IONOS has introduced the IP4-50 GPU, a new dedicated GPU server tailored for artificial intelligence and powered by the Blackwell architecture on its NVIDIA RTX PRO 6000 with 96 GB of VRAM card. The GPU server sets new benchmarks for enterprises demanding high-performance computing (HPC) for AI applications, 3D rendering, or complex scientific simulations, while also prioritizing GDPR compliance and sovereign hosting in European data centers.
High-Performance Computing: A Pillar of the Momentum Ecosystem
The launch of the IP4-50 GPU server forms part of the Momentum initiative, IONOS’ new AI ecosystem, aimed at democratizing access to the most advanced AI technology. Thanks to its pure physical “bare metal” architecture, users reap the full hardware performance without virtualization overhead, a crucial factor for demanding AI workloads such as training and deploying large language models (LLMs).
The dedicated GPU server impresses with its technical specifications:
- GPU: NVIDIA RTX PRO 6000 (Blackwell Server Edition) with 96 GB of VRAM, optimized for large data volumes and AI models.
- CPU: Dual Intel Xeon 6517P processors with 32 cores and clock speeds from 3.2 GHz to 4.2 GHz.
- RAM: 256 GB DDR5 ECC RAM.
- Storage: 2 x 1,920 GB NVMe SSDs in HW RAID 1 configuration.
The pricing model is flexible, giving customers full control over costs in a pay-as-you-go format: the rate is billed at €1.6528 per hour (roughly $1.78 per hour at current exchange rates). There is also a flat monthly maximum rate of €1,190 (about $1,285), which is especially advantageous for long-running AI training projects.
The IP4-50 GPU server is hosted in the company’s secure and sovereign European data centers, ensuring that customers’ sensitive data remains under strict European jurisdiction. In addition, IONOS fully complies with GDPR.
