Power and stability of dedicated GPU hardware with seamless integration into the Scaleway Public Cloud ecosystem.
Elastic Metal Titanium delivers dedicated, non-shared GPU resources for stable, high-powered performance essential for AI training, deep learning, and large-scale data processing, while seamlessly integrating with Scaleway’s cloud ecosystem for enhanced flexibility in scaling, storage, and data flow management.
With dual GPUs, high-bandwidth networking, and NVMe SSD storage, Elastic Metal Titanium ensures rapid data transfer, low latency, and accelerated model training. This makes it ideal for industries requiring real-time analytics and faster insights, combining powerful hardware with cloud-based scalability.
Elastic Metal Titanium offers fixed monthly pricing, giving businesses long-term budget control over their GPU-intensive projects. By avoiding the unpredictability of cloud billing, enterprises gain the stability of dedicated servers while leveraging Scaleway's public cloud infrastructure to optimize resources without compromising on performance.
| Name | CPU | GPU | vRAM | Memory | Storage | Network | Hourly price | Monthly price |
|---|---|---|---|---|---|---|---|---|
| EM-T220E-L40S | EPYC 9334 32c/64t 2.7 GHz | 2x L40S | 48GB | 256 GB DDR5 | 2 x 3.84 TB SSD NVMe | Up to 25 Gbit/s | €4.11/hour | €1499.99/month |
| EM-T620E-H100 | EPYC 9334 32c/64t 2.7 GHz | 2x H100 | 80GB | 512 GB DDR5 | 2 x 7.68 TB SSD NVMe | Up to 25 Gbit/s | €9.589/hour | €3499.99/month |
Elastic Metal’s Titanium resources ensure consistent, stable performance, while its cloud integration offers the flexibility to expand resources as business needs grow.
Compared to variable cloud costs, Elastic Metal Titanium monthly fixed pricing allows companies to budget better for sustained, predictable workloads.
With the performance capabilities of L40S and H100 GPUs, data scientists and engineers can significantly reduce the time it takes to train models and process complex datasets.
Elastic Metal Titanium with 2 x L40S GPU demonstrate superior performance, particularly in rendering and graphics tasks.
Elastic Metal Titanium with 2 x L40S GPU deliver inference performance that enables you to put big models main use cases in production efficiently.
For deep learning tasks, a client using the H100 with its 512 GB of RAM and 80 GB VRAM can reduce the training time of complex models like GPT-based transformers by 30% to 50% compared to previous-gen GPUs.
Based in France with builders and customers all over the world, our services are located within the European Union and protected by its regulations.
100% of electricity consumed in our data centers comes from renewable energy. Decommissioned hardware are securely reused & recycled.
Our technical assistance is available 24/7 to answer all your questions and assist you. Upgrade to support plan to reach directly by phone.
Our services are designed to offer you the best price/performance ratio, with straightforward billing to grow your business and keep your expenses under control.
