Stellium stated that its Newcastle facility has shifted from annual renewable matching to an hourly system, aligning energy use with local renewable generation. The company reports this change has reduced operational carbon intensity, cutting emissions by an estimated 75%. (Good Energy notes that many providers claim ‘100% renewable’ status based on annual averages, which can mask fossil fuel use during periods of low renewable output. The new model addresses this by continuously tracking energy use.)

According to data provided by the partners, the Stellium facility achieved an hourly matching score of 95.4%. This figure is significantly higher than the current UK market average, which the report estimates at approximately 43%. Achieving this involved sourcing power from over 3,300 independent UK renewable generators.

Paul Mellon, Operations Director at Stellium, acknowledged criticism of data centres for inflexible energy use. He stated that hourly matching provides transparency and demonstrates that large-scale digital operations can align with decarbonisation goals.

The partners state that directly linking demand to local renewable energy output helps the facility avoid ‘greenwashing’ concerns associated with non-time-matched Renewable Energy Guarantees of Origin (REGOs). This approach is said to encourage the expansion of local renewable assets. By creating a direct market for hourly renewable power, the model provides more predictable revenue for independent wind and solar farms across the UK.

According to a recent International Energy Agency (IEA) report, global data centre electricity consumption rose by 17% in 2025, alongside a $400bn investment from the five largest technology firms to expand digital infrastructure. The IEA notes that AI-related power use is increasing faster than general data processing. The agency projects total data centre demand will double by 2030, while AI-specific energy consumption will triple.

The IEA adds that a growing number of global projects is putting pressure on national planning systems, leading to delays in grid connections and prompting technology providers to seek reliable power solutions. And plans are being held back by tightening supply chains for key components like transformers and gas turbines.

The IEA report identifies AI as both an ‘energy taker’ and a potential ‘energy maker.’ While AI consumes significant power, it also drives innovation in long-duration energy storage. The agency states that strong demand from the tech sector is accelerating the commercialisation of next-generation energy technologies, including increased interest in carbon-free sources that provide constant baseload power to data centres.

The IEA says that the ‘scramble for solutions’ will likely persist as AI models grow more complex and energy-intensive.

Reports from Good Energy and Stellium indicate that technology already exists to decouple data growth from reliance on fossil fuels. They suggest the industry must now scale these solutions to meet future demand without compromising environmental commitments.

In related news, the Greening AI Data Centres Coalition (GADCC), a new initiative involving nine global bodies, including the World Green Building Council and the Climate Bonds Initiative, was recently launched to establish clear, credible standards for sustainable development. By defining what green means – focusing on clean energy, water recycling and heat reuse – the coalition said it aims to help investors and operators cut through greenwashing and direct capital toward facilities that meet legitimate environmental criteria.

The importance of these standards is shown by the rapid scale of investment. Sean Kidney, CEO of the Climate Bonds Initiative, warned that without such guidance, the trillions currently being poured into AI infrastructure risk becoming a ‘climate disaster’. Beyond energy, the GADCC is focused on balancing this growth with ‘responsible development’ that protects local resources and energy affordability. This framework is designed to complement local innovations, like the Newcastle hourly-matching model, by providing a consistent, data-backed benchmark that is said to protect communities and energy security worldwide.