{"id":255438,"date":"2025-11-10T19:33:14","date_gmt":"2025-11-10T19:33:14","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/255438\/"},"modified":"2025-11-10T19:33:14","modified_gmt":"2025-11-10T19:33:14","slug":"here-comes-the-energy-crunch","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/255438\/","title":{"rendered":"here comes the energy crunch"},"content":{"rendered":"

This article is an on-site version of our The State of AI newsletter. Sign up here<\/a> to get the newsletter sent straight to your inbox every Monday. To read earlier editions of the series, click here<\/a>. Explore all of our newsletters here<\/a><\/p>\n

Welcome back to The State of AI, a new collaboration between the Financial Times and MIT Technology Review. Every Monday for the next five weeks, writers from both publications will debate one aspect of the generative AI revolution reshaping global power. <\/p>\n

Last week, John Thornhill and Caiwei Chen argued that Beijing was on course to win the AI race<\/a>. Do you agree? (Nvidia\u2019s Jensen Huang seems to\u2009.\u2009.\u2009.\u2009) Join John in a live Q&A on the subject on November 13 at 1pm GMT. You can submit a question ahead of time here.<\/a><\/p>\n

This week, Casey Crownhart, MIT Technology Review\u2019s senior climate reporter,\u00a0and Pilita Clark, FT columnist and former environment correspondent, ask whether power limitations are about to cost the US its technological advantage. <\/p>\n

Casey Crownhart writes<\/p>\n

In the age of AI, the biggest barrier to progress isn\u2019t money but energy<\/a>. That should be particularly worrying here in the US, where massive data centres are waiting to come online and it doesn\u2019t look as if the country will build the steady power supply or infrastructure needed to serve them all.<\/p>\n

It wasn\u2019t always like this. For about a decade before 2020, data centres were able to offset increased demand with efficiency improvements<\/a>. Now, however, electricity demand is ticking up, with billions of queries to popular AI models each day. Efficiency gains aren\u2019t keeping pace and, with too little new power capacity coming online, the strain is starting to show. Electricity bills are ballooning for people who live in places where data centres place a growing load on the grid.<\/p>\n

If we want AI to deliver on its big promises without driving electricity prices sky high, the US needs to learn some lessons from the rest of the world on energy abundance. Just look at China.<\/p>\n

China installed 429GW of new power generation capacity in 2024, more than six times the net capacity added in the US during that time.\u00a0\u00a0<\/p>\n

It still generates much of its electricity with coal, but that makes up a declining share of the mix. Rather, the country is focused on installing solar, wind, nuclear and gas at record rates.<\/p>\n

<\/p>\n

Some content could not load. Check your internet connection or browser settings.<\/p>\n

\"\"\/<\/a><\/p>\n

The US, meanwhile, is focused on reviving its ailing coal industry. Coal-fired power plants are polluting and, crucially, expensive to run. Ageing plants in the US are also less reliable than they used to be, generating electricity just 42 per cent of the time, compared with a 61 per cent capacity factor in 2014.<\/p>\n

It\u2019s not a great situation. And unless the US changes something, it risks becoming a consumer as opposed to an innovator in both energy and AI tech. Already, China earns more from exporting renewables than the US does from its oil and gas exports.\u00a0<\/p>\n

Building and permitting new renewable power plants would certainly help, since they\u2019re currently the cheapest and fastest to bring online. But wind and solar are politically unpopular with the current administration. Natural gas is an obvious candidate, though there are concerns about delays with key equipment.<\/p>\n

One quick fix would be for data centres to be more flexible. If they agreed not to suck electricity from the grid during times of stress, new AI infrastructure might be able to come online without any new energy infrastructure.<\/p>\n

One study from Duke University found that if data centres agreed to curtail their consumption just 0.25 per cent of the time (roughly 22 hours over the course of the year), the grid could provide power for about 76GW of new demand<\/a>. That\u2019s like adding about 5 per cent of the entire grid\u2019s capacity without needing to build anything new.<\/p>\n

But flexibility wouldn\u2019t be enough to truly meet the swell in AI electricity demand. What do you think, Pilita? What would get the US out of these energy constraints? Is there anything else we should be thinking about when it comes to AI and its energy use?\u00a0<\/p>\n

Pilita Clark responds<\/p>\n

Hi Casey,\u00a0<\/p>\n

I agree. Data centres that can cut their power use at times of grid stress <\/a>should be the norm, not the exception. Likewise, we need more deals such as those giving cheaper electricity to data centres that let power utilities access their backup generators<\/a>. Both reduce the need to build more power plants, which makes sense regardless of how much electricity AI ends up using.<\/p>\n

This is a critical point for countries across the world because we still don\u2019t know exactly how much power AI is going to consume.\u00a0<\/p>\n

Forecasts for what data centres will need in as little as five years\u2019 time vary wildly, from less than twice today\u2019s rates to four times as much.<\/a><\/p>\n

This is partly because there\u2019s a lack of public data about AI systems\u2019 energy needs<\/a>. It\u2019s also because we don\u2019t know how much more efficient these systems will become. US chip designer Nvidia said last year it had delivered a 45,000 times <\/a>improvement in the energy <\/a>efficiency of its specialised chips over the previous eight years.\u00a0<\/p>\n

Moreover, we have been very wrong about tech energy needs before.\u00a0At the height of the dotcom boom in 1999, it was erroneously claimed the internet would need half the US\u2019s electricity within a decade \u2014 and therefore a lot more coal power.<\/p>\n

Still, some countries are clearly feeling the pressure already. In Ireland, data centres chew up so much power that new connections have been restricted <\/a>around Dublin to avoid straining the grid.<\/p>\n

Regulators are eyeing new rules that would force tech companies to provide enough power generation to match their demand. I hope such efforts grow. I also hope AI itself helps to boost power abundance and, crucially, accelerates the global energy transition needed to combat climate change.<\/p>\n

OpenAI\u2019s Sam Altman said <\/a>in 2023 that \u201conce we have a really powerful super intelligence, addressing climate change will not be particularly difficult\u201d.\u00a0<\/p>\n

The evidence so far is not promising, especially in the US, where renewable projects are being axed<\/a>. Still, the US may end up being an outlier in a world where ever cheaper renewables made up more than 90 per cent <\/a>of new power capacity added globally last year.\u00a0<\/p>\n

Europe is aiming to power one of its biggest data centres predominantly with renewables and batteries. But the country leading the green energy expansion revolution is clearly China.<\/p>\n

The 20th century was dominated by countries rich in the fossil fuels that the US now wants to prolong. China, in contrast, may become the world\u2019s first green electrostate. If it does this in a way that helps it win an AI race the US has so far dominated, it will mark a striking chapter in economic, technological and geopolitical history.<\/p>\n

Casey Crownhart replies<\/p>\n

I share your scepticism of claims by tech executives that AI will be a groundbreaking help in the race to address climate change. To be fair, AI is progressing rapidly. But we don\u2019t have time to wait for technologies standing on big claims with nothing to back them up.\u00a0<\/p>\n

When it comes to the grid, for example, experts say<\/a> there\u2019s potential for AI to help with planning and even operating, but these efforts are still experimental.\u00a0\u00a0<\/p>\n

Meanwhile, much of the world is making measurable progress on transitioning to newer, greener forms of energy. How that will affect the AI boom remains to be seen. What is clear is that AI is changing our grid and our world and we need to be clear-eyed about the consequences.\u00a0<\/p>\n

Further reading\u00a0Recommended newsletters for you<\/p>\n

The AI Shift \u2014 John Burn-Murdoch and Sarah O\u2019Connor dive into how AI is transforming the world of work. Sign up here<\/a><\/p>\n

Newswrap \u2014 Our business and economics round-up. Sign up here<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"This article is an on-site version of our The State of AI newsletter. Sign up here to get…\n","protected":false},"author":2,"featured_media":255439,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":{"0":"post-255438","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-artificial-intelligence","10":"tag-artificialintelligence","11":"tag-technology","12":"tag-uk","13":"tag-united-kingdom","14":"tag-unitedkingdom"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/255438","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=255438"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/255438\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/255439"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=255438"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=255438"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=255438"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}