By Pirassanth Thanapathy
Have you ever noticed how easy it is to just ask AI to rewrite a text, sum up an article, whip up an email, or even build a spreadsheet? With AI at our fingertips, speeding up our daily tasks feels effortless. But, have you ever wondered about the energy that is being consumed every time you hit “enter” on your prompt?
As AI technology advances and adoption accelerates, it is essential to examine the broader impacts of our digital choices. The exponential growth in AI-driven demand has triggered a significant expansion of data centers that are crucial for enabling and processing the sophisticated operations of AI systems.
The global electricity consumption attributed to data centers is anticipated to escalate, with annual growth rising from 12% in 2024 to a notable 17% in 2025, exceeding prior projections for that year. According to the International Energy Agency (IEA), worldwide data center electricity use is expected to double by 2030.
These figures represent projections; however, given the rapid pace of AI adoption, actual outcomes may vary significantly within a few years, causing more pressure on our grid capacity.
In 2024, data centers consumed approximately 415 terawatt hours (TWh) of electricity. For context, Canada’s total electricity consumption in 2025 was roughly 618 TWh (IEA Data Center Energy Consumption Report).
To put this into perspective, data centers alone consumed an amount of energy comparable to approximately two-thirds of Canada’s total electricity consumption. AI workloads are now the primary driver of the rapid increase in data center energy consumption.
In response to the growing energy demand due to AI, the Government of Canada has committed $10 million in federal investments toward projects leveraging AI technologies for innovative energy solutions and reduced energy costs.
This might sound counterintuitive, but while the use of AI does come with its own energy footprint, the potential it could unlock by identify new energy efficient opportunities seems to be greater!
The Canadian AI landscape is evolving rapidly. Of the 230+ data centers nationwide, Ontario is home to over 100, and data centers are expected to account for 13% of new electricity demand in the province by 2035 (Ontario Data Center Report).
In Quebec, data center electricity consumption is primarily driven by AI applications and is expected to increase by 600% by 2035. Hydro Québec is responding by proposing new rates for facilities consuming more than 5 MW annually (Quebec Data Center Energy Policy).
Alberta is emerging as a major player in AI data center operations, with a remarkable influx of proposed AI projects totalling approximately 21,000 MW from the grid. This figure is nearly double the province’s peak electricity demand (Alberta AI Data Center Report).
The primary challenge facing Canada is not a lack of investment in AI but rather the limitations of the existing electrical grid that struggles to accommodate large, concentrated loads. This highlights the urgent need to pursue and implement energy efficient solutions collectively.
Let’s face it, AI is not going anywhere. But, maybe that is exactly why we should pause and think about how our everyday choices ripple out, especially when it comes to energy.
Perhaps we ought to consider adopting sustainable AI practices that help minimize the energy required for every AI interaction while still delivering the same results.
For instance, providing direct and precise prompts minimizes redundant communication. Avoid additional courtesies such as “please” and “thank you” as one prompt. Similarly, consolidating several minor queries into a single, well-considered prompt can significantly reduce redundant AI usage.
To put this in perspective, a single AI text prompt typically consumes around 0.3 watt-hours (Wh) of electricity, making it one of the least energy-intensive AI operations. In contrast, generating an image with AI requires anywhere from 5 to 20 Wh, roughly the equivalent of charging your smartphone to full.
Sources report that ChatGPT alone processes over 2 billion prompts each day, a figure approaching the scale of Google’s 8 to 16 billion daily searches. The collective energy footprint of these billions of requests is a reminder that even small improvements in our AI habits can add up to meaningful energy savings on a global scale.
The next time you ask AI like ChatGPT or Microsoft Copilot to handle a task, consider not just how convenient it is but also the impact it has on one of our most precious resources, energy.
This is not meant to deter you from using AI; instead, it highlights the importance of improving energy efficiency and opting for sustainable practices. We should always look for ways to optimize systems, achieve more with less, and find new efficiencies even when current methods seem effective.
Would you like to be trained on a specific topic? Just let us know!
Express my interest
