Saturday, January 10, 2026

AI26004 Isambard AI - Bristol University V01 100126

 Launch this You Tube Video to learn about the most powerful AI Super Computer in the United Kingdim (2025) called the Isambard - AI at the Bristol Centre for Supercomputing based at Bristol University, Bristol, United Kingdom. 


https://youtu.be/COtQoHUgHZk?si=qzx22fuO1NipERDZ

Isambard-AI is the most powerful supercomputer in the UK, purpose-built to accelerate artificial intelligence research and scientific discovery. Launched in July 2025 at the University of Bristol, it serves as a central pillar of the UK’s AI Research Resource (AIRR).

System Overview

Performance: Capable of over 200 quadrillion calculations per second (200 Petaflops).

Hardware: Powered by 5,448 NVIDIA GH200 Grace Hopper Superchips, which combine a CPU and GPU into a single high-performance unit.

Networking: Utilizes HPE Slingshot 11 interconnects to allow for massive data transfer speeds between chips.

Sustainability: It is the 4th greenest supercomputer in the world as of early 2026, using advanced liquid cooling that reduces carbon emissions by roughly 72%.

Storage: Includes a massive 25-petabyte high-performance storage system to handle the giant datasets required for training Large Language Models (LLMs).

Key Research Projects

The system is currently hosting several "mission-critical" projects for the UK government and scientific community:

NHS Cancer Screening: Developing scalable AI models to automate prostate cancer screening and reduce diagnostic backlogs.

Climate Modeling: Simulating complex weather patterns and decarbonization strategies to help the UK reach Net Zero.

Drug Discovery: Accelerating "in silico" research to identify new medical treatments much faster than traditional lab methods.

AI Safety: Providing the AI Safety Institute with the compute power needed to stress-test frontier AI models for potential risks.

How to Access Isambard-AI

Access is managed via the AIRR Portal and is divided into specific "routes" based on the user's needs:

Gateway Route: A rolling, flexible pathway for first-time users or those exploring novel AI algorithms (up to 10,000 GPU hours).

Rapid Access Route: Specifically for UK-registered SMEs and startups to quickly prototype AI products (up to 20,000 GPU hours).

Innovator Route: For larger, multi-sector partnerships and industrial research (50,000 to 150,000 GPU hours).

AI for Science Route: A high-level tier for massive projects in fields like nuclear fusion, materials science, and engineering biology (up to 1 million GPU hours).

No comments:

Post a Comment

AI26019 Copyright and AI V01 100326

  Creative types have the upper hand in AI copyright fight Katie Prescott Kanishka Narayan is the minister for AI Next image  › ‘‘ Pimli-cod...