Nuclear Energy and AI: A Match Made in Hype Heaven?
When your phone starts asking for juice, it’s either time for a charger or to contemplate the end of humanity. But what’s cooking in the land of Big Tech while they add more clouds to your digital space? Spoiler alert: it’s not just the tech gods kicking back with cocktails!
According to recent updates buzzing around, we’re witnessing an explosive demand for nuclear power. Yes, you heard that right! The illustrious names like Google, Amazon, Meta, and Microsoft are eyeing nuclear energy not just as an environmental band-aid but as their new best friend. Why? Because powering your AI-laden world with low-carbon electricity is a tad trickier than convincing your grandma to use Venmo.
Small Modular What? Um, Reactors!
Let’s unpack this delightful little nugget: Google has tasked an American start-up, Kairos Power, to whip up some Small Modular Reactors (SMRs). Simply put, they’ve commissioned nuclear power plants that are about as big as your garden shed, or at least you’d think they were if you lived in a ‘low-carbon’ dystopia. With a total capacity of 500 megawatts —that’s approximately enough power to keep a small country running or binge-watch every episode of your favorite show— these reactors are touted as the next miracle in clean energy. But what’s the catch? Well, Google has been quite coy on whether these mini-nukes will be plugged into the grid or keep their neighbors up all night powering those data centers!
The Irony of Clean Energy
Now, let’s take a moment to pause for dramatic effect here. Google prides itself on a 15-year journey to “clean energy.” But hold on to your smartphones: emissions have skyrocketed by 13% this year alone, and a whopping 48% compared to five years ago! You could say they’re more committed to cleaner energy than ever—at least 13% more committed, anyway! And what’s sparking this sharp rise? Well, surprise, surprise, it’s the voracious appetite for AI, demanding more energy than your average toddler demands attention.
Data Centers: The Energy Gobblers
Data centers are the black holes of the tech universe, consuming energy like a kid in a candy store. With demands soaring to a staggering 100 MW or more (equal to powering 400,000 electric cars), these structures can have an environmental impact that would make a bulldozer shed a tear. If that statement doesn’t ring alarm bells, the International Energy Agency thinks you might want to lie down! Data centers currently guzzle down about 1.5% of global electricity, but hold on tight—what’s on the horizon can push this figure even higher!
The Nuclear Comeback: Watch This Space!
It seems like amidst the chaos, the nuclear industry is making a grand comeback! With less adverse effects than an awkward family gathering, tech giants are rekindling their love affair with nuclear power. Microsoft is even flirting with the formerly decommissioned Three Mile Island plant! I mean, if you can’t beat ’em, join ’em, right? It’s as if we’re witnessing a nuclear renaissance—minus the historical context of, well, you know… catastrophic failures.
A Dose of Reality: Risk and Humor
Investors are licking their lips at the high prices these tech titans seem willing to pay, anticipating a nuclear revival. But just when you thought the spotlight would settle purely on the tech giants, it’s crucial to remember: the realization isn’t as easy as it seems. The question lingers—the high-risk nuclear projects often stretch timelines into infinity and budgets into the realm of fantasy.
The Bottom Line
Are we genuinely heading for a clean-energy utopia powered by nuclear energy, or is this just another dazzling distraction? If Big Tech continues to expand data centers at this rate, I’d recommend we keep a helmet on standby… or maybe a nuclear fallout suit just to be safe. Either way, the dance between AI power-hungry giants and their nuclear ambitions is nothing short of a dramatic soap opera. Grab the popcorn; it looks like the show is just getting started!
ANSA
Is nuclear energy the zero-carbon solution to powering artificial intelligence? We are seeing a surge in demand for nuclear power plants from American big tech companies like Google, Amazon, Meta and Microsoft, who need large amounts of low-carbon electricity, around the clock, to run their data center and, consequently, everything we have on our smartphones and computers, including AI apps.
The latest news, in chronological order, concerns the Mountain View giant which has commissioned the production of some small modular nuclear reactors (Small Modular Reactors, Smr) to the American start-up Kairos Power, in order to produce and supply electricity at low carbon emissions to its energy-intensive data centers. Google and Kairos have made it known that the SMRs will have a total capacity of 500 megawatts, but have not yet clarified whether the mini-reactors will power the electricity grid or will be connected directly to the data centers. This agreement is «a milestone for us at Google in our 15-year journey towards clean energy – explained Michael Terrell, senior director for energy and climate at Google – we believe that nuclear can play an important role in helping us meet our energy demand, cleanly and around the clock.”
The key word is “clean energy”: last July data were released on the growth of the technological giant’s emissions: +13% in 2023 compared to the previous year. But even compared to five years ago, emissions had increased by 48%. And among the causes of this increase there is certainly the disruptive development that artificial intelligence (AI) is having which requires an enormous amount of energy to carry out complex calculations. Some experts argue that the rapidly expanding data centers needed to power AI threaten the entire transition to clean electricity. That’s because a new data center can delay the closure of a fossil fuel-burning power plant or push the construction of a new one. Data centers not only consume a lot of energy, but they also require high-voltage transmission lines and require significant amounts of water to stay cool. For all these reasons they have a great environmental impact.
On the investment side, suffice it to say that in 2023, the combined capital investment of Google, Microsoft and Amazon, the industry leader in AI adoption and data center deployment, was higher than that of the entire industry US oil and gas, totaling about 0.5% of US GDP. Parallel to this need for big tech has been growing excitement in the nuclear industry: while China and South Korea have been busy building nuclear reactors in recent decades, the United States and Europe have seen a decline in this infrastructure, also following the accidents at Three Mile Island in 1979, Chernobyl in 1986 and the Tsunami that hit the Fukushima power plant in 2011 which caused mistrust and hostility in public opinion towards this technology. Last month, Microsoft announced it would reactivate the decommissioned Three Mile Island nuclear power plant in Pennsylvania, while in March Amazon paid $650 million to build a data center next to the Susquehanna Steam Electric nuclear power plant, also in Pennsylvania. Investors say the high prices big tech is willing to pay could trigger a wave of investment in new nuclear power plants. While structural challenges remain behind all the hype, including the fundamental question of who will be willing to shoulder the high risk of nuclear projects, which can extend years beyond deadline and be billions over budget.
To ensure supplies, Google has invested in small reactors, while Microsoft and Amazon are hooking up to old power plants
What is certain is that large data centers, hyperscaleincreasingly common, have a power demand of 100 MW or more, with annual electricity consumption equivalent to the electricity demand of approximately 400,000 electric cars. According to the International Energy Agency (IEA), these data must be put into perspective: global sales of electric cars will reach 17 million in 2024. Today, data centers account for around 1.5% of global electricity consumption and the annual electricity consumption of data centers globally is about half the electricity consumption of household appliances, such as computers, phones and TVs. But in the coming years, the IEA has predicted a substantial increase in the number and size of data centers: consequently, the demand for electricity will grow strongly by 2030, leading to considerable pressure on local electricity networks, accentuated by the enormous discrepancy between rapid data center construction times and the often slow pace of expanding and strengthening energy infrastructure. If the increase in electricity consumption of data centers could make it more difficult to achieve the climate objectives declared by big tech companies such as Amazon and Microsoft, at the same time, the inadequacies or delays of the energy sector itself could end up slowing down the development and implementation of AI.
