During his opening address at the Adipec 2025 conference, Dr Sultan Al Jaber, UAE Minister of Industry and Advanced Technology and Adnoc's managing director and group chief executive, summed up the new reality facing the global energy system: “The true cost of AI is not just in code, it’s in kilowatts.”
Much of his speech this year focused on growth. Dr Al Jaber told officials and delegates that “the long-term outlook shows demand growth for every form of energy across every market. Our response to meet that demand should focus on the data, not the drama".
Dr Al Jaber said the world’s legacy grids and transmission systems are no longer fit for purpose. AI-driven workloads are placing stress not only on computing hardware, but across the entire energy chain, from generation and transmission to cooling and computation.
That shift underpins a profound transformation in how societies think about energy and, equally, knowledge or access to knowledge. The adage that “knowledge is power” has been turned on its head. In the age of AI, power is knowledge. Without the electricity to fuel high-performance computing, there can be no machine learning, no automation and no digital innovation.
From Moore's Law to megawatts
In 1965, Gordon Moore, co-founder of Intel, made the famous observation that the number of transistors on an integrated circuit will double every two years. At the time, few imagined it would define the trajectory of computing for the next half-century. Moore’s Law, as it became known, identified a steady rhythm that symbolised progress – faster, smaller and cheaper devices driving exponential gains in efficiency.
But as Dr Al Jaber's opening speech at Adipec 2025 made clear, that era is over. AI has shredded Moore’s Law. Computing progress is no longer measured by how many transistors can fit on a chip but by how much power can be generated to run ever-growing AI models. The relationship between computing and energy has been redrawn and Moore’s Law is no longer the benchmark for progress, power is.
Data economy’s appetite for energy
Data centres already consume more power than some countries. Most of that energy is needed to run and cool servers, preventing the heat generated by complex calculations from shutting them down. Global data centre electricity demand could reach 945 terawatt-hours by 2030, the International Energy Agency said. That is about the same as Japan's annual consumption.
The curve is becoming steeper. Compute power for training AI systems has almost doubled every six months since 2010, a pace that far outstrips Moore’s original two-year cycle. Each leap in AI capability now requires an equally significant increase in electricity.
AI leadership in the Gulf
As AI becomes integral to sectors from health care and logistics to manufacturing and media, the UAE’s energy strategy is evolving from a focus on supply and demand to a plan that enables data-driven growth. Powering intelligence is now part of the national mission.
At Adipec 2025, Dr Al Jaber’s message captured this moment of transformation. The technology driving the next industrial revolution will depend not only on code and connectivity, but on the energy that makes intelligence possible.
