UK decarbonisation progress shows mixed results as AI deployment accelerates in operational efficiency but stalls in material redesign

PROGRESS in applying artificial intelligence to decarbonise manufacturing has been uneven, with meaningful wins in operations but slower movement on fundamental material and product redesign, according to a report by Energy Systems Catapult commissioned by the UK Department for Energy Security and Net Zero.

The Catapult’s analysis found that amid a “torrent of news and hype” around AI, deployment has advanced fastest where clear commercial incentives exist , chiefly process optimisation and energy efficiency , while applications that would cut emissions by changing manufacturing inputs remain largely at the research or early-trial stage. The report notes AI has been used in exploratory work to predict material properties for lower-clinker cement and to optimise alloy use in lower-carbon steel, but “real-world adoption has been relatively limited” and the UK is “not particularly a leader” in developing or deploying these applications. The report singled out the Henry Royce Institute’s Digital Materials Foundry as a potential accelerant, describing open-source experimental materials datasets as the kind of resource that could support AI-driven property-prediction models and speed materials innovation.

There are, however, concrete examples of operational gains. According to the Catapult, a trial at Heidelberg Materials’ Mokra cement works in Czechia used an AI tool developed by Carbon Re to predict clinker quality in real time, enabling operators to adjust kiln conditions and fuel use and delivering a 2.2% reduction in specific heat consumption and a 2% reduction in overall emissions. The report also highlighted Deep.Meta’s machine-learning tool, which cut emissions in a steel rolling mill by 5% by predicting when steel had reached sufficient temperature and thereby avoiding unnecessary reheating.

The fastest progress the Catapult observed has been outside heavy industry. AI-managed tariffs for electric vehicle charging and AI-enabled home decarbonisation tools have shown rapid uptake and large demand-side impacts; the report cites a study finding AI-managed tariffs reduced EV electricity demand by 42% and notes technologies that have reduced the cost of heat pump installation.

The lead author, Sam Young, urged caution about indiscriminate deployment. “Indiscriminate use of AI can increase emissions,” he said. “But smart, targeted uses are already breaking down some of the hardest barriers to a low-carbon economy.”

The mixed picture the Catapult describes sits alongside stronger signals from other recent industry and government activity. A Rockwell Automation report cited by IT Pro found that 53% of UK manufacturers are implementing AI on the factory floor , above the European average , driven by use cases such as computer vision for quality control and predictive maintenance. The report attributed this to a robust UK ecosystem of startups, institutional funding and workforce upskilling, while also warning of difficulties in scaling AI systems across enterprises and workforce concerns about displacement.

Public investment is following that adoption. The UK government has announced several funding tranches intended to accelerate AI for decarbonisation: nearly £4 million to support a dozen green AI initiatives, additional awards totalling about £1.73 million for eight AI projects targeting emissions reductions and energy efficiency, and other strand funding and programmes described as part of the wider £1 billion Net Zero Innovation Portfolio. The government has also launched smaller targeted programmes and a £1.5 million AI for Decarbonisation initiative intended to stimulate innovation in the area.

To coordinate effort and shorten the path from prototype to industry use, new national partnerships have been established. According to The Alan Turing Institute and partners, the Artificial Intelligence for Decarbonisation’s Virtual Centre of Excellence (ADViCE) , run by Digital Catapult, Energy Systems Catapult and The Alan Turing Institute , aims to connect AI innovators with high-emitting sectors and to identify adoption barriers. The Industrial Decarbonisation AI Coalition, launched in December 2024, was billed as a platform to pool expertise and resources to accelerate industrial AI decarbonisation.

Taken together, these public and private moves reflect a two-track reality: operational efficiency and demand-side management are delivering measurable emissions cuts today, while the more transformative opportunities , swapping feedstocks, redesigning products and reconfiguring processes , require stronger data infrastructure, clearer commercial cases and sustained co-ordination between materials scientists, manufacturers and AI developers.

Industry data and project results underscore the commercial rationale for further deployment. High energy prices and the immediate cost savings from fewer defects, lower reheat energy and improved kiln performance are driving trials and early roll-outs. But the Catapult warns that scaling beyond pilot projects will need better access to high-quality experimental datasets, investment in digital skills, and mechanisms to de-risk integration into complex production environments.

The picture for UK leadership is similarly mixed. Government funding and new collaborative centres seek to position the UK at the forefront of AI-enabled decarbonisation, while industry surveys report above-average factory-floor AI adoption. Yet the Catapult cautions that for material- and design-centred decarbonisation the UK is not yet a clear leader and that wider adoption internationally will depend on the availability of open data, demonstrator projects and industry-specific pathways to market.

For policymakers and industrial decarbonisation professionals, the report underscores a pragmatic sequence: scale AI where it lowers energy use and operating cost now, while investing in the data, standards and testbeds needed to bring AI into materials design and full-process reengineering. As Sam Young put it, the task is to avoid “indiscriminate” use and focus on “smart, targeted uses” that can be verified to reduce emissions and deliver commercial value.