Covering Scientific & Technical AI | Wednesday, January 22, 2025

Nscale Introduces Sustainable GPU Cloud in Collaboration with AMD 

May 2, 2024 -- Nscale, in collaboration with AMD, is launching its high-performance GPU cloud, marking a significant advancement in cloud computing services. Originally spun out from Arkon Energy, a leading 300MW data center and hosting business in North America, Nscale is now setting a new standard for AMD technology-focused cloud solutions.

Located in Northern Norway, the N1 Glomfjord site of Nscale takes advantage of some of the world’s lowest-cost renewable energy sources. This strategic positioning makes Nscale’s N1 Glomfjord site one of the most cost-efficient hubs globally for Large Language Models (LLM) and AI training, ensuring top-tier performance and sustainability.

Key Features of Nscale Cloud:

  • Vertical AI cloud: Nscale owns and operates the full AI stack from its modular data centers to its high-performance compute clusters, allowing it to optimize each layer of the stack for performance and cost efficiency.
  • Built for sustainability: Located in Northern Norway, Nscale is powered by 100% renewable energy and leverages natural cooling solutions to deliver sustainable GPU compute services.
  • Best-in-class economics: Nscale's vertical integration and low-cost renewable power enables it to deliver one of the most affordable GPU computing solutions on the market.
  • Unrivaled user experience: Purpose-built for AI, Nscale streamlines the setup, configuration, and management of cloud-based supercomputing clusters to accelerate AI R&D.

"Over the coming years, AI will proliferate into every enterprise, product, and service, driving unprecedented demand for AI infrastructure. Nscale’s mission is to democratise high-performance computing to unleash the potential of AI. Our 100% renewable AI Cloud Services platform is vertically integrated from the power connection to the cloud, providing sustainable, high-performance, and low-cost compute for Enterprises, Research Institutions and AI Startups alike," said Josh Payne Co-Founder and CEO of Nscale. "By combining cutting-edge technology from AMD with Norway's abundant, low-cost energy, we are offering a next-gen AI Cloud platform that is not only powerful but also sustainable and cost-effective."

“At AMD, we believe in the transformative power of AI and are excited to work with Nscale to deliver cloud instances of the AMD Instinct MI300X accelerator,” said Andrew Dieckmann, corporate vice president and general manager, AMD Instinct Business. “Through Nscale’s AI Cloud Services platform, the MI300X will deliver highly competitive Generative AI inferencing performance, leveraging the AMD ROCm open software ecosystem to enable a compelling TCO advantage to other market alternatives.”

By using the enhanced memory bandwidth and capacity of the AMD Instinct MI300X accelerators, and Nscale’s extensive experience working with hardware and the proven AMD ROCm open ecosystem, Nscale can provide customers impressive price, performance, and efficiency for the most demanding LLM and GenAI workloads.

The Nscale service is poised to attract a wide range of users, from startups to large enterprises, who are looking to leverage advanced AI, deep learning, and big data analytics but have been hindered by the high costs and environmental concerns associated with traditional GPU cloud services.

Nscale is not just a service provider but a partner in innovation. With its robust and affordable GPU solutions, it aims to accelerate the development of ground-breaking technologies and research that can lead to significant advancements in various fields.

For more information about Nscale Cloud and to get started, visit Nscale.com.

About Nscale

Nscale's GPU cloud platform is engineered for the demands of AI, offering high-performance compute optimized for training, fine-tuning, and intensive workloads. From our data centres to software stack, we are vertically integrated in Europe to provide unparalleled performance, efficiency and sustainability.


Source: Nscale

AIwire