October 23, 2024 - Nuclear Energy Gives Data Centers Promise for the Future
It's no secret that global internet traffic has been on the rise - and even more remarkably in the years since the pandemic instigated many operations to transition online. A second bump in internet usage followed the introduction of chatbots and the consecutive release of publicly available AI thereafter. And while artificial intelligence shows itself as an ever-promising resource to promote efficiency across industries, there remains one eminent discussion about its heightening utilization: energy consumption.
Let's take a brief look at the use of AI today. Artificial intelligence has technically been around for decades. Yet it wasn’t until OpenAI’s chatbot release in 20221 that the demand skyrocketed - and tech responded by introducing a long line of AIs for everything from generative imaging to resume writing. Aside from its most familiar uses, AI has long since proven to be quite advantageous for applications in research, predictive modeling, process creation and more. According to one study, AI can reduce CO2 emissions by up to 40% when used in runsmart homes2. By having a larger scope of data and faster analysis times, it can streamline operations to cut emissions and costs across all industries. Another example is the EPA’s use of an AI to detect water pollution violators, which improved detection by 600%4. There are endless real-time examples of the benefits of artificial intelligence to conserve resources which seem to combat its own energy consumption and begs the question - can AI be a net positive?
There's a rush to build more data centers as demand skyrockets. In the US alone, data centers are nearly at their maximum capacity - and the US leads the world in data center numbers3. While data centers take up 1% of global energy consumption4, it’s estimated that AI takes up nearly 20% of data centers’ storage capacity3. This number has only been growing as AIs become more complex. In turn, electric companies are burdened with balancing the expanding demands of data centers with those for all other existing customers1. To mitigate this problem, several notable companies are turning towards nuclear energy: Microsoft, Amazon, and Standard Power to name a few. Just this last week, Google signed a contract with Kairos5 to buy electricity from several of their SMRs in the next decade as they are developed6. SMRs - or small modular reactors - in particular are gaining traction because of what their name implies: smaller reactors enable placement in less populated areas, and modular designs not only enable quick production times but also future scalability as power needs increase7. A huge bonus as advertised by Kairos, is the passive safety features of SMRs that make them safer alternatives to their traditional counterparts. If SMRs enable data centers to have their own energy source apart from utilities for the general public, the discussion around the resource consumption of artificial intelligence might be nearing a conclusion.
With investments from big tech, it’s safe to state that the new age of nuclear energy has begun - and we're along for the ride.
References
2 As Use of A.I. Soars, So Does the Energy and Water It Requires
3115 Data Center States You Should Know in 2024
4 AI and Environmental Challenges
5 Google to buy power for AI needs from small modular nuclear reactor company Kairos
7 Small modular reactors may have climate benefits, but they can also be climate vulnerable