Scaleway262GB VRAMBlackwell Ultraenterprise

B300 SXM6 on Scaleway

Visit Scaleway

Scaleway's NVIDIA B300 SXM6 offering brings the cutting-edge Blackwell architecture to European cloud users, featuring 262GB of HBM3e VRAM for handling massive AI models and datasets. As a major European provider, Scaleway emphasizes data sovereignty with instances hosted in France and the Netherlands, ideal for regulated industries like finance and healthcare. This enterprise-tier GPU excels in large-scale AI training, inference, and HPC workloads, delivering unprecedented performance through its 20 petaFLOPS FP8 tensor compute. Key value propositions include flexible per-hour billing for cost efficiency, integration with Scaleway's Nabu AI Supercomputer for multi-GPU scaling, and strong environmental credentials via renewable energy-powered data centers. ML engineers evaluating options will appreciate the combination of raw power, compliance with GDPR and EU regulations, and seamless integration with Object Storage, Kubernetes, and serverless services, making it a top choice for sovereignty-focused AI deployments without compromising on performance.

Why NVIDIA B300 SXM6 on Scaleway?

Choosing Scaleway for the NVIDIA B300 SXM6 leverages Europe's leading data sovereignty provider with Blackwell's immense 262GB VRAM tailored for trillion-parameter models. Scaleway's Nabu AI Supercomputer enables seamless multi-GPU clusters up to thousands of GPUs, perfect for distributed training. Per-hour billing offers flexibility for bursty workloads, reducing costs compared to commitment-based models. High-speed NVMe storage and up to 400Gbps networking complement the GPU's capabilities, while integrated services like RKE2 Kubernetes and Object Storage streamline workflows. Strong environmental focus with 100% renewable energy appeals to sustainable AI teams. For EU-based organizations, this combo ensures compliance without latency penalties from transatlantic data transfers.

Live Pricing

Real-time NVIDIA B300 SXM6 offers from Scaleway

3 listed, none in stock
Scaleway
Scaleway
Paris
Sold Out
NVIDIA B300 SXM68x
262GB VRAM
224 vCPU
3840GB RAM
22352GB Storage
20000 Mbps ↑
20000 Mbps ↓
$8.44/GPU/hr
$67.52/hr total (8×)
Scaleway
Scaleway
Paris
Sold Out
NVIDIA B300 SXM64x
262GB VRAM
112 vCPU
1920GB RAM
11176GB Storage
20000 Mbps ↑
20000 Mbps ↓
$9.90/GPU/hr
$39.61/hr total (4×)
Scaleway
Scaleway
Paris
Sold Out
NVIDIA B300 SXM62x
262GB VRAM
56 vCPU
960GB RAM
5588GB Storage
20000 Mbps ↑
20000 Mbps ↓
$11.02/GPU/hr
$22.04/hr total (2×)

Performance Notes

The NVIDIA B300 SXM6 on Scaleway delivers Blackwell's transformative performance, with up to 20 petaFLOPS in FP8 for AI training and inference on models exceeding 1 trillion parameters, thanks to 262GB HBM3e VRAM. Expect excellent multi-GPU scaling via Nabu clusters using NVLink or InfiniBand (up to 400Gbps confirmed in docs). NVMe SSD storage options provide high IOPS for data loading. Network bandwidth supports efficient all-reduce operations in distributed setups. Actual benchmarks are emerging post-launch; early reports suggest 2-3x gains over Hopper GPUs. Limitations: cooling and power density may cap dense packing—verify Scaleway's instance configs. Monitor provider updates for optimized drivers and CUDA 12.x support.

About Scaleway

A major European cloud provider emphasizing data sovereignty and integrated services.

Best For

European data sovereigntyIntegrated cloud services

Unique Features

  • Nabu AI Supercomputer
  • Strong environmental credentials
NVIDIA B300 SXM6 Specs

VRAM

262GB

Architecture

Blackwell Ultra

Tier

enterprise

Platform Features

Access Methods
SSH
Jupyter Notebooks
Web Terminal
API
Kubernetes
Containers
Billing Options
Incrementper-hour
Spot Instances
Reserved Instances
Prepaid Credits
Compliance
SOC 2
HIPAA
GDPR
ISO 27001

Getting Started

Scaleway simplifies launching NVIDIA B300 SXM6 instances via their intuitive console or API, with per-hour billing starting on-demand. New users benefit from a free trial credit and pre-configured AI images with NVIDIA drivers, CUDA, and frameworks like PyTorch. Focus on EU regions for low-latency sovereignty.

Steps

  1. 1Create a Scaleway account and Organization at console.scaleway.com.
  2. 2Set up billing and select Paris or Amsterdam region for sovereignty.
  3. 3Navigate to Instances > GPU, choose B300 SXM6, configure vCPU/RAM/storage.
  4. 4Select AI-optimized OS image (e.g., Ubuntu with CUDA), launch instance.
  5. 5Connect via SSH or Jupyter, install workloads using NGC containers.

Pro Tips

  • Use Scaleway's Elastic Metal for bare-metal B300 access if needing maximum performance without virtualization overhead.
  • Leverage Nabu House clusters for multi-GPU jobs; pre-warm instances to minimize cold-start times for iterative training.
  • Integrate with Scaleway Object Storage for datasets—enable snapshotting for quick recovery in long-running experiments.

Frequently Asked Questions

What is Scaleway's billing model for NVIDIA B300 SXM6?▾

Scaleway bills per-hour for GPU instances including NVIDIA B300 SXM6. 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 B300 SXM6?▾

No, Scaleway does not currently offer spot instances for NVIDIA B300 SXM6. 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 B300 SXM6?▾

Yes, Scaleway offers reserved instance pricing for NVIDIA B300 SXM6, 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

Compare B300 SXM6 Across Providers

The B300 SXM6 is available from 11 providers on GPUPerHour. Here is how other providers compare:

For a full comparison across all providers, see the B300 SXM6 rental page. See all GPUs on Scaleway.