Big Tech companies like Microsoft, Amazon, and Google are turning to nuclear energy to power their growing artificial intelligence (AI) operations. This shift is driven by the massive amounts of electricity required to train and run AI models. In fact, the training process for large AI models can consume hundreds of billions of words, taking months to process and requiring enormous amounts of energy.
The demand for electricity to power AI is growing exponentially, with estimates suggesting that data centers will consume around 8% of all US electricity by 2030. To put this into perspective, a single ChatGPT-powered Google query uses almost 10 times as much energy as a traditional Google search.
To meet this growing demand, tech giants are investing in nuclear energy. Microsoft, for example, plans to revive the dormant Three Mile nuclear power plant in Pennsylvania to power its data centers. Similarly, Amazon has purchased a new data center next to a nuclear power plant in Pennsylvania, with plans to increase its energy supply.
While the shift towards nuclear energy may seem counterintuitive, given the industry's focus on renewable energy, experts argue that nuclear power is necessary to meet the massive energy demands of AI. As Leslie Miley, technical advisor to Microsoft's chief technology officer, put it, "AI has a dirty secret. It's dirty".
The growing reliance on nuclear energy raises important questions about the environmental impact of AI and the tech industry's ability to meet its climate goals. As the demand for electricity to power AI continues to grow, it's clear that the industry will need to find innovative solutions to balance its energy needs with its environmental responsibilities.
