RESEARCH
Global Data Center Electricity Demand for 2030 Splits 945 to 1,587 Terawatt-Hours Across Energy Analysts
Four major forecasters -- the International Energy Agency, Gartner, Deloitte, and S&P Global -- each published 2030 projections for global data center electricity consumption, and their headline TWh figures diverge by more than 40%: from IEA's 945 TWh to S&P Global's 1,587 TWh. All four agree data center power draw will roughly double or more from current levels and that AI-optimized servers are the primary driver, but they disagree sharply on the absolute scale utilities and grid operators need to plan for. For power-sector planners and AI infrastructure investors, that gap is the difference between adding one large economy's worth of new demand versus adding closer to two.
Findings
- IEA projects global data center electricity consumption will reach 945 TWh by 2030, more than doubling from 415 TWh in 2024 -- a figure IEA says grows electricity demand from data centers roughly four times faster than total electricity demand from all other sectors combined.
- Gartner forecasts global data center electricity demand of 980 TWh by 2030, up from 448 TWh in 2025, with AI-optimized servers alone rising from 93 TWh to 432 TWh over that span -- nearly fivefold growth in the AI-specific slice.
- Deloitte's 2030 estimate of 1,065 TWh, up from roughly 536 TWh in 2025, sits 12.7% above IEA's figure and 8.7% above Gartner's, despite all three citing the same generative-AI buildout as the primary driver.
- S&P Global's 1,587 TWh 2030 estimate is the highest of the four -- 68% above IEA's figure -- built from a 2025 baseline of 860 TWh that is itself roughly double IEA's implied 2025 baseline, suggesting the firms are not starting from a shared measurement of current data center load.
Delta Engine result
↔ Divergence Detected — Δ 0.1965 (threshold 0.05)
- IEA T5 95
- S&P Global Commodity Insights T3 90
- Gartner T2 92
- Deloitte T2 78
Evidence quality
avg 68.8 · min 55 · max 95 · spread 40