When Will AI Pay for Itself?
What AI costs, and what it is starting to give back.
Five things to say
Five ready answers to the questions this page gets asked most, each one sourced and copyable.
"Isn't AI using all our electricity?"
Data centres took 23% of Ireland's metered electricity in 2025, up from 5% in 2015, more than every urban household combined. Globally it's 1.5%, heading for 3%. Small share, fastest-growing large load, severe where it clusters.
(CSO 2026; IEA 2025-26)One source: A single primary document, usually an official statistic or a peer-reviewed paper.Projection: A scenario, not a measurement."But isn't it getting more efficient?"
Yes, and that's the problem. Google cut energy per prompt 33-fold in a year and its electricity use still rose 37%. Efficiency makes AI cheaper; cheaper means more.
(Google 2025-26)Two sources: Two independent organisations agree.Vendor: The company's own figure, unaudited."Isn't this only about electricity and carbon?"
No: the ILO counts one job in four worldwide as exposed to generative AI, and women's jobs are about twice as likely as men's to sit in the most exposed group - 4.7% against 2.4%, exposure, not jobs lost.
(ILO Working Paper 140, 2025)One source: A single primary document, usually an official statistic or a peer-reviewed paper."Where's the actual benefit?"
In a randomised trial of 105,934 Swedish women, AI-supported screening found 29% more breast cancers with no rise in false alarms and 44% less radiologist reading. Not a pilot, not a promise.
(Lancet Digital Health 2025)Two sources: Two independent organisations agree."Does it net out?"
The IEA's own sums: widely used, today's AI applications could cut 1.4 billion tonnes of CO2 in 2035, three to four times all data-centre emissions. The IEA also says there's "no momentum" to make it happen. So: starting to, not yet.
(IEA 2025)One source: A single primary document, usually an official statistic or a peer-reviewed paper.Projection: A scenario, not a measurement.
