Cloud Energy Demand Surges with the Growth of AI
What’s new. The energy demand of the world’s data centers is growing at break-neck speed. AI, along with other technologies like the public cloud and cryptocurrencies, is swamping aging power systems around the world and consuming large amounts of fresh water for cooling. This is delaying the closure of fossil fuel based power plants and challenging efforts to meet sustainability goals.
Why it matters. In an age where everyone’s focus needs to be on sustainability and reducing our greenhouse gas emissions, AI and cloud computing are both helping us to tackle the climate crisis and making it worse. Public cloud data centers are 4.7 times more carbon efficient than private data centers, making them a much better target platform option for enterprises working to achieve sustainability goals for current and future workloads. At the same time, in their never ending cycles to improve sustainability, Amazon has deployed AI powered robots to build solar farms to provide energy for their public cloud data centers and Google is using AI to provide in-depth insights to improve cooling efficiency.
The risk. However, the accelerating adoption of AI, which is 100 to 1,000 times as computationally intensive as traditional workloads, has thrown us into an era of hyper-consumption. The increasing power demands are harming other efforts to reduce dependency on fossil fuels and straining power grids globally with some countries considering restrictions on data center growth to address the issue in the short term. With AI poised to drive 160% increase in data center power demand by 2030, more radical changes are required.
Go deeper. In their Electricity 2024 report, the International Energy Agency estimates that the global energy demand of global data centers is expected to grow from 460 terawatt-hours (TWh) in 2022 to more than 1,000 TWh by 2026. Those estimates do not fully reflect the impact that AI could have on computing demands, with Google up nearly 50% in emissions since 2019, and Microsoft carbon emissions up 30%, AI is already affecting the power consumption by the major players and creating potential setbacks to net zero goals.
The takeaway. To get back on track to achieve their sustainability goals by 2030, the major technology players will need to make significant investments. The cloud providers’ goals for reaching net zero by 2030 are causing a scramble for new clean energy solutions, power grid improvements and greater efficiency in AI computing. There is no one single solution, but many efforts need to be combined to both improve the carbon footprint of the energy consumed, and increase the efficiency of AI:
The bottom line. The rapid adoption of AI will deliver huge benefits to technology companies and their customers, but it is not without its costs. Organizations need to be aware of the true impact of AI on their own sustainability goals, and the potential impact future regulations and restrictions could have as they look to deploy AI in their businesses.
-------------
A FRESH APPROACH | Chiefly & Co.’s fractional experts accelerate sustainability & profitability for companies driven to do well while doing good. Learn more at chiefly-co.com
Comments