Next-generation AI infrastructure: IONOS unveils new high-performance GPU servers with NVIDIA Blackwell technology

London, 8 May 2026

IONOS, Europe’s leading digitalisation partner for small and medium-sized businesses, has launched  IP4-50 GPU, a new dedicated GPU server. It is AI-ready and equipped with the Blackwell architecture NVIDIA RTX PRO 6000 with 96 GB VRAM. The GPU server is designed for companies that require high-performance computing (HPC) for applications such as AI models, 3D rendering or complex scientific simulations and value GDPR compliance as well as fully sovereign hosting in European data centres.

High-performance computing: a building block of the Momentum ecosystem

The launch of the IP4-50 GPU server is part of the Momentum initiative, IONOS’ new AI ecosystem, which aims to democratise access to state-of-the-art AI technology. Thanks to the purely physical ‘bare metal’ architecture, users benefit from full hardware performance without virtualisation overhead – crucial for demanding AI applications such as the training and deployment of large language models (LLMs).

Technical specifications and flexible pricing

The Dedicated GPU Server impresses with its technical features:

• GPU: NVIDIA RTX PRO 6000 (Blackwell Server Edition) with 96 GB VRAM – optimised for larger data sets and AI models.
• CPU: Dual Intel Xeon 6517P with 32 cores and clock speeds from 3.2GHz to 4.2 GHz.
• RAM: 256 GB DDR5 ECC RAM
• Storage: 2 x 1,920 GB NVMe SSD in HW RAID 1

The flexible pay-as-you-go pricing model gives customers full cost control: the rate is billed at £1.8056 per hour. There is also a maximum monthly flat rate of £1,300 which is particularly beneficial for long-term AI training projects.

The IP4-50 GPU servers are securely hosted in ISO 27001- certified IONOS data centres, ensuring that sensitive customer data remains protected under top-notch security. 

Further information on the IP4-50 GPU server can be found here: GPU Server Rental | NVIDIA® Graphics Card | IONOS