Answers are grounded in this paper and cite their sources.
The rapid proliferation of hyperscale data centres (DCs) presents compounded challenges to Great Britain (GB)’s net-zero energy transition, superimposing large, continuously growing electricity loads on an energy system simultaneously aiming to be net-zero by 2050. However, existing assessments of DC sustainability rely largely on facility-level metrics such as Power Usage Effectiveness (PUE) and Water Usage Effectiveness (WUE), which provide limited insight into indirect water use and the cross-sector effects of DC integration on power-system infrastructure and decarbonisation pathways. Here, we develop a spatially explicit whole-system energy–water nexus optimisation model for 13 GB regions, representing air-cooled, water-cooled, and direct-to-chip (D2C) cooling architectures. The model uses 24-hour resolution within weighted representative days and a rolling-horizon solution approach with a one-period look-ahead, with weather-dependent PUE and WUE derived from regional typical-meteorological-year data. Under the modelled scenarios, 21.7 GW of installed DC IT capacity adds 124–133 TWh/yr of electricity demand to the GB energy system by 2050, and water-cooled configurations consume 297 × 10 9 litres/yr of direct, on-site cooling water, compared with 31 × 10 9 litres/yr for air cooling. DC integration also increases indirect water consumption by 229 -- 255 × 10 9 litres/yr through the additional generation required, almost entirely from nuclear generation. Relative to water cooling, D2C cooling reduces total system water consumption by 322 × 10 9 litres/yr, almost entirely through the elimination of direct water use, while differences between the three configurations in electricity generation ( 6 TWh/yr between D2C and water cooling), indirect water and heat supply lie within the numerical variability of the optimisation. Under hypothetical water-availability constraints, the air-cooled configuration remains feasible down to 70% of its unconstrained regional water consumption, at which system cost rises by 4.7% (from £593 to £621 billion), solar capacity increases by 25 GW, nuclear capacity falls by 11 GW, and DC capacity relocates from NT and SW towards SO, WN, NO and SC. The water-cooled configuration remains feasible only down to 95% availability, because its direct cooling water cannot be reduced without changing the installed DC capacity. In an exploratory analysis, integrating DC waste heat with Direct Air Capture (DAC) installs 59 MtCO 2 /yr of fully utilised DAC capacity by 2050, using 87% of the available waste heat and adding 11 TWh/yr of electricity demand. Relative to the D2C, net CO 2 emissions decrease from − 9 to − 68 MtCO 2 /yr, system cost (excluding the reserve-margin penalty) rises by £4–5 billion, while changes in generation and installed capacity remain within the variability between independent solutions, reflecting that DAC deployment is driven by the carbon price applied to net emissions rather than by the net-zero constraint, which is already met without DAC. These results indicate that the system-level consequences of DC deployment can differ from those inferred from facility-level PUE and WUE metrics alone, highlighting the importance of jointly considering electricity, water, cooling, heating, and negative-emissions interactions in planning future DC infrastructure.
No takes yet. Share an insight, caveat, or question.
Hemmati et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: