Electricity is by far the largest ongoing expense for data center operators, accounting for 46% of total spending for enterprise data centers and 60% for service provider data centers. Electricity consumption is growing rapidly as data centers take on more workloads and more energy-intensive workloads, such as artificial intelligence (AI), according to a new International Data Corporation (IDC) report.
The IDC expects the surging demand for AI workloads will lead to a significant increase in data center capacity, energy consumption, and carbon emissions, with AI data center capacity projected to reach a compound annual growth rate (CAGR) of 40.5% through 2027. Similarly, AI data center energy consumption is forecast to grow at a CAGR of 44.7%, reaching 146.2 terawatt hours (TWh) by 2027 with AI workloads consuming a growing portion of total data center electricity use.
Overall, the IDC expects global data center electricity consumption to more than double between 2023 and 2028, noting a five-year CAGR of 19.5%, reaching 857 Terawatt hours (TWh) through 2028.
At the same time, electricity prices are rising due to supply and demand dynamics, environmental regulations, geopolitical events, and sensitivity to extreme weather events fueled, in part, by climate change. The IDC believes the trends that have caused electricity prices to increase over the last five years are likely to continue.
Rising consumption and increased energy costs will make data centers considerably more expensive to operate, however, the exact CapEx is uncertain.
As part of the study, the IDC conducted scenario planning for a data center with 1 MW of IT load, running at 50% capacity, with a power usage effectiveness (PUE) of 1.5. The study looked at three energy price growth scenarios using energy pricing and growth rates for the United States, Germany, and Japan. In all three scenarios, the percentage growth in electricity spend exceeded a CAGR of 15% in all cases, with most scenarios showing growth of over 20%. The study also revealed that an additional 10% in energy efficiency can offer considerable savings to data center operators.
Latest Wholesale and Capacity News: Omantel Launches New Data Center in Salalah, Dhofar Region
Interestingly, Ooredoo Group recently announced a landmark QAR 2 billion financing deal with QNB, Doha Bank, and Masraf Al Rayan, to accelerate the growth of its data center and AI business.
Oil rich gulf nations are heavily investing in AI to diversify its economies from fossil-fuel dependency. They believe they can provide the cheap power needed to run the energy-guzzling data centers for AI uses.
Turning to Renewables
Apart from technological solutions like improved chip efficiency, liquid cooling and rethinking data center design as options to increase data center efficiency, experts suggest that data center providers, including cloud and colocation services, should continue to prioritize investment in renewable energy sources.
Solar and wind power, in particular, offer significant environmental advantages while also providing the lowest levelized cost of electricity (LCOE), which reflects the average net present cost of electricity generation over a generator's lifetime. By collocating facilities at (or near) the source of renewable energy generation, providers can reduce both construction costs and energy losses associated with distribution, enhancing overall efficiency and sustainability while also improving resiliency by removing grid reliability parameters.
Also Read: Digital Dubai Pioneers Journey Towards AI-Powered Data Center Economy
Nuclear Option
Seen as a more consistent source of power than solar and wind, many tech companies are betting on nuclear energy's rapid development to meet AI's electricity demands.
In recent news, Microsoft signed a 20-year-deal with US electricity company, Constellation Energy, to use power from the Three Mile Island nuclear station—the site of America’s worst nuclear accident in 1979. This deal will help restart operations at the plant to supply electricity for Microsoft’s growing AI and cloud services.
Similarly, in March, Amazon's AWS agreed to invest USD 650 million in a data center campus powered by another 40-year-old Pennsylvania nuclear plant.
These needs are not easy to be met. US data center power use is expected to roughly triple between 2023 and 2030 and will require about 47 gigawatts of new generation capacity, according to Goldman Sachs estimates, which assumed natural gas, wind and solar would fill the gap.
Tech companies are also interested in small modular reactors (SMRs), which are more compact and potentially easier to deploy. However, this technology is still in its infancy and lacks regulatory approval, leading companies to seek out existing nuclear power options.
Data centers are the heart of the digital economy and the demand for data centers is expected to rise substantially, positioning them as a primary focus for growth and investment. But the cost of operating data centers is also increasing substantially due to rising electricity prices and increased data center consumption. Climate conscious investors and regulators are keen to ensure that such spike in power does not trigger a huge rise in greenhouse gas emissions.
Continue Reading:
Data Centers: Marking AI Dominance in the Middle East
Umniah Constructs Largest Tier III Data Center in Jordan
AI in the Data Center: Are MTP® Connectors Revolutionizing the Technological Landscape?
