For years, the tech world has sold us a clean, almost ethereal vision of the future. A world run on elegant algorithms and powered by the ‘cloud’—a term so soft and intangible it sounds like it floats on air. Well, it’s time for a reality check. The explosion in Artificial Intelligence isn’t happening in the ether; it’s happening in hulking, humming data centres, and they are unimaginably thirsty for power. The dirty little secret is that the AI revolution might just be powered by the fuels we were desperately trying to leave behind.
Forget the neat and tidy story of ‘energy transition’. According to energy executives speaking at the recent LNG 2026 conference, as reported by Al Jazeera, we are now in the age of ‘energy addition’. The soaring demand from AI is rewriting the rules, and frankly, the electricity grid is being pushed to its breaking point.
So, What’s the Big Deal with AI Energy Consumption?
At its heart, AI energy consumption is simply the electricity required to train and run complex machine learning models. But “simply” is doing a lot of work in that sentence. Training a large language model is not like charging your phone; it’s like launching a rocket. It involves staggering computational resource demands, with thousands of specialised chips running at full tilt for weeks or even months on end.
Think of it this way: your brain is incredibly efficient, running on about 20 watts – the power of a dim lightbulb. An AI model doing a similar, albeit less complex, task can require megawatts. That’s the power consumption of a small town, all focused on generating text, images, or code. And that’s just one model. Now, multiply that across every company rushing to slap “AI-powered” on its products. See the problem?
Data Centres: The Unseen Factories of the Digital Age
This voracious appetite for electricity leads us directly to the doorstep of the data centre. These aren’t just storage depots for your holiday photos. They are the physical heart of AI, packed with servers that generate immense heat and require constant, powerful cooling systems. The growth in data center power requirements is nothing short of astonishing.
We’re witnessing a fundamental shift where data centres are becoming one of the most significant drivers of new electricity demand globally. The infrastructure that underpins our digital lives is straining, and the AI boom is throwing fuel on the fire.
The Grid Can’t Keep Up
The challenge is that our electrical grids were built for a different era—one with predictable, centralised power generation. Now, they face a double whammy: the intermittent nature of renewables and the relentless, 24/7 pull from AI-driven data centres. This creates massive instability.
As Shell CEO Wael Sawan noted, “When the wind isn’t blowing and the sun isn’t shining, gas fills the gap.” This statement perfectly captures the pragmatic, if uncomfortable, reality that the energy industry is now championing.
The Myth of Perfectly Sustainable AI Development
This brings us to the thorny issue of sustainable AI development. While a laudable goal, the current reality looks very different. The tech industry loves to talk about its commitment to renewables, but AI’s constant power needs don’t pause when it gets cloudy.
The energy industry’s solution? Natural gas, specifically Liquefied Natural Gas (LNG). They argue it’s the only reliable, scalable source capable of balancing the grid and meeting AI’s insatiable demand. Global LNG demand is projected to soar from 400 million tonnes per year today to 600 million by 2030. That’s not a transition; that’s an acceleration.
A Case Study in Realpolitik: Qatar’s Big Bet
Look no further than QatarEnergy. As detailed in the Al Jazeera report, the state-owned giant is commissioning a fleet of nearly 200 new LNG carriers—one of the largest fleet expansions in energy history. They, along with partners like ExxonMobil, are betting big that the world will need huge amounts of gas for decades to come, largely to power our digital ambitions. Their argument is that gas is a cleaner alternative to coal, reducing emissions by around 40% in power generation. It’s a compelling argument, but one that still locks us into a fossil fuel future.
Bottlenecks, Bureaucracy, and the Real World
Even with this massive pivot to gas, there’s another problem: the physical infrastructure can’t be built overnight. In many Western countries, getting a permit for a new power line or a gas plant is a bureaucratic nightmare that can take a decade. The speed of digital innovation is colliding with the slow, grinding pace of building things in the real world. We want AI-powered everything, but we haven’t figured out how to plug it all in.
This disconnect between digital ambition and physical reality is perhaps the biggest story in tech and energy right now. The AI industry has created a demand curve that shoots straight up, while the energy supply curve is struggling under the weight of regulation, infrastructure limits, and capital investment cycles.
The Moral Dimension: Powering AI vs. Powering Homes
There is also a profound ethical question at play here. As we debate how to power the next generation of chatbots, we can’t forget what Qatar’s energy minister, Saad al-Kaabi, powerfully stated: “Nearly a billion people still do not have basic electricity. We cannot deprive them of growth.”
Is it right for the developed world’s demand for more sophisticated AI to consume vast energy resources that could otherwise be used to bring basic electricity to millions? The energy industry is positioning itself as the solution to both problems—powering AI and ending energy poverty—with gas as the common denominator. It’s a clever strategic narrative, but it conveniently sidelines the climate implications.
The comfortable consensus around a slow, steady move to renewables has been shattered. The AI industry, in its race for supremacy, has inadvertently thrown a lifeline to fossil fuels, creating a new and complex energy paradigm. The conversation is no longer just about algorithms and efficiency; it’s about pipelines, power grids, and geopolitics.
So, as we marvel at the latest AI creation, it’s worth asking: what is the true cost of this progress, and are we prepared to pay the price? What do you think?
