In his 2023 Spring Budget speech, the Chancellor Jeremy Hunt announced a raft of measures to boost the high-performance computing (HPC), artificial intelligence (AI), and supercomputing sectors, with billions of pounds of investment aimed at making the UK a world-leading, quantum-enabled economy within a decade. Recent research from McKinsey and Company has estimated that emerging technologies such as generative AI could add the equivalent of up-to $4.4 trillion to the global economy, with life sciences among the industries that could see the biggest impact – a catalyst to help support that ambition.
Why does this matter? Firstly, it’s not just because the UK is both an AI and supercomputing pioneer and deserves to reclaim its tech leadership reputation. A second and more important reason is that industrial-scale computing systems are vital to the work of the country’s research bodies, scientists, enterprises, and start-ups, who are working at an accelerated pace to make the world as we know it, a better place.
Funding the UK’s supercomputing revolution
Reflecting on the Government announcements of this year, and the news that the Prime Minister has approved a £100m budget to secure components from leading chipmakers, both the headline figures and the recent action on investments is encouraging.
Initially, they have included £900m to fund an exascale supercomputer facility, £370m for AI research, and a cool £2.5bn devoted to turning the UK into a global leader in quantum and supercomputing, and Jeremy Hunt has confirmed he was accepting “all nine of the digital technology recommendations” made by the former Chief Scientific Adviser, Sir Patrick Vallance in the Digital Technologies Review.
Other pledged projects detail a new quantum computing research centre, and an AI sandbox strategy to help UK companies get their products to market faster. Furthermore, the Department for Science and Technology has committed £3.5 billion to the future of science and tech, with around £1BN of the funding pledged directly to support the next generation of supercomputing and artificial intelligence research – something well placed to drive new economic growth in Greater Manchester’s and underpin its tech, research and AI revolution.
Today the North West, for example, is home to some of the UK’s most advanced computing organisations including The Hartree National Centre for Digital Innovation, the SKA Telescope and the N8 Research Partnership – a collaboration of eight of the North of England’s most research-intensive Universities, all of whom are working together to establish innovative programmes of national and international prominence to help fuel growth. Other organisations harnessing the power of supercomputing for the betterment of humankind include Manchester University, and the Christie NHS Foundation Trust, the largest single site cancer centre in Europe, treating more than 60,000 patients each year.
In Cambridge, a world-renowned location for start-up innovation, Michael Gove is reported to be creating proposals that will turn the city into Britain’s version of Silicon Valley. Here billions of pounds of investment are expected to help build up-to 250,00 homes, further supporting the region’s pharmaceutical, tech and life sciences ecosystems and its desirability as a destination for skilled professionals. The Wellcome Genome Campus has also shared plans to invest millions of pounds in an expansion to secure its status as a world-leading hub for genomics and biodata - adding new capacity for biosciences at a time when the UK is facing a lab space shortage.
At Kao Data our own investments in a new £350M Greater Manchester data centre, and in the continued expansion of our high performance Harlow campus, have positioned us perfectly to underpin the growth and research ambitions of these two prestigious regions. I truly believe that investment strategies such as these are critical to help the UK retain its reputation as a cornerstone of the tech industry, but if the government’s commitments are to be realised, further funding needs to be urgently prioritised.
Supercomputing-powered medical advances
The processing of data both securely, and at speed is one of the most important functions in technology today. By mining vast quantities of data, applications such as AI, machine learning and complex data modelling can fast-track game-changing progress in sectors that range from logistics to climate science.
Data processing at this scale requires the support of intensive computing architectures that are far more powerful than the traditional servers. In short, that’s why we need supercomputers. A key example of this is NVIDIA’s Cambridge-1 supercomputer, which is hosted at Kao Data’s Harlow Campus in the UK Innovation Corridor.
In less than two years, Cambridge-1 has been instrumental in some of the most important and influential medical breakthroughs seen by this country – enabling AstraZeneca, King’s College London, and Oxford Nanopore, among others, to develop new drugs, identify disease-causing variations in the human genome, and gain a better understanding of complex illnesses such as dementia and viruses that include Covid-19.
Organisations such as BioNtech, who recently acquired AI start-up InstaDeep, and which locates its supercomputer at the Kao Data campus in Harlow, has also been appointed to develop ground-breaking new cancer treatments, providing up-to 10,000 patients with precision cancer immunotherapies by 2030. This is another important example of the power of supercomputing for good, and we at Kao Data are delighted to see the continued investment from our Government.
Why supercomputers?
Unlike traditional computers, supercomputers use multiple central processing units (CPUs) grouped into tens of thousands of compute nodes and communicating with systems like data storage and networking, to collaborate on solving specific problems in record time. When deployed within high-performance data centres, supercomputers can achieve these processes incredibly fast and efficiently, even when working at massive scale.
In terms of speed, NVIDIA’s Cambridge-1 can deliver 400 petaflops of AI compute. That’s 400,000 times faster than Intel’s ASCI Red, which became the world’s fastest computer at 1 teraflop (1,000 gigaflops) and is exactly the level of performance that the world’s leading AI scientists are embracing to deliver new breakthroughs. Thanks to supercomputing and its availability to researchers, AI for healthcare is booming in the UK and presents a key opportunity for the country to retain its leadership position.
The future of UK supercomputing: Speed vs sustainability?
The most obvious way in which the UK can compete with foreign supercomputing superpowers is to deliver faster compute speeds with greater efficiency. With enough investment and innovation, teams like those behind Cambridge-1 will be able to achieve exascale speeds in record time.
A dilemma remains, however. Greater computing speeds means greater energy consumption, and the next generation of industrial-scale computing architectures must be built with computational performance, sustainability, and energy efficiency with equal priority. Here, organisations such as NVIDIA are already challenging whether data centres are truly ready to accommodate the next generation of AI, and those who are NVIDIA-DGX Ready, OCP-Readyä, and built to accommodate liquid cooling technologies at-scale, are likely to emerge as frontrunners.
Projects such as NVIDIA’s Earth-2, for example, could theoretically require the same level of power as projects such as Cambridge University's Open Zettascale Lab, whose energy consumption needs are vast. Much of this could be alleviated, at least in some respects, by ensuring that the architectures which support them are precision engineered for sustainability and efficiency, and that both user and operator remain committed to reducing their environmental impact wherever possible – be that through design efficiencies, renewables, or indeed by accommodating direct-to-chip liquid cooling.
Solving the energy consumption requirement of exascale supercomputers is the next big task for the tech industry, and for a country committed to net zero, this is an area which is ripe for investment. By doubling down on government efforts to support HPC, Generative AI and supercomputing research efforts, and ensuring that future high performance platforms are built to the highest standards of sustainability, the UK now has a real opportunity to take a leadership position on the tech world stage.
This article was first published on the Tech Capital in September 2023. Read it here.
Spencer Lamb
Spencer Lamb is Chief Commercial Officer at Kao Data and an experienced figure within the industry. He has a keen interest in HPC, AI and hyperscale computing.