The UK government outlines policy approaches to advance high-performance computing and AI infrastructure sustainably.
In May 2024, the UK entered the European High Performance Computing Joint Undertaking (EuroHPC) program, bringing together supercomputing resources from 35 countries, including Norway, Turkey and all 27 European Union member states. By joining this prestigious program, the UK aims to bolster its scientific and technological leadership, foster international cooperation, and leverage HPC to drive innovation and economic growth. However, membership of the EuroHPC will accelerate demand for energy in the UK, potentially exceeding the nation's current capabilities. This energy challenge is compounded by the UK's need to contend with the formidable power demand surge created by increased use of Artificial Intelligence. John Pettigrew, Chief Executive Officer of the UK's National Grid, commented that AI will consume 500% more power in the UK during the next decade.
The scale of this global energy challenge is significant. Data centers worldwide consumed 460 terawatt-hours in 2022, accounting for almost two percentage points of total global electricity demand. With the added power demand for AI, electricity consumption from data centers in the EU in 2026 is predicted to be 30 percent higher than 2023. According to recent research by Goldman Sachs, AI is set to increase data center power demand by 160% by the end of the decade. Google's emissions have nearly doubled in the past five years alone thanks to AI, raising critical questions about the possibility of achieving net-zero emissions by the end of the decade.
The key to sustainably meeting this demand lies in harnessing renewable energy to power data centers, supported by energy-efficient strategies at AI facilities. While renewables like solar and wind are inherently intermittent and present challenges around consistent power supply and high initial investment costs, solutions exist. NexGen Cloud hosts its AI Supercloud in a data center just outside Oslo, which uses hydroelectric energy to power and cool its infrastructure, enabling round-the-clock clean energy supply. Additionally, direct government investment in GPU chips is vital, as the UK's investment in Nvidia GPU chips is dwarfed by orders from tech giants like Elon Musk's X and China. According to Pitchbook, $122 billion has been invested in generative AI businesses in the US, while European generative AI businesses have received only $3.8 billion.
Adopting sovereign cloud solutions can help UK businesses access high-powered GPUs while ensuring compliance with data protection laws and European data sovereignty regulations. By powering data centers with 100% renewable energy and implementing efficient infrastructure investment, the UK and EU can significantly mitigate the environmental impact associated with High Performance Computing and Artificial Intelligence, thus enabling innovation to continue without jeopardizing the possibility of net-zero by 2030. This approach can provide both a blueprint for how the UK can sustainably support HPC and AI, and a path to becoming a true global leader in these sectors.