The UK government has tripled its funding for supercomputers to equip researchers with the necessary tools to evaluate the safety of advanced AI models. Originally budgeted at £100 million during the Spring Budget in March, the AI Research Resource, designed for AI safety assessment, will now receive an impressive £300 million.
This announcement coincided with the UK hosting a summit on AI safety at the historic Bletchley Park, where 28 nations came together to establish a shared agreement on the risks associated with AI.
Michelle Donelan, the Secretary of State for Science and Technology, expressed the urgency of adopting AI technology safely. She emphasized that the UK is poised to lead the world in this endeavor, providing researchers and scientific talent with the tools required to understand and harness this complex technology.
The University of Cambridge will be home to the second supercomputer, aptly named ‘Dawn.’ Powered by 1,000 water-cooled chips, Dawn will focus on climate modeling, healthcare, and fusion energy. It is expected to be operational within the next two months. Dawn complements the Isambard-AI supercomputer, which will be located at the University of Bristol.
Costing £225 million, this supercomputer, set to be ten times faster than the UK’s current fastest machine, will be supplied by technology company Hewlett Packard Enterprise (HPE) and feature over 5,000 Nvidia chips.
Ian Buck, Vice President of Hyperscale and HPC at Nvidia, emphasized the importance of nations creating their own AI infrastructure, enabling researchers to access cutting-edge AI and HPC compute resources used by global AI pioneers. The Isambard-AI supercomputer is poised to drive AI and scientific breakthroughs.
However, concerns have emerged regarding the availability of electricity for the grid to power the supercomputer’s GPUs, as reported by The Telegraph. This ambitious project signifies the UK’s commitment to advancing AI technology safely and driving scientific progress.