AI’s Power Surge: Will Our Future Energy Demands Spark a Global Crisis?

In an era where digital transformation is rapidly reshaping every facet of our lives, the world stands at a critical crossroads. On the one hand, the push towards a sustainable, fossil fuel-free future is more urgent than ever, driven by the pressing need to combat climate change. On the other hand, the rise of artificial intelligence (AI) promises to revolutionise industries, economies, and societies, unlocking unprecedented levels of efficiency and innovation. However, this AI-driven future comes with a hefty price tag—one that could strain global energy resources and drive up consumer costs in the decades to come.

Today, data centers are the backbone of our digital economy, housing the servers that power everything from social media to cloud computing. According to recent estimates, these server farms consume up to 2% of the world’s total electricity—a staggering figure that rivals the energy consumption of mid-sized countries like Malaysia or South Africa. This number is only expected to grow as demand for digital services, cloud storage, and data processing continues to soar.

But while these figures are alarming on their own, they pale in comparison to the energy demands of AI data centers. Specialised AI servers, designed to handle the immense computational load of training and running complex machine learning models, can consume up to 50 times more energy than conventional data centers. Can you imagine 50 times more energy consumption! The sheer scale of this difference raises serious questions about the feasibility of maintaining both a low-carbon future and an AI-driven economy.

So what is the cost for a digital future powered by AI? My fear is that consumers may end up being caught in the crossfire where governments are pushing for a greener world and at the same time, racing to see who will be 1st in an AI-enabled future.

The ongoing transition away from fossil fuels—still the cheapest and most reliable source of energy in many parts of the world—could lead to higher electricity prices as renewable sources struggle to meet rising demand. We've seen that happen in Europe and elsewhere. This, combined with the energy-intensive nature of AI, suggests that the cost of living in a AI transformed world may be much higher than anticipated.

If this is the case then the implications are profound. If AI adoption continues at its current pace, consumers could face significant increases in their energy bills, as utilities pass on the costs of upgrading infrastructure and incorporating more expensive, less reliable renewable energy sources. This would disproportionately affect low-income households and small businesses, widening the economic divide.

I believe we should be having this conversation since the potential for energy shortages is serious, real and cannot be ignored. As we know many countries are planning to phase out fossil fuels in favour of renewables, like wind and solar, which in my view could struggle to keep up with the constant, massive energy demands of AI data centres. As our dependency on AI grows and it is already happening, it could lead to rolling blackouts, energy rationing, or even more severe consequences in regions where energy grids are already under strain.

I believe that the simultaneous push for AI and at the same time calls for the rapid phase-out of fossil fuels without an efficient substitute is a policy contradiction. AI is now unstoppable and without significant advancements in energy storage, grid infrastructure, and new power technologies, we could be setting ourselves up for an energy crisis that undermines the very progress AI promises to deliver.

Given these challenges, I think it’s fair to ask: Why are governments and industries so eager to embrace AI? The answer lies in the potential economic and societal benefits AI is meant to offer. From improving healthcare outcomes and reducing waste in manufacturing to driving innovations in transportation and finance, AI has the power to transform the global economy in ways we are only beginning to understand.

By the same token these benefits are downplaying or ignoring the surge in energy consumption. As things stand and knowing how technology tends to have a mind of its own once it is unleashed and AI is no different, I believe we cannot have both! So my next question is; Are we facing an inconvenient truth in that we cannot simultaneously phase out fossil fuels and build an AI-driven economy without significant consequences?

I believe that we are and something has to give. Either we slow down the enthusiasm for AI, allowing energy infrastructure and renewable technologies time to catch up, or we reconsider the timeline for phasing out fossil fuels to ensure we can meet the energy demands of the future. This is not a choice between progress and stagnation, but rather a decision about the pace and sustainability of that progress since all too often the most vulnerable in our society and the world at large pay the price of any transformation. I honestly believe that policymakers, industry leaders, and consumers alike must engage in a serious conversation about what kind of future we want—and how much we’re willing to pay for it. Additionally, I believe we should also be seriously evaluating the effects of a less aggressive timeline to phasing out fossil fuels.

Sources:

Masanet, E., et al. (2020). Recalibrating global data center energy-use estimates. Science 367(6481), 984-986.

Hao, K. (2019). Training a single AI model can emit as much carbon as five cars in their lifetimes. MIT Technology Review.

Jones, N. (2018). How to stop data centres from gobbling up the world’s electricity. Nature 561, 163-166.

IEA. (2021). Data centres and data transmission networks. International Energy Agency.