AI and high-performance compute can transform engineering and manufacturing processes, with real impact. But we’re a long way from plug-and-play solutions.
When Isambard-AI was first powered up last year, it marked a step-change in UK AI capability. At launch in July 2025, it became the UK’s most powerful AI supercomputer.
Located at NCC, it's built and managed by the Bristol Centre for Supercomputing (BriCS), part of the University of Bristol. After 12 months in service it’s accelerating AI-led innovation in sectors from climate science to healthcare.
Through NCC’s access to this national research asset, the advanced engineering and manufacturing sectors also stand to benefit. The challenge is making it happen.
Engineering processes are complex, data held by most firms is unstructured and unreadable by AI, and we need safe digital and physical spaces to test new solutions. Success requires sustained coordination, between industrial partners and organisations like the High Value Manufacturing Catapult, to lay the groundwork for successful AI adoption and prove its worth.
When we get it right, the results are compelling: AI for adaptive manufacturing can reduce part defects by up to 80%, other models used for automated inspection can realise human-level accuracy in a fraction of the time.
Here’s what we’ve learnt about applying AI to engineering – and where we’re headed next.
Where and when to use AI
While businesses can roll out administrative AI at pace, deploying it on the factory floor is an entirely different challenge. This means we need to be savvy about identifying where AI can really make a difference – and deliver a return on the resources invested in developing and scaling it.
Our digital engineers have been exploring the applications of Isambard-AI which show genuine promise. A key part of the puzzle is ensuring we can give AI the data it needs, but also recognising where AI is not the most appropriate response to an engineering challenge.
Read more: AI in engineering: Turning promise intro practise
How to build AI models which actually work
For advanced engineering and manufacturing, off-the-shelf AI models won’t cut it. We need to build our own.
Isambard-AI gives us the technical capability, but success depends on harnessing the expertise of both data scientists and domain experts (engineers) who understand complex processes inside out.
Over the last 12 months, NCC has been determining what building AI models actually involves, and which teams you need around the table to do it.
Read more: The AI models that count will be shaped by engineers
What AI can deliver for engineering and manufacturing
It’s easy to be sceptical about AI: discussion about its potential is everywhere, examples of it working in industrial settings is harder to find.
However, with the right data, the right partners, and pointed at the right problem, we’re starting to see how it can add real value.
In the Horizon Europe TURBO project, AI-controlled infusion is projected to reduce wind turbine blade defects by up to 80%. It can even be done remotely, NCC machine learning models in the UK have been used to control large-scale infusions for Siemens Gamesa in Denmark.
We’ve also used AI to identify defects in aerospace wing preforms, flagging issues before a costly infusion process takes place.
The potential impact shouldn’t be underestimated. This is AI significantly improving the processes on which key growth sectors like clean energy and aerospace depend, in doing so tackling some of their most critical manufacturing challenges.
The work continues
Over the next twelve months, NCC will continue to work with industrial partners and the UK's AI Champion for Advanced Manufacturing, Chris Dungey of HVM Catapult, to maintain momentum and deliver the Advanced Manufacturing AI Adoption Plan. Our priorities include:
- Industrialising AI tools which are proven to work in demonstrator environments.
- Identifying new applications with a focus on engineering design, certification, inspection, and adaptive manufacturing.
- Building an industry consortium devoted to developing new standards and certification processes in product assurance.
- Supporting SME digital adoption through the Made Smarter programme, helping businesses identify where AI can improve process, reduce waste, and quicken time-to-market.
- Supporting our regional partners to position the West of England as the UK’s first AI Supercluster with a comprehensive research push and industry pull offer.
With thanks to the University of Bristol and Bristol Centre for Supercomputing for facilitating access to Isambard-AI, helping advanced engineering and manufacturing realise the benefits of AI.