deploybase

Scaleway GPU Cloud Review: European Alternative

Deploybase · November 11, 2025 · GPU Cloud

Contents

Scaleway Overview

As of March 2026, Scaleway is a French cloud provider owned by Iliad, operating data centers across Europe. Founded in 2014, Scaleway has grown to serve thousands of customers prioritizing European data residency. The platform emphasizes simplicity, transparency, and straightforward pricing.

Scaleway differentiates through European focus. All data centers reside within EU borders. Pricing displays in EUR with clear tax implications. Support operates in French and English. Customer service is based in Europe.

The company has invested heavily in GPU infrastructure over the past three years. GPU cloud services launched with A100 and H100 availability. Pricing undercuts North American alternatives by 15-25% on most configurations.

GPU Offerings and Availability

Scaleway provides A100 SXM GPUs in their Paris data centers. Single A100 instances cost approximately EUR 2.50 per hour ($2.70 at current rates). Eight-GPU bundles cost EUR 20 per hour. This pricing undercuts AWS by 30% on A100.

H100 SXM systems are available with slightly higher pricing. Single H100 costs EUR 3.50 per hour. Eight-GPU clusters cost EUR 28 per hour. This represents competitive pricing for Hopper-generation performance.

RTX 4090 GPUs serve smaller workloads and research projects. Scaleway offers RTX 4090s at approximately EUR 1.80 per hour. This positions RTX 4090 as viable alternative for inference workloads.

Availability varies by GPU type. A100s have consistent availability in Paris. H100 capacity shows seasonal variation with higher demand during fall. RTX 4090s offer best availability. Capacity planning is recommended for production workloads.

Pricing Analysis

Scaleway pricing is transparent with no hidden fees. Compute hourly rates include all resources. Storage costs EUR 0.08 per GB-month for persistent volumes. Bandwidth egress costs EUR 0.01 per GB.

Annual contract discounts of 10-20% apply to committed capacity. Teams committing to 12-month rentals save substantially. This pricing structure favors stable workloads.

Comparison to AWS shows Scaleway advantages. A100 on AWS costs around $6.88 per hour for single-GPU instances. Scaleway charges $2.70 per hour for A100, a significant discount. H100 on AWS (p5 instances) costs approximately $6.88 per GPU per hour (8-GPU p5.48xlarge at $55.04/hr). Scaleway's H100 at EUR 3.50 (~$3.80) is cheaper, while also including native EU data residency and GDPR compliance. RTX 4090 on Scaleway costs $1.80 versus Lambda's $0.92 for Quadro RTX 6000.

Comparison to Lambda Labs shows mixed results. Lambda A100 costs $1.48 per hour. Scaleway's $2.70 is nearly 2x higher. For teams prioritizing European residency, the cost difference is justified.

GDPR and Data Compliance

Scaleway operates all infrastructure within EU borders. GDPR compliance is built-in, not an afterthought. Data never leaves European jurisdictions. This meets strict requirements for teams processing EU personal data.

Scaleway Data Protection Officer certifies compliance. Annual audits verify adherence to privacy standards. Documentation provides clear evidence of GDPR compliance for audit purposes.

Subprocessor agreements are straightforward. Scaleway publishes lists of subprocessors. No surprise third-party data sharing occurs. This transparency appeals to privacy-conscious teams.

Service Level Agreements guarantee 99.99% uptime for production workloads. Compensation applies for SLA breaches. This contractual obligation protects against unexpected downtime.

Performance and Reliability

GPU performance matches specifications across regions. A100 systems deliver consistent 312 teraflops of tensor performance. Network latency within Paris data center averages 2-3 milliseconds. Cross-region latency reaches 50-80 milliseconds.

Reliability is solid but not exceptional. Uptime typically exceeds 99.95% monthly. Scheduled maintenance windows occur quarterly. Unplanned outages are rare, averaging less than 30 minutes annually.

Support response times average 4 hours for technical issues. Priority support reduces response times to 1 hour. Email-based support is the primary channel. Phone support is available for production customers.

Kubernetes integration simplifies GPU management. Scaleway provides managed Kubernetes with GPU node pools. This integration reduces operational complexity for container deployments.

Comparison with Alternatives

Paperspace provides European alternatives through partners but doesn't operate EU data centers. Pricing is competitive but lacks GDPR-native compliance.

CoreWeave operates in Amsterdam serving EMEA. Eight A100 GPUs cost $21.60 per hour globally. CoreWeave's pricing is comparable to Scaleway's on A100.

AWS offers EU data centers but with regional pricing markups. Irish data center pricing exceeds Paris by 5-10%. Contractual flexibility is superior to Scaleway.

Google Cloud operates in Belgium and Finland. Pricing matches AWS with regional variations. G Cloud integrates with Vertex AI for managed ML.

Scaleway combines competitive pricing, GDPR compliance, and European presence effectively. Teams requiring EU residency save money compared to AWS. Teams comfortable with US-based infrastructure find better pricing elsewhere.

FAQ

Q: Is Scaleway good for GDPR compliance? A: Yes. Scaleway operates exclusively within EU borders. GDPR compliance is built-in. All data processing occurs within EU jurisdiction.

Q: How much cheaper is Scaleway than AWS? A: A100 on Scaleway costs $2.70 per hour. For H100, Scaleway ($3.80/hr) is cheaper than AWS p5 instances ($6.88/GPU/hr from the 8-GPU $55.04/hr cluster). For teams prioritizing European residency, Scaleway offers clear cost advantages especially on H100.

Q: What GPUs does Scaleway offer? A: Scaleway offers A100 SXM, H100 SXM, and RTX 4090. Availability is good for A100 and RTX 4090. H100 capacity varies seasonally.

Q: Does Scaleway provide managed services? A: Scaleway offers managed Kubernetes with GPU node pools. Direct API access enables custom deployments. Limited managed ML services compared to AWS or Google Cloud.

Q: What is Scaleway's uptime SLA? A: Scaleway guarantees 99.99% uptime for production workloads. Historical uptime exceeds this target.

Sources