-01.jpg)
The widespread boom in the domain of artificial intelligence (AI) across the globe is transforming global commercial and industrial landscape, causing a transformative shift in development and operations of data centers. The inclusion of data analytics, industrial IoT, smart manufacturing & robotics and others necessitate demand for specialized AI data centers which are to handle the complex requirements of AI workloads.
The enormous amount of data generated by AI applications require robust processing and storage facilities with high capacity and throughput to handle the data overflow, which incorporates leveraging a combination of high-speed SSDs, large-capacity HDDs, and distributed storage architectures hat eventually cause rise in the electricity consumption by the data center facilities.
The fast incorporation of AI into software programming across diverse sectors, increase the overall electricity demand of data centres. On comparing with the average electricity demand of a standard Google search (0.3 Wh of electricity) to OpenAI’s ChatGPT Electricity (2.9 Wh per request), and considering 9 billion searches daily, results in the necessity of approximately 10 TWh of additional electricity annually as reported by International Energy Agency.
The advancements in adoption of High-Performance Computing (HPC) systems which are designed to handle massive amounts of data and perform complex calculations at lightning speeds, further contribute towards the surge in the electricity consumption owing to the ability to utilize cutting-edge GPUs, FPGA (Field-Programmable Gate Arrays), and ASIC (Application-Specific Integrated Circuit) which are optimized for AI and machine-learning tasks, and require specialized cooling (such as direct-to-chip cooling or immersion cooling) and sophisticated power management systems which amplify the power requirements of AI-optimized data centers, leading to a significant increment in electricity consumption.
According to International Energy Agency (IEA), data centres and artificial intelligence (AI) consumed approximately 460 TWh of electricity across the globe in 2022, which comprises of nearly 2% of total global electricity demand. The surge in quantity of digital data is attributed majorly to the expansion and evolution of AI optimised data center processing and storage facilities across the globe.
The U.S Energy Information Administration (EIA) forecast predicts rise in electricity demand partially due to data centers, with overall power demand anticipated to rise to 4,193 billion kilowatt hours (kWh) in 2025 and 4,283bn kWh in 2026 from a record 4,097bn kWh in 2024.
Moreover, the International Energy Agency (IEA) anticipates the evolution of digital services and technological advancements in artificial intelligence sector as key drivers of surging electricity consumption by data centers, which is anticipated to fall in range between 620-1,050 TWh by 2026, with base case for demand at just over 800 TWh – up from 460 TWh in 2022 represents an additional 160 TWh up to 590 TWh of electricity demand in 2026 compared to 2022 which is roughly equivalent to adding at least one Sweden or at most one Germany.