Digital Catapult urges UK datacentre planning overhaul to fix 50GW queue

The shift toward treating compute as a flexible asset directly impacts high-bandwidth services like video streaming and AI-driven media processing. If the UK adopts carbon-aware scheduling, streaming platforms may need to architect workloads to be time-shiftable to avoid peak energy costs and grid constraints. This integrated approach mirrors challenges seen in Ireland, where datacentres now consume 21% of national electricity, suggesting that infrastructure availability will soon dictate regional deployment strategies for global platforms. As the UK targets 6GW of AI-ready capacity, the industry should watch for new government frameworks that prioritize grid-flexible datacentre projects over traditional hyperscale developments.

Digital Catapult's push to integrate datacentre planning with energy infrastructure arrives amid a broader scramble for grid capacity across the UK and Europe. In March 2025, the UK government published its Data Centre Strategy, committing to a 6GW target for AI-ready capacity by 2030, a figure that would require significant new grid connections and has drawn scrutiny from National Grid ESO, which manages the connection queue. The strategy also designated datacentres as Critical National Infrastructure, elevating their priority in planning decisions. Google, one of the entities mentioned in the source report, has been among the most active hyperscalers in the UK market, announcing a £5 billion investment in UK datacentre expansion in January 2025 to support AI workloads, a move that directly increases demand on the constrained grid. The regulatory and policy landscape around datacentre energy use is tightening across multiple jurisdictions. Ireland's Commission for Regulation of Utilities introduced a moratorium on new datacentre connections in Dublin in 2022, which has since been extended and now serves as a cautionary precedent for the UK. In the UK itself, Ofgem and National Grid ESO have been exploring flexible connection agreements that would allow datacentres to curtail consumption during peak demand periods. The UK's National Grid reported in late 2024 that the datacentre connection queue had grown to over 50GW, far exceeding available capacity. Digital Catapult's recommendation to treat compute workloads as flexible assets aligns with these emerging regulatory mechanisms and could influence how streaming and media processing workloads are classified under future grid-access rules. Technical benchmarks for flexible datacentre operation remain limited but are growing. A 2025 study by the International Energy Agency estimated that global datacentre electricity demand could reach 945 TWh by 2030, roughly double current levels, with AI workloads driving much of the growth. In the UK specifically, the Digital Catapult's own earlier research on carbon-aware computing demonstrated that shifting batch workloads to periods of high renewable generation could reduce carbon intensity by up to 40%. For streaming platforms, this suggests that transcoding, encoding, and content delivery caching could become candidates for time-shifting under new grid-flexibility frameworks. Luke Sperrin, who leads Digital Catapult's work on AI infrastructure, has publicly argued that are particularly well-suited to flexible scheduling because they are often latency-tolerant compared to real-time inference tasks.