, Japan, Singapore, US West and Germany; 3rd-generation KVM-based servers are available only in the Korea, Singapore and Japan VPC environments

Virtual machines on Nscale GPU or CPU infrastructure, each running in a VPC behind security groups and booting from pre-provisioned or custom images or snapshots; managed bare-metal nodes are also offered for intensive workloads.

Sales-led, reserved bare-metal NVIDIA GPU capacity (e.g. GB300 NVL72 racks and B200 hosts) ring-fenced per project and region on InfiniBand/NVLink fabrics, delivered via Nscale Kubernetes Service, Managed Slurm or bare metal for large-scale training and inference.

Managed Slurm batch scheduler running on Kubernetes, providing HPC-style scheduled queues for large-scale AI training and mixed workloads on Nscale GPU capacity.

SecNumCloud-qualified virtual machines from French sovereign data centres. Built on Outscale's TINA compute stack — general-purpose, memory-optimised, and compute-optimised instance families with Linux and Windows images, cloud-init bootstrap, and integration with NumSpot block storage + networking.

KubeVirt-based VM workloads running alongside containers on the same OpenShift cluster, enabling lift-and-shift of legacy VMs without leaving the platform

Compute service that schedules and manages virtual machine instances across hypervisors (KVM/QEMU via libvirt, VMware vSphere, z/VM, Ironic for bare metal). The reference implementation behind OpenStack-based private clouds, equivalent to EC2 / Azure VMs / Compute Engine

Bare-metal and virtual machine compute instances on Oracle Cloud Infrastructure — Intel/AMD x86 plus Ampere A1 ARM shapes, GPU instances (NVIDIA A10/A100/H100/RTX PRO), per-second billing, and an always-free tier covering 4 OCPU ARM allocation

On-demand KVM virtual machines in T Cloud Public's EU data centres (Biere/Magdeburg DE, Amsterdam NL), with general-purpose, dedicated, memory-optimised, large-memory, disk-intensive, high-I/O and GPU flavours, hourly or reserved monthly billing, and VPC/EVS integration

Flexible Compute Unit - virtual machines with AWS EC2-compatible API: custom vCore/memory shapes, OUTSCALE-specific and AWS-compatible predefined types, optional dedicated VMs, and reserved VMs at discounted rates for a set duration

Flexible GPUs (fGPUs) allocated and attached to VMs (up to two of the same model), with NVIDIA models from K2, P6, P100 and V100 to A100, L40, H100 and H200 depending on Region, for AI/ML training and inference

OpenStack-based virtual machines in OVHcloud regions across Europe, North America and Asia-Pacific: Discovery, General Purpose, Compute Optimised, Memory Optimised, GPU, Storage Optimised and Metal flavours. Billed hourly or monthly; outbound traffic is included except in Singapore and Sydney.

Public Cloud GPU instances with NVIDIA GPUs, including H200, H100, L40S, L4, V100S and Quadro RTX 5000, plus NVMe storage and up to 25 Gbps of public and private networking. Used for AI training and inference, video processing and rendering.

Low-cost virtual private servers on monthly plans, with anti-DDoS protection, daily backup and unlimited bandwidth up to 3 Gbps. Used for websites, applications, game servers, VPNs and development.

Google Compute Engine VMs in S3NS's SecNumCloud-qualified Cloud de Confiance (PREMI3NS), run from the French u-france-east1 region. Subset of GCP families (C4, C3, M3, and A3 with NVIDIA H100); Debian, Ubuntu, Rocky Linux and COS images, with Windows only as bring-your-own-licence.

Virtual machines with shared and dedicated CPU machine types, GPU flavours, custom images and STACKIT private networking, billed hourly or monthly in the EU01 (Germany) and EU02 (Austria) regions

Elastic compute service offering scalable, secure virtual machines across 30+ global regions with a range of instance families, spot instances, dedicated hosts, and bare-metal options, plus per-second billing on eligible SKUs

KVM cloud servers in Starter (Standard storage, 99.99% SLA), Premium (MaxIOPS storage, 99.999% SLA) and Cloud Native plans, priced monthly but billed hourly (capped at 672 hours), across 15 zones in Europe, North America and APAC

Cloud servers with NVIDIA GPUs for AI/ML training and inference, rendering and media workloads

Hourly-billed virtual machines with shared and dedicated CPU, custom ISO uploads, snapshots, and a global private network spanning every Vultr region

On-demand NVIDIA GPU instances (H100, H200, A100, L40S, A40, A16) with hourly billing for AI training and inference workloads, with reserved capacity available via the Vultr sales team