DevOps

Scaleway vs Hetzner: the definitive 2026 comparison for European cloud

Scaleway vs Hetzner:

Scaleway vs Hetzner is one of the most common comparisons for teams evaluating European cloud providers in 2026. If you’ve been researching EU-based infrastructure, you’ve likely already encountered the Scaleway vs Hetzner debate — and realized that the choice is not as straightforward as pricing tables suggest.

The honest answer is that they’re solving different problems. Hetzner is built for engineers who want maximum performance per euro and are comfortable managing their own stack. Scaleway is built for teams that want the managed-service depth of a hyperscaler inside a sovereign European legal framework. Getting the choice wrong means either paying for abstraction you don’t need or burning senior engineering hours on infrastructure plumbing that a managed service would have handled in minutes.

At Gart Solutions, we’ve deployed workloads on both platforms across dozens of client engagements — FinTech, HealthTech, SaaS, and high-traffic media. This article lays out exactly what each provider does well, where each falls short, and which type of organisation should choose which.

TL;DR Summary

Choose Hetzner

If cost is the overriding constraint, you have strong DevOps capability, and you run compute-heavy workloads without needing managed services.

Choose Scaleway

If you need managed Kubernetes, AI/GPU infra, or specific EU compliance (HDS/SecNumCloud) with a developer experience closer to AWS.

Company overview: who is behind each platform?

Hetzner Online GmbH was founded in 1997 in Germany and remains 100% privately held. Its business model is vertically integrated: Hetzner builds and owns its own data centres, designs its own server configurations, and operates everything from the hardware up. This ownership structure is what makes its pricing so aggressive — there are no middlemen, no hyperscaler margins.

Scaleway was founded in 1999 and operates as a subsidiary of the French Iliad Group (parent company behind the Free mobile and broadband network). It has evolved from a budget VPS provider into a full-stack cloud platform with 100+ products spanning compute, storage, serverless, Kubernetes, and AI inference. Its ambition is explicit: to become the sovereign AI cloud of Europe.

Neither company has a US parent, which is the key structural fact for compliance teams. Both are outside the scope of the US CLOUD Act — meaning US authorities cannot compel either provider to hand over European customer data through a US court order.

Scaleway vs Hetzner: data sovereignty and regulatory compliance

Governance attribute Hetzner Scaleway
Jurisdiction Germany (EU) France (EU)
CLOUD Act exposure None None
GAIA-X participation Non-participant Active member
Certifications ISO 27001, GDPR, BSI C5 ISO 27001, HDS, SecNumCloud
Management console infra Primarily EU-hosted Partially uses US-based services

For most organisations, both providers clear the sovereignty bar. The nuances matter at the margin. Scaleway holds HDS (French healthcare data hosting) and SecNumCloud certifications, making it the stronger choice for regulated sectors like HealthTech and FinTech operating under French or EU-specific frameworks. Hetzner holds the BSI C5 attestation — the German Federal Office for Information Security’s cloud compliance standard, increasingly required for public-sector and critical infrastructure contracts in the DACH region.

One caveat worth flagging: independent analysis from 2025 noted that Scaleway’s management console relies on some US-hosted infrastructure. This doesn’t affect the residency of your stored data or workloads, but organisations with the most extreme sovereignty mandates should verify this against their specific requirements. Hetzner’s console infrastructure is primarily EU-hosted throughout.

Navigating GDPR, HDS, or BSI C5 requirements? Gart Solutions’ compliance architecture practice helps regulated organisations map provider capabilities to specific legal frameworks — before a regulator does it for you.
Learn about our data sovereignty practice

Pricing and cost structure

This is where the comparison becomes striking. Both providers undercut US hyperscalers by a significant margin, but they do it differently.

Hetzner’s model is simple: fixed monthly or hourly pricing for VMs, dedicated servers, and storage, with a generous included traffic allowance. Cloud instances ship with 20 TB to 60 TB of free monthly egress depending on the node. Overage is billed at approximately €1.18 per TB — a fraction of AWS’s egress rates.

Scaleway’s model offers free egress for most services, but applies bandwidth caps based on instance size. For workloads that exceed those caps, traffic is throttled rather than billed.

14.3×

Hetzner value vs AWS (compute)

4.8×

Scaleway value vs AWS (compute)

~10%

Hetzner CPX32 cost vs comparable AWS instance

In practical terms, a 300 TB monthly traffic profile that costs roughly $21,000 on AWS can be served for effectively zero on Hetzner (distributed across multiple high-traffic nodes within the free allowance), or for close to zero on Scaleway within its throttling regime. For an engineering team running cost-sensitive workloads — batch processing, media transcoding, large-scale data pipelines — the Hetzner economics are genuinely transformative.

Compute performance and infrastructure

Both providers run AMD EPYC processors as their primary x86 compute platform. Hetzner has introduced the CAX line (Ampere ARM64), offering excellent performance-per-watt for parallelisable workloads. Scaleway offers Apple Silicon on bare metal for macOS CI/CD use cases — an unusual differentiator.

In comparative testing, Hetzner instances show lower CPU variance across runs, indicating a tighter hypervisor environment with less aggressive oversubscription. Scaleway integrates instances into a more complex VPC networking stack by default, which introduces marginal overhead but provides greater architectural flexibility.

Performance metric Hetzner CCX/CPX Winner Scaleway PRO/PLAY
Peak sequential read ~3,728 MB/s ~2,298 MB/s
Network download speed ~1,811 Mbps ~1,673 Mbps
Performance consistency 71/100 65/100
Instance provisioning time ~25 seconds ~30 seconds
CPU architecture options AMD EPYC, Intel Xeon, Ampere ARM AMD EPYC, Apple Silicon (bare metal)

Scaleway vs Hetzner: Storage

Hetzner’s storage portfolio covers Cloud Volumes (block), Object Storage (S3-compatible), and Storage Boxes — a distinctive multi-protocol network storage service supporting SMB, WebDAV, and SSH. Storage Boxes are particularly well-suited for centralised backup and file storage outside the standard cloud VM paradigm.

Scaleway’s object storage offers three tiers: Standard, One Zone-IA, and Glacier — a deep archival tier under 100% European jurisdiction. For organisations that previously relied on AWS Glacier for cold storage, Scaleway’s Glacier removes the legal exposure of a US-jurisdiction archive with no sacrifice in cost efficiency.

Storage metric Hetzner Cloud Volumes Scaleway Block Storage
Maximum volume size 10 TB 15 TB
Larger capacity
IOPS tiers Standard SSD 5k or 15k
Snapshot support Yes (manual/scheduled) Yes (encryption-ready)
Scaling Vertical expansion Resize on-demand
No downtime

Scaleway’s ability to resize block volumes without downtime is a meaningful operational advantage for production databases with growing datasets. Hetzner’s storage is reliable but has faced criticism for longer recovery times during storage-related incidents — a product of its more centralised architecture compared to Scaleway’s Multi-AZ approach.

Networking

Scaleway’s Virtual Private Cloud (VPC) is the more mature offering, supporting full regional isolation, Multi-AZ VPC architectures, VPC Peering, and integrated IP Address Management (IPAM) — capabilities significantly enhanced in 2025.

Hetzner made significant strides with the launch of the unified Hetzner Console, which introduced the ability to provision servers without public IP addresses — important for high-security database clusters. The remaining friction is the split between Hetzner Cloud (VMs) and Hetzner Robot (dedicated servers): teams operating across both must still manage vSwitch configurations and MTU settings manually.

Bandwidth metric Hetzner (EU regions) Scaleway
Included traffic 20–60 TB per instance
Winner
Free (most services)
Subject to FUP
Overage cost ~€1.18/TB N/A (throttled)
No unexpected bills
DDoS mitigation Advanced (AI-powered)
Winner
Standard

Managed Kubernetes and container orchestration

Scaleway Kapsule is a mature managed Kubernetes service with a free control plane, native integration with Scaleway load balancers and block storage, and support for Cilium and Calico as CNI options. Scaleway Kosmos extends this to a multi-cloud Kubernetes engine — you can manage nodes from Hetzner, AWS, or bare metal from a single Scaleway control plane.

Hetzner does not offer a native managed Kubernetes service. Teams typically use community-driven tooling like kube-hetzner and hetzner-k3s, which integrate the Hetzner Cloud Controller Manager and CSI driver. These tools are production-grade, but they shift the operational burden to your team: control plane upgrades, etcd backups, certificate rotation, and node auto-healing all require explicit automation.

Kubernetes feature Scaleway Kapsule Winner Hetzner (DIY)
Control plane management Managed (free) Self-managed
Multi-cloud nodes Yes (via Kosmos) Native support for external nodes Manual (Wireguard)
Auto-healing nodes Native Community-automated
Serverless integration Yes No
CNI options Cilium, Calico User-defined

When advising clients on Kubernetes platform selection, we at Gart Solutions weigh this trade-off carefully against team composition. A platform engineering team of 3+ can comfortably run Hetzner-based Kubernetes at significant cost savings. A 1–2 person DevOps function that also owns application deployments generally benefits more from the reduced cognitive load of a managed control plane.

Evaluating Kubernetes platform options for your team?

Gart Solutions’ DevOps and platform engineering practice has implemented production Kubernetes on both Hetzner and Scaleway — we can model the right trade-off for your team size and operational model.

Learn about our platform engineering practice

Managed databases and application platform

Scaleway’s managed database portfolio covers PostgreSQL (versions 12–17), MySQL 8, Redis, and MongoDB, with high availability through standby nodes and automated logical backups. In 2025, Scaleway added managed ClickHouse for analytics and a serverless SQL database that scales to zero when idle — an economical option for development environments or low-traffic APIs.

Hetzner does not offer a managed cloud database API. Databases on Hetzner are self-managed VMs — you’re responsible for clustering, backups, point-in-time recovery, and version upgrades. The upside is total control and lower cost; the downside is that Day 2 database operations are entirely your problem.

Database capability Scaleway Managed RDB Winner Hetzner Cloud VM (DIY)
One-click cloning Yes Manual volume snapshot
Point-in-time recovery Yes Manual setup required
High availability Standby nodes (Multi-AZ) Automatic failover Manual replication
Storage hot-resize Up to 10–15 TB On-demand expansion Manual volume resize
Serverless SQL Yes Scales to zero when idle No

AI and GPU infrastructure

The GPU landscape is where Scaleway’s product strategy comes into its own. As of 2026, Scaleway operates NVIDIA H100 SXM, L40S, and L4 instances on-demand, alongside a Model-as-a-Service (MaaS) layer exposing foundation model inference via API — similar to OpenAI, but hosted under European legal jurisdiction. Its partnership with Hugging Face for sovereign inference makes it the de facto platform for European AI workloads requiring data residency guarantees.

Hetzner’s GPU offering is focused on dedicated hardware: the GEX44 (RTX 4000 Ada) and GEX46 (RTX 6000 Ada) available on monthly leases. This suits long-running inference deployments or render farms where monthly lease economics beat hourly on-demand pricing. But Hetzner provides no AI software stack and no managed inference.

GPU metric Scaleway Winner (AI) Hetzner
Deployment model On-demand cloud instance Dedicated server (monthly)
H100 availability Yes (SXM and PCIe) Available for immediate scale Limited / specialised
AI software stack Native MaaS + Hugging Face Ready-to-use inference APIs Manual driver/stack setup
Starting price (entry GPU) ~€0.83/hr (L4) ~€0.34/hr (RTX 4000)
Sovereign inference Yes 100% EU legal framework Hardware-level only

Developer experience and Infrastructure as Code

Scaleway is the stronger DX platform by most measures. Its Terraform provider covers virtually every product — DNS, serverless functions, Kubernetes, object storage, managed databases — and the credential flow generates ready-to-use Terraform snippets. A developer can go from account creation to a deployed S3-compatible bucket in approximately 20 minutes.

Hetzner’s DX has improved significantly with the 2025–2026 unification of the Hetzner Console, merging Storage Boxes, DNS, and cloud VMs into a single interface. The Hetzner Terraform provider is stable and fast, but narrower in scope given the smaller managed services catalogue. Hetzner’s API is particularly praised for its simplicity and provisioning speed — VMs are typically available in approximately 25 seconds.

Documentation quality diverges clearly: Scaleway’s documentation is closer to the hyperscaler standard, with architectural guides and detailed tutorials. Hetzner’s documentation is more reference-oriented and assumes a higher baseline of systems administration knowledge.

Support, SLA, and operational risk

Reliability factor Hetzner Scaleway
SLA guarantee 99.9% 99.9%
Support channels Tickets & Phone Phone for dedicated servers Tickets, 24/7 Phone, Slack Winner
Standard response time Variable (several hours) 2–4 hours
Account suspension risk High
Aggressive KYC
Moderate
Resilience model Host failure = downtime Multi-AZ failover
Winner

The most significant operational risk with Hetzner is its account verification process. To combat abuse at its price point, Hetzner operates an aggressive KYC and fraud detection system that has resulted in account terminations — sometimes during active migrations, with minimal explanation. For a business with a hard migration deadline, this is a real risk to plan around. Mitigation: verify your account fully before beginning any critical cutover, and maintain a warm standby on a secondary provider during the transition window.

Scaleway’s primary operational risk is platform maturity: some newer managed products have experienced instability as they’ve grown. Standard support response times can be variable. For mission-critical workloads, Scaleway’s 24/7 phone support and Multi-AZ architecture provide adequate business continuity guarantees — but thorough failover testing before go-live is recommended.

Who should choose Hetzner?

  • Cost-sensitive startups running stateless web applications, API servers, or batch processing at scale
  • High-bandwidth services (media, CDN origins, data distribution) that benefit from the 20–60 TB free egress allowance
  • Strong DevOps organisations comfortable managing their own Kubernetes, databases, and monitoring
  • Data archiving and backup use cases leveraging Storage Boxes and Object Storage
  • Long-running GPU compute (rendering, ML training) where a monthly dedicated lease makes economic sense

Who should choose Scaleway?

  • Cloud-native development teams that want managed Kubernetes, serverless, and databases without operational overhead
  • Regulated industries (HealthTech, FinTech, public sector) requiring HDS, SecNumCloud, or GAIA-X certification
  • AI and ML workloads requiring H100 on-demand or sovereign model inference under EU jurisdiction
  • Organisations migrating from AWS who want a familiar managed-service model inside EU legal boundaries
  • Small teams (1–2 DevOps engineers) where reducing the managed services footprint materially lowers operational burden

Conclusion: Scaleway vs Hetzner

The Scaleway vs Hetzner decision is ultimately a question of how much infrastructure complexity your team can absorb — and at what cost.

Hetzner is the right choice when raw compute efficiency is the primary objective and your engineering team has the skills to build and maintain the services that Hetzner deliberately doesn’t offer. The economics are genuinely extraordinary and can change a company’s unit economics at scale.

Scaleway is the right choice when you need managed service depth that lets developers move fast — Kubernetes without control plane overhead, databases without Day 2 operations, AI inference without building your own stack — all within a sovereign European legal framework with meaningful compliance certifications.

Both platforms represent a sophisticated, mature alternative to US hyperscalers. The “European cloud is only for budget workloads” narrative is outdated. In 2026, both Hetzner and Scaleway are capable of supporting demanding enterprise and AI workloads with legal protections and pricing transparency that AWS and GCP cannot match. Choose based on team composition and operational capability, not just the price list.

Let’s work together!

See how we can help to overcome your challenges

FAQ

Is Hetzner cheaper than Scaleway?

For raw compute, yes — Hetzner typically offers the lowest price-per-vCPU in Europe. A comparable virtual machine on Hetzner can cost 30–50% less than the equivalent on Scaleway. However, when you factor in the engineering time required to self-manage Kubernetes, databases, and networking on Hetzner, the total cost of ownership can favour Scaleway for teams without deep infrastructure expertise. Price-per-euro of compute is only one variable in the full equation.

Which provider is better for GDPR compliance?

Both are structurally sound for GDPR: neither has a US parent company and neither is subject to the US CLOUD Act. For regulated industries requiring specific certifications — French healthcare (HDS), French national security cloud (SecNumCloud), or GAIA-X interoperability — Scaleway holds the relevant accreditations. For DACH public sector and critical infrastructure, Hetzner's BSI C5 attestation is often the required certification. Match the provider to your specific regulatory framework, not just the GDPR checkbox.

Can I run Kubernetes on Hetzner in production?

Yes, and many organisations do successfully. Community tooling like kube-hetzner and hetzner-k3s provides production-grade automation for cluster lifecycle management, including the Hetzner Cloud Controller Manager and CSI driver. The trade-off is that your team owns control plane upgrades, etcd backups, and certificate rotation. For teams with a dedicated platform engineering function, this is a manageable and cost-effective approach. For smaller teams, Scaleway's managed Kapsule service removes significant operational burden.

How does Scaleway's GPU offering compare to Hetzner for AI workloads?

Scaleway has a substantial lead for AI workloads requiring on-demand GPU access and managed inference. It offers NVIDIA H100 SXM and L40S instances, a Model-as-a-Service layer for sovereign inference, and a partnership with Hugging Face. Hetzner's GPU offering is focused on monthly-lease dedicated servers (RTX 4000/6000 Ada), which suits long-running persistent workloads but lacks the agility of on-demand scaling and provides no managed AI software stack. For LLM training or high-volume inference under GDPR, Scaleway is the stronger choice in Europe.

What is the risk of account suspension on Hetzner?

It is a documented operational risk. Hetzner's pricing attracts high volumes of abuse attempts, and its automated KYC system has terminated legitimate accounts, sometimes during active migrations and without detailed explanation. This risk is more pronounced for businesses based in South Asia or for newly created accounts provisioning significant resources quickly. Mitigation strategies include verifying your account fully before beginning a critical migration and maintaining a warm standby deployment on a secondary provider during cutover.

How do I decide between Scaleway and Hetzner for a microservices architecture?

The key variables are team size and managed service requirements. If your microservices architecture relies on managed databases, managed Kubernetes, serverless functions, and observability — Scaleway's ecosystem supports all of this natively. If your team prefers to own the infrastructure layer and wants to optimise cost aggressively, Hetzner as the compute substrate with community tooling for orchestration is a viable and well-proven pattern. Gart Solutions' cloud architects regularly help teams map their specific architectural requirements to the right European provider — reach out for a vendor-neutral assessment at gartsolutions.com/services/cloud-architecture.

Is Scaleway stable enough for production workloads?

Scaleway's core infrastructure — compute, object storage, managed Kubernetes — is production-grade and used by large European organisations. Some newer managed products have experienced instability as they've matured. Standard support response times can be variable. For mission-critical workloads, Scaleway's 99.9% SLA, 24/7 phone support, and Multi-AZ architecture provide adequate business continuity guarantees — but thorough testing of failover behaviour before go-live is recommended, as it is with any provider.
arrow arrow

Thank you
for contacting us!

Please, check your email

arrow arrow

Thank you

You've been subscribed

We use cookies to enhance your browsing experience. By clicking "Accept," you consent to the use of cookies. To learn more, read our Privacy Policy