OVHcloud unwraps NVIDIA GPUs & AI models for diverse use

OVHcloud, the European provider of cloud services, has unveiled its latest portfolio of NVIDIA GPUs and open-source Large Language Models (LLMs) such as Mixtral 8x22B and Llama 3. The announcement at the Adopt AI event in France renews OVHcloud's commitment to making artificial intelligence accessible for organisations, aligning with its values of openness, transparency, and building a trusted cloud environment.

Yaniv Fdida, Chief Product and Technology Officer at OVHcloud, expressed enthusiasm for the broader accessibility of AI technologies. "We are incredibly excited to further democratise AI for our Public Cloud and Bare Metal customers by offering a broad choice of NVIDIA GPUs, strategically selected and priced for a variety of AI workloads ranging from model training to inferencing," he said.

The company aims to address businesses' evolving AI requirements through a diverse array of solutions available at different price points. The newest selections of NVIDIA GPUs, which include the NVIDIA H100 Tensor Core GPU, NVIDIA L4 Tensor Core GPU, and NVIDIA L40S GPU, are now available in OVHcloud's Public Cloud. The GPUs are designed to accelerate AI projects from model creation to inferencing, housed in OVHcloud's eco-friendly data centres.

OVHcloud's introduction of the NVIDIA L40S GPU expands its Public Cloud offerings beyond the previously available NVIDIA H100 and L4 GPUs. The L40S combines powerful AI computing capabilities with superior graphics and media acceleration, supporting workloads from LLM inference and training to 3D graphics and video rendering.

In addition to the Public Cloud options, OVHcloud has launched new Bare Metal servers equipped with NVIDIA L40S GPUs, specifically tailored for deep learning, machine learning, and high-performance computing. The new HGR-AI-2 servers feature AMD EPYC 9354 processors in a dual configuration, with 64 cores, 128 threads, and up to 2.3 TB of DDR5 memory, providing substantial computational power. These servers offer configuration options for up to four NVIDIA L40S 48 GB GPUs per server.

OVHcloud's AI solutions extend across the full AI value chain, providing tools from data experimentation to model deployment. Their serverless tools, including AI Notebooks, AI Training, and AI Deploy services, enable companies to experiment with data, train models, and bring AI projects to production without the complexity of managing infrastructure. The services support minute-by-minute billing, rendering them accessible and economically efficient for businesses of all sizes.

The latest suite of open-source models, comprising Mixtral 8x22b and Llama 3, is also available via OVHcloud's AI Endpoints. This serverless solution allows developers to access various AI models through API endpoints, offering a testing playground before actual API calls are made. This ensures ease of use and data privacy adherence, bolstering OVHcloud's reputation for secure data handling.

To complement these advancements, OVHcloud's Public Cloud universe now includes the newly integrated NVIDIA GPU options, ready to drive AI-powered initiatives. Additionally, the new HGR-AI-2 servers with NVIDIA L40S GPUs are available in France and Canada, with prices starting at EUR €2,969.99 per month. Australian customers can also order these servers from data centres located in their region.