Intelligent, responsive, and flexible computing systems are essential for meeting global climate goals, as they can help optimize energy production and consumption. As the world keeps consuming more and more energy, the likelihood that we’ll be able to reach net zero by midcentury is seeming more and more like a pipe dream. The U.S. Energy Information Administration (EIA) has projected that global energy consumption will continue to rise through 2050 at a faster pace than efficiency gains and renewable energy installation capacities, meaning that mid-century climate goals will be difficult – if not impossible – to reach without significant shifts in consumer behavior and policy measures.
“Global population growth, increased regional manufacturing, and higher living standards push growth in energy consumption beyond advances in energy efficiency,” found the IEA’s International Energy Outlook 2023. The bulk of this population growth and rise in living standards is being driven by the global south, where birth rates remain high and economies are rapidly developing. At present, developing countries use much less power than richer nations per capita, but around 85% of all new energy demand in the near future is expected to come from outside the developed world according to the World Economic Forum.
It has been expected that developed countries will need to support the clean energy transition of poorer nations who will have to ‘leapfrog’ over typical fossil-fuel powered development trajectories and go straight to building out costly mass-scale renewable energy infrastructure. Developed nations, which have predominantly already peaked in terms of population growth and related infrastructure buildout, could then support these nations in terms of climate financing and clean energy exports.
But for the first time in a long time, the energy consumption of developed nations is growing at a significant rate at the same time that renewable energy buildout is facing slowdowns. While there’s no one single market trend or industrial sector that is reversing the curve of energy consumption in the developed world, a recent Forbes report notes that it can essentially be boiled down to four key factors: 1) “accelerating AI power demand,” 2) “a rush for metals for the energy transition and as an asset class,” 3) “inelastic demand from retirees”, and 4) “a rebuilding of the North American housing supply.”
Of these four key factors, the massive uptick in energy demand for Artificial Intelligence, not to mention that staggering energy consumption of data centers as a whole, looms the largest. “Currently, the entire IT industry is responsible for around 2 percent of global CO2 emissions,” Science Alert reported last year. And it’s growing at a breakneck pace. Technological research and consulting firm Gartner projects that in a business-as-usual scenario, the AI sector will be solely responsible for 3.5 percent of global electricity consumption by 2030. Already, the carbon footprint of Artificial Intelligence is almost as large as that of Bitcoin – in other words, it already consumes as much energy as many entire countries.
The energy demands of an increasingly AI-powered world are a sort of runaway train. The full scope of AI’s growth trajectory – as well as the potential of machine learning to help pave the way for smarter and more efficient energy use and production – is still poorly understood. Pandora’s box has only just been opened. It’s true that AI could become a net positive for the world’s carbon footprint if we can properly harness and regulate it.
Meeting global climate goals will require sweeping and unprecedented systems transformation. This imperative will be impossible without intelligent, responsive, and flexible computing systems able to rapidly recognize, respond to, and predict complex patterns of production and consumption. Already, AI is integral to renewable energy forecasting, smart grids, coordination of energy demand and distribution, maximizing efficiency of power production, and research and development of new materials.
“Fundamentally speaking, if you do want to save the planet with AI, you have to consider also the environmental footprint,” Sasha Luccioni, researcher of ethics at the open-source machine learning platform Hugging Face, told the Guardian last year. “It doesn’t make sense to burn a forest and then use AI to track deforestation.”
The growing energy needs of mining and metals exploration in the first world is also tied with decarbonization in complex and often ironic ways. The vast amount of manufacturing needed to build utility-scale solar and wind farms, electric vehicles, and an electric grid capable of handling a boom in both, will require massive amounts of metals and other rare earth elements that don’t yet have sufficiently established supply chains. As such, developed countries that don’t typically waste their time and money on primary materials markets are looking to get into the mining game to shore up their own stores of crucial components like lithium and copper. And all of that will contribute a whole lot of emissions to the atmosphere in the name of lowering them. At present, mining consumes between 5% to 10% of global energy.
The last two factors identified by Forbes – the static energy consumption of Baby Boomers and the desperate need for more housing in the United States and Canada – are also major factors keeping rates of energy consumption and greenhouse gas emissions high in developing countries. This poses a big problem for the global bottom line, as pathways toward net-zero were based on the assumption that energy demand in the developed world would remain relatively static. It’s also an issue for global inflation rates, which are likely to remain high. The only option is for global policymakers to get very serious about ramping up support for new and bigger clean energy projects in the immediate term.
By Haley Zaremba for Oilprice.com
Lead image (Credit: Reuters)
