A drilling rig stands at a Fervo Energy geothermal site under construction near Milford, Utah, in this 2023 photo. Experts say growing demand for energy to power AI is also fuelling development of alternative forms of energy. (Ellen Schmidt/The Associated Press)
The Energy Impact of AI: Balancing Progress and Sustainability
Using artificial intelligence (AI) can often feel like witnessing a magic trick. With a simple request to ChatGPT, users can receive an image, no matter how imaginative or unconventional the prompt may be. However, this seemingly effortless technology comes at a significant cost. Behind the scenes, powerful servers are working tirelessly, consuming vast amounts of energy. Estimates suggest that a single query to ChatGPT may use about ten times more power than a Google search, while generating an image consumes energy comparable to charging a smartphone.
As demand for AI grows, concerns have arisen regarding the environmental impact of the energy required to support these technologies. Investment bank Goldman Sachs predicts that carbon dioxide emissions from data centers could more than double from 2022 to 2030, raising alarms about a potential surge in fossil fuel-related emissions.
Nevertheless, not all experts view this trend as entirely negative. The increasing energy demands of AI are also propelling advancements in alternative energy sources such as wind, solar, geothermal, and nuclear. These developments could play a crucial role in the global energy transition.
The Need for Power
With AI's energy requirements projected to soar, tech companies are heavily investing in data centers, the sprawling facilities that house servers, routers, and cooling systems essential for running AI technologies. According to Bloomberg News, the five leading U.S. companies invested approximately $105 billion in data centers in 2023, a figure expected to rise to $187 billion by 2028. Globally, power demand from data centers is anticipated to increase by 160% by 2030, according to the Canada Energy Regulator.
Tech firms are faced with the challenge of sourcing this energy while balancing costs, reliability, and environmental considerations, especially in light of their climate commitments. Ed Crooks, vice-chair for the Americas with Wood Mackenzie, emphasized the complexity of these competing priorities during a recent energy conference in Calgary. He noted that the quickest way to meet immediate energy needs is often through gas-fired power plants, despite their carbon emissions, which contradict many companies' goals for carbon neutrality.
In Canada, the distribution of data centers varies, with most located in provinces abundant in hydroelectric power. Ontario hosts 105 centers, while Quebec and British Columbia have 57 and 35, respectively. In contrast, Alberta, which primarily relies on natural gas, has 22 data centers but is positioning itself as a prime location for new facilities. The Alberta Electric System Operator (AESO) recently reported a surge in applications for data center projects, representing around 5,000 megawatts of capacity at preliminary stages.
One solution to mitigate the carbon footprint of natural gas plants is the integration of carbon capture technologies. However, this approach can increase project costs, raising questions about its economic viability within the industry.
The Potential for Renewable Energy
As tech companies strive to secure reliable energy sources, many are investing in renewable energy projects, even those with high initial costs, such as geothermal energy. Google, for instance, has launched a groundbreaking geothermal power project in Nevada to supply energy to its nearby data centers and is exploring the use of small modular reactors (SMRs) for similar purposes. Amazon has also emerged as the world’s largest corporate buyer of solar and wind energy, while Microsoft recently made headlines for its agreement to restart the Three Mile Island nuclear power plant in Pennsylvania.
Experts believe that as major tech companies develop large-scale alternative energy projects, they can streamline processes and reduce costs, making it easier for others to follow suit. Keaton Horner, director of analytics at Orennia, a Calgary-based analytics firm, noted that if companies like Microsoft invest heavily in innovative energy technologies, they could significantly lower costs within five to ten years, making them competitive with traditional energy sources.
However, securing sufficient renewable energy remains a challenge. A recent carbon removal megaproject in Wyoming was canceled due to fierce competition for renewable energy from data center operators and cryptocurrency miners, highlighting the difficulties of sharing resources in a growing energy market.
Looking Ahead
Crooks pointed out that the impact of AI on energy and the environment cannot be categorized as solely positive or negative. Nevertheless, he believes that as long as stakeholders continue to push for investments in technologies that promote decarbonization, AI could ultimately contribute to a more sustainable energy future.
“What AI does is it creates a power industry and energy industry that is looking for the future,” he stated. “And I think, at the end of the day, that is going to make the energy transition more possible and drive more progress than if we didn’t have AI.”
