Insight

One atomic statement made by an expert, a student or a publication, with its evidence status.

Expert Disputed · What do these labels mean?

Roughly 11,000–12,000 data centres exist globally, consuming about 1.5–2% of the world’s electricity and accounting for somewhere in the region of 1–3% of global CO2; the market is estimated to grow 16% per annum to 2030.

— Paraphrased from the lecture

Where in the sourceearly (~00:14) — anchor: “grow at 16% per annum”
Confidencecheck — the electricity share reads “1.52%” in the transcript, and the CO2 share is explicitly hedged (“very difficult to provide an estimate”)

Context

In his introduction to what data centres are and where they are, Martins sizes the sector globally and argues it is the fastest-growing building or real-estate segment, putting pressure on land, power, water, biodiversity and communities. The electricity share reads “1.52%” in the transcript (most likely 1.5–2%), and he hedges the CO2 share as very difficult to estimate.

From the transcript

Current estimates is something in the region of like 11, 12,000 data centres globally. […] And it’s give or take about [1.5–2, garbled in the transcript as “1.52”]% of the world’s electricity and it keeps growing massively. And also in terms of carbon footprint, again, very difficult to provide an estimate, but we’re somewhere in a region of 1, 2, 3% of global CO2 emissions. […] Current estimates are essentially saying that between now and the end of the decade in 2030, this is a market that’s going to grow at 16% per annum, which is effectively huge.

Automatic transcript, presentation, 00:17:01. Lightly cleaned: repeated words and fillers removed, omissions marked […], editorial insertions in [brackets].

What needs checking

A verification pass against public sources (claim bank, fact-check item 2) found:

Electricity and growth confirmed (~415 TWh = ~1.5% in 2024, → ~945 TWh / just under 3% by 2030 at ~15% p.a.). The CO2 claim is wrong: IEA puts data-centre emissions at 180 Mt CO2 = ~0.5% of global combustion emissions. IEA, Energy and AI

Transcription

The number or wording may have been garbled in the transcript, or the speaker hedged it (see Part 6; the CO2 share is materially lower than stated). To be checked against the recording.

Connections


Something wrong or missing? Suggest a correction