--- title: Choosing the best Scaleway Instance type for your workload description: Find out how to select the ideal Scaleway Instance type for your specific needs, whether it is for learning, testing, production, or workload optimization. dates: validation: 2025-11-18 posted: 2023-02-20 tags: instance comparison vcpu ram core --- Scaleway **CPU & GPU Instances** are virtual machines in the cloud. You can create and manage Instances via our [console](https://console.scaleway.com/), [API](https://www.scaleway.com/en/developers/api/), [CLI](https://www.scaleway.com/en/cli/), or [other developer tools](https://www.scaleway.com/en/developers/). When you [create an Instance](/instances/how-to/create-an-instance/), you must select the **Instance type** you want to create. This page explains the different ranges of Instances available at Scaleway and helps you to choose the best one for your needs. ## Different Instances for different use cases Different Instance types have different prices and are designed for different use cases. They offer different levels of power and performance, based on their **vCPU** (cores), **memory**, **storage**, and **bandwidth**. You may not need a super powerful Instance if you just want to play around and do some experiments for personal projects, so a **Development** Instance could be perfect for you in this case. But if you want to use your Instance to host a business-critical application in production, you need the power and reliability of a **General Purpose** Instance, precisely designed to reliably handle this type of demanding workload. ## Instance technical specifications Instances are virtual machines: multiple Instances can run on one "real" physical host machine in a data center. The physical host has a **hypervisor**, which **virtualizes** the real physical resources of the host (CPU, memory, storage, etc.) and ensures that each Instance receives its share of those resources. The different types of resources that go into specifying an Instance type are: * **vCPU**: A **v**irtual **C**entralized **P**rocessing **U**nit represents a part of the underlying physical CPU that is assigned to a particular Instance. Each vCPU may be [shared or dedicated](/instances/reference-content/choosing-shared-vs-dedicated-cpus/). - **Shared**: A [context-switching mechanism](https://en.wikipedia.org/wiki/Context_switch) allows a physical core to be shared between multiple vCPUs for multitasking. Instances with shared vCPU can be more cost-effective for running small, non-resource-intensive applications than dedicated vCPU Instances. Our different types of shared vCPU Instances offer a variety of options to balance price and performance depending on your use case. - **Dedicated**: Instances with dedicated vCPU do not share their compute resources with other Instances (1 vCPU = 1 CPU thread dedicated to that Instance). This type of Instance is particularly recommended for running production-grade compute-intensive applications. Scaleway provides CPU Instances equipped with either x86 or ARM CPUs. ARM Instances generally offer higher energy and cost efficiency, while x86 Instances provide wider software compatibility and stronger single-thread performance. [Learn more about the differences between ARM and x86-based Instances](/instances/reference-content/understanding-differences-x86-arm/). * **Core**: A core is a physical unit of vCPU. More cores give an Instance more computing power. A description of an Instance's cores may look something like `4 x86 64bit`, where **4** refers to the number of cores, **x86** refers to the microprocessor family, and **64bit** refers to the chunk size used by the microprocessor for processing data. * **Memory** (aka **R**andom **A**ccess **M**emory): RAM is like an Instance's short-term memory. Data that is currently being used by the vPCU is stored in RAM. More RAM enables faster computing. * **Bandwidth**: Bandwidth refers to the volume of data that can be sent to/from an Instance in a given amount of time (usually **p**er **s**econd). * **Storage**: Different Instance types may support **Local** and/or **Block** Storage: - **Local** storage is an all-SSD-based storage solution, hosted on the local hypervisor. - **Block** storage is more flexible, network-attached storage that can be plugged in and out of Instances like virtual hard drives. Some Instances only support Block Storage, and boot themselves from their Block Storage. Depending on the Instance type, you can attach additional storage volumes and increase their size according to your needs. ## Instance ranges The table below shows the different ranges of Instances at Scaleway and their specifications. Within each range, there are different types of Instances. Use this table to help identify the right Instance range for your use case and computing needs. | **Instance range** | [Development](/instances/reference-content/development/) | [General Purpose](/instances/reference-content/general-purpose/) | [Compute Optimized](/instances/reference-content/specialized/) | [Memory Optimized](/instances/reference-content/specialized/) | |-----------------------|----------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | **Use cases** | Scaling your development and testing environment
Hosting CI/CD runners and containerized worker nodes | Hosting production workloads and business-critical applications
Ensuring predictable CPU performance in the face of high traffic | Designed for high-performance web-serving, video encoding, machine learning, batch processing, CI/CD | Hosting high-demanding analysis, in-memory calculation, big-data processing, high-performance or cache databases | | **Supported storage** | Resilient Block Storage or Local Storage | Resilient Block Storage | Resilient Block Storage | Resilient Block Storage | | **vCPU** | From 1 to 48 cores | From 1 to 64 cores | From 2 to 64 cores | From 2 to 64 cores | | **Shared/Dedicated** | [Shared vCPU](/instances/reference-content/choosing-shared-vs-dedicated-cpus/#shared-vcpu-instances) | [Shared vCPU](/instances/reference-content/choosing-shared-vs-dedicated-cpus/#shared-vcpu-instances) or [Dedicated vCPU](/instances/reference-content/choosing-shared-vs-dedicated-cpus/#dedicated-vcpu-instances) | [Dedicated vCPU](/instances/reference-content/choosing-shared-vs-dedicated-cpus/#dedicated-vcpu-instances) | [Dedicated vCPU](/instances/reference-content/choosing-shared-vs-dedicated-cpus/#dedicated-vcpu-instances) | | **RAM** | From 2 GiB to 256 GiB | From 2 GiB to 384 GiB | From 4 GiB to 512 GiB | From 4 GiB to 512 GiB | | **Maximum Bandwidth** | From 200 Mbps to 10 Gbps | From 100 Mbps to 12.8 Gbps | From 400 Mbps to 12.8 Gbps | From 400 Mbps to 12.8 Gbps Refer to the tabs below to compare the different Instance ranges and learn more about their key use cases: Scaleway Development Instances are highly reliable and affordably priced, making them an ideal choice for businesses looking to balance operational flexibility, cost-efficiency, and performance. Designed for lightweight applications, development environments, and low-intensity workloads, these Instances offer a budget-friendly cloud solution. Key use cases include: * Larger development and testing environments * Low-traffic website hosting * Web application hosting * Hosting worker nodes in container ecosystems * CI/CD runners Development Instances provide a cost-efficient yet powerful compute platform suitable for both development and light production use. Equipped with shared AMD EPYC™ vCPUs and ranging from 2 GiB to 128 GiB of RAM, they cater to a wide array of workloads with solid performance. Learn more about [Development Instances](/instances/reference-content/development/) Scaleway General Purpose Instances are reliable, high-performance Instances with shared or dedicated vCPUs, designed for the most demanding workloads and mission-critical applications. Leveraging the state-of-the-art AMD EPYC™ 7003 series processors, they deliver exceptional performance and production-grade computing. Designed for critical production environments and high-traffic applications, these Instances prioritize performance and reliability with an optimal mix of compute, memory, and network resources. Key use cases include: * Critical production environments * High-traffic applications * Data analytics and business intelligence applications * Hosting of Software-as-a-Service (SaaS) platforms General Purpose Instances provide high performance, security, and scalability, enabling organizations to efficiently manage high-traffic applications and build resilient architectures. Features like CPU pinning ensure optimized performance and reliability for your production workloads. Learn more about [General Purpose Instances](/instances/reference-content/general-purpose/) Scaleway Specialized Instances are designed to handle specific types of workloads efficiently, meeting the demands of applications that require high CPU or memory resources. These Instances are ideal for running compute-intensive or memory-intensive applications such as data analytics, high-performance databases, and machine learning inference. Tailored for businesses with fluctuating needs, these Instances allow for optimal use of resources and cost-effectiveness. Two variants are on offer: * **POP2-HC**: Designed for compute-intensive tasks, this Instance features a high-performance CPU/RAM ratio (AMD EPYC 7543) and is ideal for compute-heavy workloads. * **POP2-HM**: Engineered for memory-demanding applications, this Instance offers abundant RAM and is suitable for memory-intensive workloads. Key use cases include: * RAM-intensive workloads: * High-demanding analysis * In-memory calculations * Big-data processing * CPU-intensive workloads: * Video encoding * Machine learning * Batch processing Specialized Instances precisely tailor your computing resources to your application’s needs, ensuring you neither overspend nor underperform. Whether managing CPU-intensive data analytics or memory-hungry databases, these Instances adapt to your workloads, providing a balanced, cost-effective, and high-performing cloud computing solution ready to meet your specific business challenges. Learn more about [Specialized Instances](/instances/reference-content/specialized/) ## Instance types Check out the [Instances page](https://www.scaleway.com/en/virtual-instances/) of our website for full details on each Instance type. Alternatively, head over to the [Instances datasheet](/instances/reference-content/instances-datasheet/) to see all our available Instance types within each range. ## Instance pricing Instances are priced by the hour, with different Instance types having different hourly pricing. You can check out the current prices for Instances either on the [Instance creation page of the Scaleway console](https://console.scaleway.com/instance/servers/create) or on the [dedicated pricing page of our website](https://www.scaleway.com/en/pricing/?tags=available).