A recent report from IDC has revealed that electricity consumption has become the main expense for enterprise data centers, representing 46% of their total operating costs. This percentage rises to 60% in the case of service provider data centers. According to the consultancy, the increase in electricity consumption is directly related to the growing adoption of high-energy-consuming workloads, such as artificial intelligence (AI), which is driving an increasing demand for data center capacity.
Artificial intelligence drives energy consumption growth
The IDC report highlights that artificial intelligence is having a significant impact on the increase in electricity consumption. As data centers take on more tasks related to AI, not only does the demand for processing capacity grow, but so does energy consumption and associated carbon emissions. IDC forecasts that by 2027, the capacity of data centers dedicated to AI will increase at a compound annual growth rate (CAGR) of 40.5%, while the energy consumption of these centers will grow at a rate of 44.7%, reaching 146.2 Terawatt hours (TWh) in that year. This trend indicates that AI workloads will represent an increasingly larger portion of total electricity consumption in data centers.
Additionally, IDC projects that global data center electricity consumption will double between 2023 and 2028, with an average annual growth of 19.5%, reaching 857 TWh in 2028. This increase in energy demand will have significant implications for both environmental sustainability and business operating costs.
Electricity prices and their impact on data centers
The growth in electricity consumption in data centers occurs in a context of continuous rise in energy prices, driven by various factors such as energy supply and demand, environmental regulations, geopolitical tensions, and extreme weather phenomena linked to climate change. IDC points out that this upward trend is likely to continue in the coming years, significantly raising the operational costs of data centers.
To illustrate the impact of these cost increases, IDC conducted a scenario planning study on a data center with 1 MW of IT load operating at 50% capacity in 2023, with a Power Usage Effectiveness (PUE) of 1.5. The analysis showed that in all scenarios studied, electricity spending will increase at an annual rate exceeding 15%, with some scenarios surpassing 20%.
Energy efficiency and the use of renewable energy
Sean Graham, Director of Cloud to Edge Data Center Trends Research at IDC, believes that there are various technological options to improve energy efficiency in data centers. These options include improving chip efficiency, adopting liquid cooling systems, and redesigning energy infrastructure.
However, Graham also emphasizes the importance of data center providers, especially those in the cloud and colocation services, prioritizing investment in renewable energy. According to IDC, solar and wind energy offer significant advantages both environmentally and economically, as they provide the lowest levelized cost of electricity (LCOE), which measures the average net cost of electricity generation over the lifespan of a generator.
In summary, electricity consumption in data centers is growing rapidly, driven by the rise of artificial intelligence and increasing processing demands. This trend is not only leading to an increase in operating costs but also in carbon emissions, highlighting the urgency of adopting energy-efficient and sustainable solutions. Investment in renewable energy sources emerges as a key solution to reduce both costs and environmental impact, while ensuring the long-term operational viability of data centers in an increasingly demanding environment.
Source: IDC (https://www.idc.com/getdoc.jsp?containerId=prUS52611224)