deploybase

Scaleway GPU Cloud Pricing: Complete Guide for Every GPU (March 2026)

Deploybase · November 10, 2025 · GPU Pricing

Contents

Scaleway GPU Pricing Overview

Scaleway is a European cloud provider with GPUs at competitive rates. Mid-range globally, excellent in Europe.

EU data residency? Scaleway makes sense. GPU instances are newer than AWS but supply is growing. This covers pricing and optimization.

Who Scaleway Works For

  • European users (compliance)
  • Budget-conscious teams
  • Privacy-first developers
  • EU infrastructure requirements

Smaller than the big clouds but growing fast. Reliability has improved. Multi-language support.

GPU Instance Pricing

H100 GPU Instances

Single H100 (80GB):

  • Hourly: ~$2.80/hr
  • Monthly: ~$1,904/month (18% discount vs hourly)
  • Memory: 128GB RAM
  • vCPUs: 32 cores
  • Storage: 200GB SSD

Scaleway H100 pricing is competitive. $2.80/hr is comparable to RunPod and significantly cheaper than AWS (~$6.88/GPU/hr on p5 instances) and Vultr ($4+/hr).

Cost comparison:

  • Scaleway: $2.80/hr
  • RunPod: $2.69/hr (similar)
  • AWS p5 H100: ~$6.88/GPU/hr (8-GPU p5.48xlarge $55.04/hr total, ~2.5x more)
  • Vast.AI: $2.00-3.50/hr (variable)

Scaleway is competitive globally for H100.

A100 GPU Instances

Single A100 (40GB):

  • Hourly: ~$0.89/hr
  • Monthly: ~$605/month
  • Memory: 64GB RAM
  • vCPUs: 16 cores
  • Storage: 100GB SSD

A100 pricing at $0.89/hr is exceptional for a managed EU cloud provider.

Cost comparison:

  • Scaleway: $0.89/hr
  • Google Cloud: ~$2.93/hr (3x more)
  • RunPod: $1.39/hr (56% more)
  • DigitalOcean: $2.68/hr (3x more)
  • Vast.AI: $0.80-1.50/hr (variable, less reliable)

Scaleway A100 is outstanding value among managed cloud providers.

L4 GPU Instances

Single L4 (24GB):

  • Hourly: ~$0.87/hr
  • Monthly: ~$591/month
  • Memory: 32GB RAM
  • vCPUs: 8 cores

Inference GPU at competitive price for a managed EU provider.

Cost comparison:

  • Scaleway: $0.87/hr
  • Vast.AI: $0.12-0.20/hr (cheaper but less reliable)
  • Google Cloud g2: ~$0.71-1.00/hr
  • AWS g4dn: ~$0.80/hr

Scaleway L4 is competitively priced for a managed EU cloud provider, though budget marketplaces like Vast.AI are cheaper.

Multi-GPU Instances

Scaleway offers configurations with 2-4 GPUs:

Dual H100:

  • Hourly: ~$5.60/hr (2x single)
  • Monthly: ~$3,808/month

Quad A100:

  • Hourly: ~$3.56/hr (4x single)
  • Monthly: ~$2,420/month

Pricing scales linearly. No bulk discount.

Regional Pricing

Scaleway has data centers in Europe, North America, and Asia:

European regions (Paris, Amsterdam):

  • Base pricing (most competitive)

North American regions:

  • +10-15% vs Europe

Asia-Pacific regions:

  • +20-25% vs Europe

Europe is cheapest. Scaleway's strength in EU market is reflected in pricing.

For global teams, EU data centers offer better pricing than US options. Privacy-conscious projects benefit.

Pricing Models

Hourly Billing

Pay per hour, stop anytime. No commitment required. Most flexible.

Monthly Commitment

18-20% discount for monthly commitment. Instance is "reserved" for the month. Early termination loses remainder.

A100 example:

  • Hourly: $0.89/hr = $649.70/month (730 hours)
  • Monthly: $605/month
  • Savings: $44.70/month (7%)

Monthly is better for continuous workloads.

Annual Commitment

Multi-year commitments are less common on Scaleway than competitors. Inquire for custom pricing on large orders.

Real-World Pricing Examples

Small inference project (10 hours/week):

  • Scaleway L4: 10 × $0.87 = $8.70/week
  • Vast.AI L4: 10 × $0.15 = $1.50/week
  • AWS g4dn: 10 × $0.80 = $8.00/week

Scaleway is comparable to AWS for managed EU infrastructure. Vast.AI is cheaper but less reliable.

Medium training (100 hours/month):

  • Scaleway A100: 100 × $0.89 = $89/month
  • Vast.AI A100: 100 × $1.00 = $100/month (variable)
  • Google Cloud A100: 100 × $2.93 = $293/month
  • RunPod A100: 100 × $1.39 = $139/month

Scaleway is very competitive. RunPod and Vast.AI are also affordable alternatives. Google Cloud is significantly more expensive for managed A100.

Heavy training (24/7 for 3 months):

  • Scaleway A100: 730 × $0.89 × 3 = $1,815
  • Scaleway (monthly): $605 × 3 = $1,815
  • Vast.AI (average): $730 × 3 = $2,190
  • Google Cloud A100: ~$2,139/month × 3 = $6,417
  • AWS p4: ~$1,825/month × 3 = $5,475

Scaleway is an excellent value for longer commitments, significantly undercutting both AWS and Google Cloud for A100 workloads.

Competitive Pricing Analysis

A100 GPU hourly:

  • Scaleway: $0.89
  • Vast.AI: $0.80-1.50 (variable, less reliable)
  • RunPod: $1.39 (56% more)
  • Google Cloud: ~$2.93 (3x more)
  • DigitalOcean: $2.68 (3x more)
  • Vultr: $2.80 (3x more)

H100 GPU hourly:

  • Scaleway: $2.80
  • RunPod: $2.69 (4% cheaper)
  • AWS p5 (per GPU): ~$12.25 (4x+ more)
  • Vast.AI: $2.00-3.50
  • DigitalOcean: N/A

L4 GPU hourly:

  • Scaleway: $0.87
  • Vast.AI: $0.12-0.20 (significantly cheaper but less reliable)
  • Google Cloud g2: ~$0.71-1.00/hr
  • AWS g4dn: ~$0.80/hr

Scaleway L4 is competitive for a managed EU provider.

Cost Optimization Strategies

1. Use monthly commitment

7-10% discount for monthly billing. Worth it for continuous workloads.

2. Right-size the instance

L4 at $0.87/hr for inference. Pick based on actual needs.

3. Use European pricing

If teams are in Europe, Scaleway's pricing is best in region. EU data residency is a bonus.

4. Batch processing

Run multiple jobs sequentially. Reduces startup overhead.

5. Stop instances when idle

Stopped instances don't charge for compute. Storage charges continue but are minimal.

6. Monitor consumption

Scaleway's dashboard is transparent. Set alerts for budget overruns.

7. Consider multi-month commitments

For projects running 3+ months, bulk pricing might be available. Inquire with sales.

Storage and Bandwidth

Block Storage: $0.06/GB/month (cheaper than most competitors)

Data Transfer:

  • Outbound: $0.10/GB (higher than some providers)
  • Between Scaleway zones: Free
  • Inbound: Free

Keep data in Scaleway. Transfer costs add up for large datasets.

European Data Residency Advantage

Scaleway's key differentiator is EU data centers. For teams requiring:

  • GDPR compliance
  • EU data residency
  • European infrastructure isolation

Scaleway is natural choice. Cost is secondary to compliance.

German teams especially benefit from Scaleway's strong EU presence.

Reliability and Support

Scaleway offers:

  • 99.9% SLA for Compute instances
  • Email support (free) and priority support (paid)
  • Growing documentation
  • Community forums

Support is less mature than AWS/Azure but adequate. SLA is solid.

Uptime is generally good. Less documentation than major clouds but improving.

Integration with Scaleway Ecosystem

Scaleway offers complementary services:

  • Object Storage: S3-compatible, $0.015/GB/month
  • Managed Databases: PostgreSQL, MySQL
  • Container Registry: Docker image hosting
  • Managed Kubernetes: Container orchestration

Integration is good for teams already using Scaleway infrastructure.

Limitations

Smaller ecosystem: Fewer pre-built integrations vs AWS/GCP

Regional spread: Primarily EU. Limited US/Asia options.

Newer platform: Less mature than decade-old competitors

Documentation: Good but less extensive than major clouds

Support: Adequate but not 24/7 premium support like AWS

These limitations are acceptable for many workloads. For bleeding-edge deployments, major clouds are safer.

When to Use Scaleway

Use Scaleway for:

  • European teams (best EU pricing)
  • GDPR/data residency requirements
  • Cost-conscious budgets
  • A100 inference/training
  • Teams already on Scaleway platform
  • Projects willing to trade support for cost savings

Avoid Scaleway for:

  • Mission-critical production (use AWS/GCP)
  • latest hardware requirements (older batches)
  • 24/7 premium support needs
  • Complex multi-region deployments
  • Highly regulated industries requiring extensive audits

Scaleway is excellent for European teams with standard ML needs and moderate budgets.

FAQ

Is Scaleway cheaper than AWS?

For A100, yes — Scaleway at $0.89/hr versus AWS at ~$2.93/hr is significantly cheaper. For H100, Scaleway at $2.80/hr is also far more affordable than AWS's ~$12.25/GPU/hr on p5 instances. AWS has more production features and global coverage.

Should I use Scaleway or Vast.AI?

Vast.AI is cheaper but less reliable. Scaleway is more reliable but more expensive. Choice depends on risk tolerance.

Can I run production ML on Scaleway?

Yes. SLA is 99.9%. Reliability is solid. Suitable for production.

What GPU should I start with on Scaleway?

L4 at $0.87/hr for inference/learning. A100 at $0.89/hr for training. H100 at ~$2.80/hr for heavy lifting.

Does Scaleway support PyTorch?

Yes. All major ML frameworks work. Scaleway provides Linux instances; you install your tools.

Is data stored in Scaleway encrypted?

Storage can be encrypted. Check security settings during instance creation.

Can I scale horizontally on Scaleway?

Yes. Launch multiple instances. Orchestrate with Kubernetes or manually. Multi-GPU instances available.

How do I minimize my Scaleway bill?

Use monthly commitment. Right-size instances. Stop when idle. Use L4 for small jobs.

Sources

  • Scaleway Instance Pricing (as of March 2026)
  • Scaleway Documentation and Specifications
  • Cloud Pricing Comparisons (March 2026)
  • European Cloud Provider Analysis

Last updated: March 2026. Pricing reflects market rates as of March 22, 2026.