Since its groundbreaking release on November 30, 2022, OpenAI's highly acclaimed project, ChatGPT, has revolutionized information access, placing knowledge just a few keystrokes away. But this accessibility to information must come at a cost, right?

Since 2022, we have seen the proliferation of AI data centers across North America to keep pace with increasing human reliance on AI. According to Digital Reality, “An AI data center is a facility equipped with a vast amount of computing resources specifically designed for AI workloads. It provides the infrastructure necessary to train and deploy complex machine learning models and algorithms” (The Impact of AI on Data Centers, n.d). These AI data centers are housed in large warehouses that demand energy efficiency, with large-scale centers continuously consuming between 20MW - 100MW of power. This is equivalent to the energy it would take to power a small city (Trus, 2024). In essence, AI is an energy-hungry behemoth. 

If the large amount of energy required to power data centers is not sourced with the environment in mind, this rapid surge in demand for energy could significantly spike worldwide pollution levels. This growing concern has led big tech companies to explore a slightly unorthodox approach. 

So why nuclear? Nuclear power plants not only provide a consistent, near-zero-emission power source through the nuclear fission of uranium, modern facilities also have a functional operating life spanning 40 to 60 years, ensuring decades of clean and reliable energy for power-intensive data centers (Novak, n.d). While nuclear plants inevitably always involve some level of risk, modern reactors are designed with multiple layers of security, such as containment structures, emergency cooling systems, and immediate shutdown mechanisms to reduce and nearly eliminate risk correlated to accidents and meltdowns. 

To strengthen their AI-related energy security, companies such as Google and Amazon already have deals with Kairos Power and X-Energy underway to complete the building of small modular reactors (SMRs) - efficient, innovative, and theoretically safer nuclear power plants with building completed before the year 2040 (Mazhar, 2024). Given big tech’s astronomical energy needs, other widespread renewables, like solar and wind power don’t cut it. Being too reliant on meteorological outcomes to produce energy and their occupation of seemingly endless land, they lack the consistency and resource efficiency that nuclear plants and AI need. 

With Canada holding 10% of the global uranium reserves and being the third country with the most reserves, trailing behind Australia and Kazakhstan, Canada is poised to lead the global shift towards efficient energy (Lu, 2024). Cameco Corporation (CCO.TO), the leading Canadian uranium mining player on the market, has exhibited significant market growth year-over-year, indicating an increasing demand for uranium and, in turn, nuclear power (Yahoo, 2025). 

Canada’s uranium-rich province, Saskatchewan, likely benefits from this trend as its northern mines, such as Cigar Lake, are central to the country’s mineral production. With the increasing interest in nuclear power, Saskatchewan will be the Canadian province to increase the country’s uranium production, further cementing Canada as a leader in the field (Lu, 2024). 

Whilst there is no simple or overnight solution to addressing AI’s monumental energy needs, the United States and its recent President-elect, Donald Trump, have recently announced their own 500$ billion project, “Stargate." Stargate aims to develop critical infrastructure to power the next generation of AI technology advancements. The initiative is focused on building massive data centers nationwide, powered in part by nuclear energy (Leparmentier, 2025).

Nuclear power is becoming the most reliable solution to the world’s growing energy needs. As big tech and government continue to invest in nuclear-backed projects such as Stargate, it seems as if the atom will likely power the future.

Works Cited:

The Impact of AI on Data Centers. (n.d.). Digital Realty. Retrieved January 31, 2025, from https://www.digitalrealty.com/resources/articles/data-center-ai

Introducing ChatGPT. (2022, November 30). OpenAI. Retrieved January 31, 2025, from https://openai.com/index/chatgpt/

Leparmentier, A. (2025, 01 22). Stargate, Trump's $500 billion project to boost artificial intelligence. Le Monde. Retrieved 02 07, 2025, from https://www.lemonde.fr/en/economy/article/2025/01/22/stargate-trump-s-500-billion-project-to-boost-artificial-intelligence_6737299_19.html

Lu, M. (2024, 02 13). Charted: Global Uranium Reserves, by Country. Visual Capitalist. Retrieved 02 01, 2025, from https://www.visualcapitalist.com/charted-global-uranium-reserves-by-country/

Mazhar, M. (2024, October 30). Microsoft, Google and Amazon turn to nuclear energy to fuel the AI boom. CBC. Retrieved February 3, 2025, from https://www.cbc.ca/radio/thecurrent/generative-ai-and-nuclear-energy-1.7362127

Novak, S., & Podest, M. (n.d.). Nuclear power plant ageing and life extension: Safety aspects. International Atomic Energy Agency. Retrieved January 31, 2025, from https://www.iaea.org/sites/default/files/29402043133.pdf

Trus, T. (2024, 08 01). Integrity Energy. Is Data Center Energy Usage Driving Up My Energy Bills? Retrieved 01 31, 2025, from https://www.integrityenergy.com/blog/is-data-center-energy-usage-driving-up-my-energy-bills/#:~:text=They%20typically%20consume%20between%201,than%20100%20MW%20of%20power.

Yahoo. (2025, 02 02). Cameco Corporation. Yahoo Finance. Retrieved 02 02, 2025, from https://ca.finance.yahoo.com/quote/CCO.TO/profile/