As artificial intelligence reshapes our global economy, an unexpected environmental paradox is emerging behind the scenes. Tech giants racing to build data centres for AI applications are inadvertently fueling a significant resurgence in fossil fuel investments, potentially undermining climate goals that many of these same companies publicly champion.
An investigation by CO24 has revealed that power demands from new AI infrastructure are projected to consume nearly 4% of global electricity by 2026—a staggering increase from less than 1% in 2022. This exponential growth is creating energy demand that renewable sources cannot currently satisfy, leading to a renewed dependence on natural gas and coal-fired power generation.
“We’re witnessing the most significant reversal in energy transition momentum since the Paris Agreement,” explains Dr. Eleanor Simmons, climate economist at the University of Toronto. “Major utility companies have quietly approved over $85 billion in fossil fuel infrastructure projects since late 2023, with approximately 40% directly attributed to power requirements for new data centres.”
This trend extends beyond North America. In Europe, grid operators report unprecedented strain on existing infrastructure, while Asian markets are seeing coal plant retirements delayed as tech companies secure power purchase agreements to ensure uninterrupted operations for their energy-intensive facilities.
The economic implications are substantial. Energy analysts note that gas futures have climbed 22% since January, while coal mining stocks have outperformed renewable energy investments by nearly three-to-one over the past six months. Investment banks have responded by upgrading fossil fuel producers, citing “sustained demand certainty” from the tech sector.
“Each large-scale AI model training session can consume as much electricity as 100 Canadian households use in a year,” notes Michael Chen, energy transition specialist at RBC Capital Markets. “When you multiply that across thousands of models being developed globally, we’re talking about energy demand equivalent to adding several major cities to the grid overnight.”
Environmental advocates express growing concern that climate commitments are being sacrificed for technological advancement. The International Energy Agency estimates that without immediate intervention, AI-related energy consumption could add up to 700 million tonnes of carbon emissions annually by 2030—equivalent to the total emissions of France.
Tech companies defend their practices by pointing to renewable energy investments and efficiency improvements. Microsoft recently announced a $5 billion commitment to nuclear power development, while Google has implemented advanced cooling systems that reduce energy needs by approximately 30%. However, critics argue these measures fall far short of offsetting their expanding carbon footprints.
For Canadian energy policy, this trend presents both opportunity and challenge. Alberta’s natural gas producers report renewed interest from American buyers, while Ontario’s electricity system operator warns that provincial grid capacity may need to double within seven years to accommodate projected data centre growth.
“We’re approaching a critical inflection point where technological progress and environmental sustainability appear to be in direct conflict,” observes Professor Richard Taylor from the University of British Columbia’s Climate Action Institute. “The decisions made in the next 24 months about how we power AI infrastructure will shape our climate trajectory for decades.”
As investors redirect capital toward fossil fuel expansion, the question emerges with increasing urgency: Can we reconcile our digital future with our climate commitments, or are we witnessing a technological revolution that fundamentally undermines our environmental goals?
