Round up | Computing for Sustainable Innovation: 3rd Exascale and Scalable AI Workshop

For the third year in a row, our Exascale and Scalable AI Workshop brought together representatives from industry, academia and the scientific community to discuss how to help the UK move forward in the exascale landscape.

Jointly organised with the Lawrence Livermore National Laboratory (LLNL) and Argonne National Laboratory (ANL) from the United States, the event marked a continued commitment to fostering international collaboration at the forefront of high performance computing (HPC) and artificial intelligence (AI). The aim was to explore how exascale computing and scalable AI can address global sustainability challenges and unlock new scientific and economic opportunities.

“The exascale and scalable AI workshop enables us to meet with our U.S. counterparts, share what we’re working on and explore ways to collaborate in solving the grand challenges we face. A major focus was on sustainable innovation like the fusion energy sector, which holds great promise for producing renewable energy and tackling climate change.”

Sue Thorne, STFC Hartree Centre

The workshop started by highlighting the strength of collaboration to advance computing at scale. With active involvement from the Science and Technology Facilities Council (STFC), ANL, LLNL and other UK research organisations, the event showed how shared ambitions are shaping national strategies.

Talks by DiRAC and Network Plus highlighted how co-design of HPC and AI services maximises return on investment. Speakers also stressed the growing importance of training programmes and the need to decentralise compute infrastructure to improve accessibility and support knowledge exchange.

“Federation will be important for how we leverage our existing and future investments to be sustainable and efficient and accessible for our large-scale computing infrastructure overall. So our Network Plus project will produce a roadmap for UKRI on how to achieve that over the next year and a half. That hopefully will allow us in 2026 to be able to get on with delivering a federated infrastructure for the UK.”

Jonathan Hayes, Queen Mary University of London

Speakers from UK Atomic Energy Authority (UKAEA), Lucideon and the NHS provided examples of how advanced computing is delivering real-world impact. Whether accelerating material innovation, supporting public health through AI-driven diagnostics or enabling progress in fusion energy, scalable compute and AI are essential for competitiveness and resilience.

“The NHS is quite good at looking inwards. What we need to do is look outwards and recognise that the solutions to many of our problems have probably already been developed somewhere else. Being in the room with people from different industries and sectors allows you to strip away your assumptions based on your own knowledge of a particular industry and be open to a new way of thinking.”

Janet King, NHS England

IBM added further insight by addressing the sustainability challenges facing future computing systems. As demand for computational power grows, so too does the responsibility to ensure that innovation aligns with environmental and energy goals.

Sustainability was not just a topic of discussion, but the central theme of the entire workshop. As computational capabilities increase, so must our efforts to reduce environmental impact. Discussions covered system design, energy efficiency and the development of data-driven solutions to address sustainability challenges.

A dedicated breakout session on sustainability and computing brought together perspectives from national laboratories, academia and industry. Discussions included the use of AI to automate aspects of research software engineering, emerging techniques for reducing HPC energy consumption, and the potential of custom hardware such as LBANN (Livermore Big Artificial Neural Network) to improve efficiency. The session underscored the importance and urgency of integrating sustainability principles into computing research and infrastructure.

With the increasingly heterogeneous systems emerging to satisfy the demand for various AI and HPC workloads and the combination of those, we will need a sustainable approach to enable this heterogeneity to be exploited efficiently. One such approach can be the composable architectures one based on the current collaboration between Hartree Centre and IBM and presented by Jim Sexton (IBM) on the 3rd Exascale and Scalable AI Workshop.

Vassil Alexandrov, STFC Hartree Centre

One stream of breakout sessions explored how exascale computing is driving progress in life sciences and materials research. Presentations from leading UK facilities demonstrated how national infrastructure, powered by HPC, is enabling faster drug discovery, advanced materials design and more efficient scientific experimentation.

While the other stream focused on specialist areas including battery technologies, electric vehicles and alternative energy. These discussions illustrated the wide-ranging impact that scalable AI and simulation can have in supporting sustainable innovation, such as improving battery safety through AI-powered battery management systems.

There is a continued growing need for more sustainable and efficient EV batteries to meet net-zero goals. Exascale compute and scalable AI are crucial to support these goals, workshops like these bring together experts from across the board to work together towards more sustainable innovation.

Tim Powell, STFC Hartree Centre

The workshop concluded with a forward-looking panel discussion on how exascale computing and scalable AI can help future-proof businesses, research and public services. The key message was that while computing is becoming more powerful, its potential to serve society depends on responsible development and use.

As exascale systems come online and AI continues to evolve, the future of computing will require deeper integration with societal goals. Events like this workshop show that with strong partnerships and a focus on sustainable outcomes, we are well-positioned to lead the way.