H100 PCIe on Scaleway
Visit ScalewayScaleway's NVIDIA H100 PCIe offering delivers enterprise-grade AI acceleration through the Hopper architecture GPU with 80GB HBM3 VRAM, optimized for large-scale ML training, inference, and HPC workloads. As a leading European cloud provider, Scaleway emphasizes data sovereignty with instances hosted in France, ensuring GDPR compliance and low-latency access for EU-based teams. This combination stands out via integration with the Nabu AI Supercomputer for scalable clusters, alongside strong environmental credentials using low-carbon energy sources. Key value propositions include per-hour billing for cost flexibility, seamless integration with Scaleway's object storage, block storage, and Kubernetes services, and high performance for transformer models and generative AI. Ideal for ML engineers and data scientists prioritizing regulatory compliance, sustainability, and integrated European cloud ecosystems without sacrificing Hopper's 4x inference speedup over A100 or FP8 precision advantages. While PCIe form factor suits single-node efficiency, it pairs well with Scaleway's networking for distributed training.
Why NVIDIA H100 PCIe on Scaleway?
Choose Scaleway for NVIDIA H100 PCIe if data sovereignty is paramount—EU-hosted instances in Paris ensure GDPR adherence and minimal latency for European users. The provider's Nabu AI Supercomputer enables seamless scaling to H100 clusters, complementing the PCIe GPU's single-node prowess for prototyping or cost-sensitive workloads. Hourly billing offers granular control, reducing waste compared to longer commitments elsewhere. Scaleway's integrated services like Rclone-compatible object storage and managed Kubernetes enhance workflows, while strong environmental credentials (renewable energy focus) appeal to sustainability-conscious teams. This setup leverages Hopper's Transformer Engine without US provider dependencies, ideal for regulated industries like finance and healthcare.
Live Pricing
Real-time NVIDIA H100 PCIe offers from Scaleway
| Provider | GPU Model | VRAM | Host Specs | Region | Price | Status | Action | |
|---|---|---|---|---|---|---|---|---|
 Scaleway | 2×NVIDIA H100 PCIe 80GB VRAM | 80GB | 48 vCPU 480GB RAM 5588GB Storage | Warsaw | $3.24/GPU/hr $6.48/hr total (2×) | Available | ||
Scaleway | 2×NVIDIA H100 PCIe 80GB VRAM | 80GB | 48 vCPU 480GB RAM 5588GB Storage | Paris | $3.24/GPU/hr $6.48/hr total (2×) | Sold Out | ||
Scaleway | NVIDIA H100 PCIe 80GB VRAM | 80GB | 24 vCPU 240GB RAM 2794GB Storage | Paris | $3.24/GPU/hr | Available | ||
Scaleway | NVIDIA H100 PCIe 80GB VRAM | 80GB | 24 vCPU 240GB RAM 2794GB Storage | Warsaw | $3.24/GPU/hr | Available | ||
Scaleway | 2×NVIDIA H100 PCIe 80GB VRAM | 80GB | 48 vCPU 480GB RAM 5588GB Storage | Paris | $4.20/GPU/hr $8.40/hr total (2×) | Sold Out |
Performance Notes
Expect top-tier Hopper performance: up to 4P FLOPS FP8 for inference, 2P FP16 for training, excelling in LLMs up to 70B parameters on 80GB VRAM. Scaleway's GPU instances feature NVMe SSD storage (up to 15TB local) and 100-200Gbps networking, supporting efficient multi-GPU via NVLink or InfiniBand in Nabu clusters. PCIe variant offers solid single-node throughput but lower inter-GPU bandwidth than SXM (avoid for massive scaling without cluster). Benchmarks show H100 PCIe competitive for fine-tuning/inference; provider-specific results align with NVIDIA specs, though real-world varies by workload. No public Scaleway H100 PCIe benchmarks yet—test via short instances for your use case.
A major European cloud provider emphasizing data sovereignty and integrated services.
Best For
Unique Features
- Nabu AI Supercomputer
- Strong environmental credentials
VRAM
80GB
Architecture
Hopper
Tier
enterprise
Platform Features
Getting Started
Launching NVIDIA H100 PCIe on Scaleway is straightforward via their console, with per-hour billing and EU regions. New users can spin up instances in minutes, pre-loaded with NVIDIA drivers and CUDA for immediate ML workloads. Focus on Paris DC for sovereignty.
Steps
- 1Create a Scaleway account and add payment method for hourly billing.
- 2Navigate to Instances > GPU, select NVIDIA H100 PCIe (80GB), choose Paris region.
- 3Configure instance: select OS (Ubuntu), storage (NVMe), and networking (public IP).
- 4Launch instance and SSH in using provided key or console access.
- 5Install CUDA toolkit if needed, pull Docker images for PyTorch/TensorFlow.
Pro Tips
- Leverage Scaleway's Object Storage with Rclone for datasets; mount via FUSE for low-latency access during training.
- Use Nabu House for managed JupyterLab on H100 clusters to scale beyond single instances seamlessly.
- Monitor costs and performance via Scaleway Console metrics; start with spot instances for non-urgent jobs to save 50-70%.
Frequently Asked Questions
What is Scaleway's billing model for NVIDIA H100 PCIe?▾
Scaleway bills per-hour for GPU instances including NVIDIA H100 PCIe. Hourly billing means you pay for full hours even if your job completes mid-hour. Plan your workloads accordingly to maximize cost efficiency.
Does Scaleway offer spot instances for NVIDIA H100 PCIe?▾
No, Scaleway does not currently offer spot instances for NVIDIA H100 PCIe. All instances are billed at on-demand rates. However, they do offer reserved instances for committed usage, which can provide significant discounts for long-term workloads.
Does Scaleway offer reserved instances for NVIDIA H100 PCIe?▾
Yes, Scaleway offers reserved instance pricing for NVIDIA H100 PCIe, which can provide significant discounts (typically 20-40% off on-demand rates) for committed usage periods. Reserved instances are ideal for predictable, long-running workloads like production inference services, ongoing training pipelines, or development environments that run continuously. Contact Scaleway for current reserved pricing and commitment terms.
Related Pages
Rent NVIDIA H100 PCIe
Atlantic.net vs Scaleway: GPU Cloud Comparison
AWS vs Scaleway: GPU Cloud Comparison
Cirrascale vs Scaleway: GPU Cloud Comparison
NVIDIA B300 SXM6 on Scaleway - Pricing & Availability
NVIDIA H100 SXM5 on Scaleway - Pricing & Availability
NVIDIA L4 on Scaleway - Pricing & Availability
NVIDIA GeForce RTX 3070 on Scaleway - Pricing & Availability
NVIDIA H100 PCIe in Amsterdam, Netherlands - Pricing & Availability
NVIDIA H100 PCIe in Canada - Pricing & Availability
NVIDIA H100 PCIe in California, United States - Pricing & Availability
NVIDIA H100 PCIe in Czechia - Pricing & Availability
NVIDIA H100 PCIe in France - Pricing & Availability