What Is Scaleway? Features, Pricing & Alternatives (2026)

What Is Scaleway?

Scaleway is a European cloud infrastructure provider offering bare-metal servers, virtual private servers (VPS), managed Kubernetes, object storage, and serverless computing. Founded in 1999 and headquartered in Paris, France, Scaleway operates as part of the Iliad Group and positions itself as a cost-effective alternative to hyperscale providers like AWS, Google Cloud, and Azure.

The platform targets developers, startups, and small-to-medium enterprises that need reliable cloud compute without vendor lock-in or enterprise-level complexity. Unlike many competitors, Scaleway emphasizes transparent hourly billing with no long-term contracts, making it particularly appealing for projects with variable workloads or teams testing infrastructure configurations before committing to larger deployments.

Scaleway's infrastructure spans multiple European data centers in France and the Netherlands, with a growing presence in Poland. The company prioritizes EU data sovereignty compliance, making it a natural choice for organizations subject to GDPR requirements or those preferring to keep data within European borders. While it lacks the global footprint of AWS or Azure, Scaleway compensates with competitive pricing and a developer-friendly approach that strips away unnecessary complexity.

Key Features and Specs

Scaleway offers granular control over instance configurations, allowing users to select specific CPU, RAM, and storage combinations rather than accepting pre-packaged tiers. This flexibility means you can provision a virtual instance with exactly 4GB RAM and 2 vCPUs if that matches your workload, rather than jumping to a predefined 8GB tier.

Compute Options:

• Instances (VPS): Range from Development instances with 2 vCPUs and 2GB RAM to GPU instances with NVIDIA cards for machine learning workloads. General-purpose instances use AMD EPYC processors.
• Bare Metal (Elastic Metal): Dedicated servers starting with 4-core Intel Xeon processors and 32GB RAM, scaling to 64-core configurations with 512GB RAM. These bypass the virtualization layer entirely for workloads requiring predictable performance.
• Apple Silicon (Mac Mini M1): Bare-metal Mac minis for iOS and macOS development, billed hourly.

Managed Services:

• Kubernetes Kapsule: Fully managed Kubernetes clusters with automatic updates and node pool management. Control plane free; you pay only for worker nodes.
• Serverless Functions and Containers: Run code without managing servers, billed per invocation and compute time.
• Object Storage: S3-compatible storage with multiple storage classes (Standard, Glacier for archives). Pricing starts around €0.01 per GB/month.
• Managed Databases: PostgreSQL and MySQL instances with automated backups and point-in-time recovery.

Network Features: Scaleway provides public IPv4 and IPv6 addresses, private networking via VPC, and load balancers that distribute traffic across instances. Bandwidth is metered, with egress costs typically lower than AWS equivalents. Instances within the same availability zone communicate over 10Gbps private networking at no additional cost.

Storage: Block storage volumes attach to instances with SSD or HDD options. SSD block storage delivers around 5,000 IOPS for general workloads, while local NVMe storage on certain instance types pushes significantly higher throughput for databases or caching layers.

Scaleway Pricing

Scaleway operates on a freemium model with hourly billing and no upfront commitments. Free tier offerings include limited compute credits, object storage, and managed database instances for testing purposes. Pricing is transparent and published per resource, avoiding the hidden costs that sometimes appear in other providers' billing.

Instance Pricing Examples:

• DEV1-S (Development instance): 2 vCPUs, 2GB RAM, 20GB SSD – approximately €0.01/hour (~€7/month)
• GP1-S (General Purpose): 4 vCPUs, 8GB RAM, 150GB SSD – approximately €0.06/hour (~€45/month)
• GPU instances: Starting around €0.90/hour for configurations with NVIDIA GPUs suitable for inference or training workloads

Bare Metal Pricing: Elastic Metal servers range from roughly €50/month for entry-level 4-core configurations to €300+/month for high-memory or GPU-equipped machines. These prices undercut equivalent dedicated servers from competitors like OVHcloud or Hetzner in some configurations, though direct comparisons depend on specific CPU generations and RAM amounts.

Bandwidth and Storage: Outbound bandwidth is metered beyond included allowances. Typically, each instance includes some free egress (often 100GB-200GB/month depending on instance type), with additional transfer costing around €0.01 per GB. Block storage volumes cost approximately €0.08-€0.10 per GB/month for SSD, less for HDD.

Managed Services: Kubernetes Kapsule control plane is free; you pay only for the worker node instances. Managed databases start around €15/month for small PostgreSQL instances with 1GB RAM. Serverless functions charge per 128MB memory increment and execution time, competitive with Google Cloud Functions pricing.

Billing granularity is hourly, with resources stopped or deleted ceasing charges immediately. No contract minimums or termination fees apply, making Scaleway suitable for short-lived test environments or burst workloads.

Performance and Locations

Scaleway operates data centers in three European countries: France (Paris), the Netherlands (Amsterdam), and Poland (Warsaw). This concentration in Europe means low-latency access for users across the EU but higher latency for requests from North America or Asia-Pacific regions. The provider does not currently offer US East or West Coast locations, limiting its appeal for applications requiring North American proximity.

Data Center Regions:

• Paris (PAR1, PAR2, PAR3): Multiple availability zones with redundant power and networking. Primary hub for most Scaleway services.
• Amsterdam (AMS1, AMS2): Two availability zones serving Northern European customers.
• Warsaw (WAW1, WAW2): Newer region expanding Eastern European coverage.

Workload Fit: Scaleway's infrastructure suits latency-tolerant workloads like batch processing, development environments, CI/CD pipelines, and European-facing web applications. Static site hosting backed by object storage and CDN works well. GPU instances handle machine learning inference and training for European research teams. The platform is less ideal for global content delivery requiring sub-50ms latency from multiple continents—those use cases benefit from providers with broader geographic distribution.

Network performance between availability zones within a region is generally strong, with sub-millisecond latency and high throughput over private networking. Public internet bandwidth from Scaleway data centers typically delivers 1-10Gbps depending on instance type, sufficient for most web applications but not optimized for high-frequency trading or real-time financial data where microsecond latency matters.

Storage IOPS on standard SSD block volumes reach 5,000-15,000 depending on volume size, adequate for web servers and small databases. Workloads requiring sustained 50,000+ IOPS (large-scale analytics, high-traffic database clusters) may need local NVMe-backed instances or consideration of competitors with higher-tier storage.

Who Is Scaleway Best For?

Scaleway excels for specific user profiles prioritizing cost control, EU data residency, or freedom from long-term contracts.

European Startups and SMBs: Companies building SaaS products for European markets benefit from GDPR-compliant infrastructure without the complexity of enterprise cloud contracts. A startup in Berlin running a PostgreSQL database, Node.js API, and static frontend can provision everything on Scaleway for €100-200/month, scaling hourly as user load increases.

Development and Testing Environments: Teams needing ephemeral infrastructure for CI/CD pipelines, staging environments, or proof-of-concept projects appreciate hourly billing. Spin up 10 instances for integration testing, run tests for three hours, delete everything—paying only €3-5 instead of monthly minimums.

Compliance-Focused Organizations: Healthcare providers, financial services, or government contractors subject to EU data residency rules find Scaleway's Paris and Amsterdam regions straightforward for compliance documentation. The provider publishes clear data processing agreements and maintains ISO 27001 and other certifications.

Developers Avoiding Vendor Lock-In: Scaleway uses standard technologies (S3-compatible object storage, Kubernetes, PostgreSQL) rather than proprietary services like AWS Lambda or DynamoDB. Migrating workloads to another provider requires less re-architecting than leaving ecosystems with custom APIs.

Bare-Metal Workloads: Applications requiring dedicated hardware for performance predictability—game servers, video encoding, database clusters—access Elastic Metal servers at lower price points than Dell or HPE dedicated hosting. Hourly billing even for bare metal allows temporary high-performance compute without purchasing physical hardware.

Pros and Cons of Scaleway

Pros:

Granular Configuration: Unlike providers locking you into t2.medium or n1-standard-4 instance tiers, Scaleway lets you select exact CPU and RAM combinations. A workload needing 6GB RAM gets exactly that, not 8GB with wasted capacity.

Bare-Metal Accessibility: Elastic Metal servers provide dedicated hardware at prices often 30-50% below comparable offerings from dedicated server providers. Hourly billing removes the traditional barrier of multi-month contracts for bare-metal access.

No Long-Term Contracts: Pay hourly, scale down or delete resources anytime. No termination fees, no negotiated enterprise agreements requiring minimum spend commitments.

Developer-Friendly Tooling: The web console is intuitive without overwhelming features. API documentation is clear, and Terraform providers work reliably. CLI tools simplify automation for DevOps teams.

EU Data Sovereignty: Multiple regions within Europe, transparent GDPR compliance, and parent company (Iliad Group) headquartered in France reduce regulatory complexity for EU-based organizations.

Cons:

Limited Ecosystem: Compared to AWS's hundreds of services or DigitalOcean's marketplace integrations, Scaleway offers fewer managed services. No direct equivalents to AWS RDS Aurora, Kinesis, or SageMaker. Third-party tools sometimes lack Scaleway-specific plugins.

Support Response Times: Standard support operates during European business hours with response times measured in hours to days, not minutes. Premium support tiers exist but cost extra. This matters less for experienced teams but frustrates users needing hand-holding.

Geographic Concentration: Zero presence in North America, Asia-Pacific, or South America means high latency for global applications. A user in San Francisco accessing a Scaleway Paris instance experiences 150-180ms round-trip latency versus 10-20ms for US-West infrastructure.

Smaller Community: Fewer Stack Overflow answers, tutorials, and open-source integrations specifically for Scaleway compared to ubiquitous documentation for AWS or Azure. You'll often adapt generic Kubernetes or Terraform examples rather than finding Scaleway-specific guides.

Service Maturity: Some managed services are newer and less battle-tested than equivalents from established providers. Managed Kubernetes (Kapsule) works well but lacks the maturity of Google Kubernetes Engine or Amazon EKS.

Scaleway Alternatives

OVHcloud operates European data centers with similar pricing and bare-metal focus. OVH provides more global regions (including North America and Asia) but has historically faced criticism for support quality. Pricing is competitive, often undercutting Scaleway on dedicated servers while matching VPS costs. OVH's larger footprint suits users needing multi-continent infrastructure.

Hetzner Cloud delivers exceptional price-to-performance ratios for European VPS, often beating Scaleway on pure cost for similar specifications. Hetzner operates data centers in Germany and Finland, with recent expansion to the US. The platform is simpler—fewer managed services, more focus on raw compute. Choose Hetzner for maximum cost efficiency on straightforward workloads; choose Scaleway for managed Kubernetes or broader service catalog.

DigitalOcean offers a middle ground between Scaleway's European focus and AWS's global scale. DigitalOcean provides US, European, and Asian regions with a strong developer community and marketplace. Pricing is slightly higher than Scaleway for equivalent compute, but DigitalOcean's documentation, tutorials, and third-party integrations are more extensive. DigitalOcean suits developers prioritizing ease of use and community resources over absolute lowest cost.

Final Verdict

Scaleway delivers solid value for teams prioritizing EU data residency, transparent pricing, and freedom from long-term contracts. The platform's strengths—granular instance configuration, accessible bare-metal servers, and honest hourly billing—make it a practical choice for European startups, development environments, and compliance-focused organizations.

The trade-offs are real: limited geographic reach, a smaller ecosystem than hyperscale providers, and support that won't hold your hand like premium enterprise tiers. Teams building global applications requiring sub-50ms latency worldwide should look elsewhere. Organizations deeply invested in AWS-specific services face migration friction.

For European SaaS companies, agencies hosting client projects, or development teams running CI/CD pipelines, Scaleway offers infrastructure that works without unnecessary complexity. The €7/month entry point for small instances and hourly bare-metal access lower barriers compared to traditional dedicated hosting. Data sovereignty compliance is straightforward, avoiding the multi-region gymnastics required with US-based providers.

Scaleway won't replace AWS for enterprises running thousands of microservices across six continents. It will save money and reduce vendor lock-in for focused use cases where European infrastructure, transparent billing, and standard technologies align with actual requirements rather than theoretical scaling needs.