{"id":124681,"date":"2025-11-20T07:02:40","date_gmt":"2025-11-20T07:02:40","guid":{"rendered":"https:\/\/alj.com\/?post_type=spotlight-fadyjameel&p=124681"},"modified":"2026-05-08T06:11:25","modified_gmt":"2026-05-08T06:11:25","slug":"can-ai-help-or-hinder-a-greener-future","status":"publish","type":"spotlight-fadyjameel","link":"https:\/\/alj.com\/en\/spotlight-by-fady-jameel\/can-ai-help-or-hinder-a-greener-future\/","title":{"rendered":"Can AI help or hinder a greener future?"},"content":{"rendered":"

Artificial intelligence (AI) \u2013 we are told \u2013 is the<\/em> secret weapon that could save us from a climate timebomb.\u00a0 It could potentially enable cleaner ways of living, reduce needless carbon emissions, and devise new ways of restoring our atmosphere to its natural balance.<\/h4>\n

There is just one problem . . . .\u00a0 \u00a0\u00a0AI itself has a significant carbon impact.<\/h4>\n

Developing and running AI models is emerging as one of the most energy-intensive activities in modern society.\u00a0 Today, one standard data center uses an amount of electricity equivalent to 100,000 homes.[1]<\/a>\u00a0 If that sounds alarming, it is just the beginning.\u00a0 The largest data centers currently in development are likely to use 20 times that much power, equal to an astonishing two million homes.\u00a0 Soon, data centers will be sucking as much energy from the grid as, for example, a power-hungry aluminum smelter.<\/p>\n

Anxiety about where all this electricity will come from is hardly impeding ambitions.\u00a0 Investment in data centers globally reached US$ 500,000,000,000 \u2013 half a trillion dollars \u2013 in 2024, doubling in the space of just two years.[2]<\/a><\/p>\n

In 2024, data centers consumed 1.5% of the world\u2019s energy output and emitted 1% of all greenhouse gases.[3]<\/a>\u00a0 Yet data center energy usage is growing four times quicker than the rate of total electricity consumption, with no sign of slowing down.\u00a0 In fact, two-thirds of organizations globally plan to invest heavily in generative AI over the next 24 months.[4]<\/a>\u00a0 With this trajectory, AI could be consuming up to 8% of the world\u2019s total energy output by 2030.[5]<\/a><\/p>\n

By that point, the International Energy Agency (IEA<\/a>) predicts AI\u2019s energy appetite will have risen to 945 TWh \u2013 rivalling the annual energy consumption of a mid-sized industrial nation.[6]<\/a>\u00a0 In the USA, the epicenter of international AI development, data centers will by 2030 consume more electricity than is used for the production of chemicals, cement and steel combined.<\/p>\n

\"\"<\/p>\n

Wall Street, as ever, is watching.\u00a0 AI is currently the rocket fuel driving US stock markets.\u00a0 Since 2022, when the first working AI models such as ChatGPT emerged, 65% of S&P 500 market capitalization growth (approximately US$ 12 trillion) has been achieved by companies openly fusing AI into their key functions.[7]<\/a>\u00a0 Last year, AI-focused startups in the USA reached median valuations of US$ 5.5 billion \u2013 five times more than other generic startups.<\/p>\n

\"\"<\/p>\n

Of course, this is not exclusively an American narrative.\u00a0 Beyond the USA, data centers in all mature economies are predicted to account for more than one-fifth of all growth in electricity demand this decade.[8]<\/a><\/p>\n

Data centers compromise the environment not just directly \u2013 through electrical consumption, refrigerants, fuel cells and generators \u2013 but indirectly too, via the manufacture of high-end equipment, associated cloud services, and the reprocessing of end-of-life equipment.\u00a0 Together, all these outputs exacerbate climate change, air and water pollution, and resource depletion.\u00a0 So, why do we still consider AI as such a pivotal asset in the existential battle against global warming?<\/p>\n

\"\"<\/p>\n

How can we reduce the environmental impact of AI?<\/h2>\n

Despite data centers representing such a significant drain on global power grids, scientists remain cautiously optimistic about the prospects of AI turning the tables on climate change.<\/p>\n

Four billion people currently live in areas deemed climate-vulnerable, and an estimated 250,000 extra deaths are forecast per year from 2030 due to heat stress, malnutrition and disease.[9]<\/a>\u00a0 As such, we need all the help we can get, even if our chief ally is digital rather than flesh and blood.\u00a0 The rapid evolution of AI heralds a whole range of game-changing climate interventions.<\/p>\n