AI-driven surge in data centre energy demand sparks scepticism over sustainability

Global data centre electricity consumption is projected to nearly double by 2030, driven by AI expansion, raising urgent questions about sustainable power solutions amid mounting environmental concerns.

Electricity demand for data centres worldwide is set to surge dramatically over the coming decade, driven in large part by the rapid expansion of artificial intelligence (AI) technologies. According to Gartner, Inc., global data centre electricity consumption is forecast to rise 16% in 2025 alone, reaching 448 terawatt hours (TWh), and then nearly double to 980 TWh by 2030. This sharp increase is predominantly due to the growing use of AI-optimised servers, which are expected to see their power consumption surge from 93 TWh in 2025 to 432 TWh by 2030. These AI-specific servers will account for 21% of global data centre electricity use in 2025 and expand to 44% by 2030, comprising 64% of the incremental power demand for data centres at that point.

The US and China stand as the leading hubs of this rapidly growing infrastructure, together accounting for more than two-thirds of global data centre power demand. China is better positioned to manage electricity consumption through the deployment of more power-efficient servers and careful infrastructure planning, while the US data centre consumption is projected to rise from 4% to nearly 8% of the country’s total regional electricity use by 2030. Europe is also experiencing growth, with its share of regional power usage by data centres expected to climb from 2.7% to 5% in the same period.

Other major research organisations corroborate these findings, though with some variations in their projections. Goldman Sachs, for example, anticipates a 165% rise in global data centre power demand by 2030 relative to 2023 levels, a figure notably higher than Gartner’s forecast, underscoring the intense energy appetite of AI developments. Similarly, the International Energy Agency (IEA) predicts that data centre electricity usage could more than double by 2030, reaching approximately 945 TWh, with the US and China responsible for nearly 80% of this growth. Additionally, McKinsey reports that power demand for US data centres alone is expected to soar to 606 TWh by 2030, representing nearly 12% of the nation’s total electricity consumption.

This unprecedented growth in power demand raises significant concerns regarding sustainability. Currently, fossil fuels dominate on-site power generation for data centres, a situation deemed unsustainable in the long term. Industry experts highlight emerging clean energy alternatives, such as green hydrogen, geothermal energy, and small modular nuclear reactors (SMRs), as promising options for powering future data centre microgrids. However, these technologies face barriers including high initial costs and regulatory challenges. In the nearer term, natural gas is expected to remain the primary energy source, while rapid adoption of battery energy storage systems (BESS) is forecast within three to five years to help balance intermittent renewable sources like solar and wind.

The predicted increase in data centre consumption emphasizes how AI’s growing computational demands are reshaping global electricity landscapes. These trends underline the urgent need for energy-efficient AI hardware innovation and scalable renewable energy integration to prevent environmental and infrastructural strain as digital economies expand. While some projections differ in scale, the consensus is clear: data centres, especially those supporting AI, will be major drivers of electricity demand in the coming decade, prompting critical discussions about sustainable power solutions.

📌 Reference Map:

• ^[1] Communicationstoday – Paragraphs 1, 2, 3, 4, 5
• ^[2] Gartner – Paragraphs 1, 3
• ^[3] Goldman Sachs – Paragraph 2
• ^[5] McKinsey – Paragraph 3
• ^[7] IEA/Datacenterdynamics – Paragraph 3
• ^[1] Communicationstoday (Tony Harvey quote) – Paragraph 4

Source: Fuse Wire