Computers and energy use
Computers and their energy use have become a significant concern as demand continues to rise, particularly with advancements in cryptocurrency and artificial intelligence (AI). Since the mid-twentieth century, the development and proliferation of computers have transformed energy consumption patterns, making it challenging to assess their overall energy impact. While computers consume energy during use and in their manufacturing, they also influence human behavior and societal organization, leading to potential energy savings through innovations like satellite navigation and video conferencing.
The integration of smart technologies into energy markets is changing the way consumers interact with their energy systems, allowing for better monitoring and management of energy use. Historically, early computers were large industrial devices that consumed considerable energy, yet their limited numbers meant their overall impact was minimal. With the advent of microchip technology and personal computers, energy efficiency improved, but the sheer increase in device numbers has led to greater energy consumption overall.
In the twenty-first century, while personal computers have become more efficient, the energy demands of server farms and data centers have surged, accounting for a notable percentage of global electricity use. Additionally, energy-intensive activities like cryptocurrency transactions and AI processing further complicate the landscape of computer energy use, raising questions about the balance between energy savings and increased consumption.
Subject Terms
Computers and energy use
Summary: The use of energy for computers continues to grow as of 2024, and experts predict that cryptocurrency and artificial intelligence (AI) will further increase the need for energy.
The emergence of computing since the mid-twentieth century has had multiple consequences for the consumption, management, and development of energy resources. Although traditional wisdom may assert that computers are essentially electronic products that consume energy, a comprehensive analysis of computing’s overall energy impact is far more difficult to calculate.
Computers draw energy when used and represent considerable energy expenditure in their creation; however, media theorists and social theorists would also point out that the rise of the “network society” is changing human behavior, with potentially profound implications for energy use. Commentators such as Lev Manovich and Manuel Castells have argued that the computerization of culture is in the process of radically changing the ways in which society is organized.
Simple examples of this can be found in the way in which computer-based satellite navigation technology can lead to mass savings (or increases) in energy use as driving patterns are changed. Similarly, the increased possibilities for video conferencing and emailing have allowed for the possibility of great savings in the energy-intensive transportation of people. At a more direct level, technologies are now emerging that allow homeowners to monitor and track their own energy use. From smart meters to smart appliances, the trend in energy markets globally is for an integration of computer network technologies with consumer appliances. In this case, it is not simply human behaviors that computing affects: The behaviors of appliances themselves start to change as they are connected to smart grids that operate to balance supply and demand. This facilitates the adoption of intermittent renewable energy technologies that previously were infeasible on energy grids that required a consistently reliable flow of power to meet variable demand. Thus, social scientists would point out that while computer technologies have the potential to change behavior and save energy, it is extremely difficult to determine the savings versus the additional costs.
Calculating devices such as the abacus date back to at least 2400 BCE and used negligible amounts of energy. Following the Industrial Revolution, the computer, like many other tools, was conceived as a mechanical device. Charles Babbage’s difference engine (1822) and subsequent analytical engine (1837) are widely held to be the first major attempts to develop what can be regarded as a computer in the modern sense. Although he never finished it, Babbage initially conceived of the difference engine as a steam-powered device, famously protesting the tedium of manual calculations with the proclamation, “I wish to God these calculations had been executed by steam.” Because Babbage’s machines were never realized, it was not until the development of electronic digital computation that the modern era of computing began in earnest and significant quantities of energy were used for computing.
Twentieth-Century Energy Use
Like their unrealized predecessors, twentieth-century computers were conceived as very large industrial devices. This meant that while they often consumed considerable quantities of energy in their operation, only a few devices were in operation relative to other industrial processes. In other words, early computers consumed a great deal of energy but were few in number, so they had a negligible impact on the power consumption of their host nations. However, the development of microchip technology led to the possibility and development of personal computers. In a reversal of the previous logic, the production of personal computers led to great savings in the energy consumption of individual machines but was followed by the exponential growth in the number of devices. Over time, this has meant that computers have come to consume an ever-increasing proportion of many nations’ energy budgets. For the duration of computing history, machines have been widely governed by Moore’s law, which states that the transistor density (and therefore the capacity of the computer’s central processing unit, or CPU) doubles every two years at no increase in cost.
Twentieth-First-Century Energy Use
While personal computers became increasingly efficient in the twenty-first century, the server farms used as the backbone of the Internet are growing in quantity and energy demand. Server farms are centralized facilities that use linked servers to process and distribute data. Modern computers also rely on data centers, which house data and IT infrastructure. According to the Frontier Group, in 2022, data centers used 240-340 terawatt-hours of electricity, which is about 1.2 percent of the world's total electricity use.
Cryptocurrency is a virtual payment system powered by encryption algorithms. It uses a great deal of electricity. For example, verifying a single Bitcoin transaction uses about as much electricity as an American home over twenty-seven days. Artificial intelligence (AI) learns information from source data. Generative AI, such as ChatGTP, is able to learn complex information. It usually takes more energy to use AI than to submit a query using a traditional search engine.
Bibliography
Cubitt, Sean, et al. “Does Cloud Computing Have a Silver Lining?” Media Culture and Society 32, no. 6 (November 2010).
Fen, Wu-Chun. Green Computing: Large-Scale Energy Efficiency. Boca Raton, FL: CRC Press, 2011.
Lovelock, James. The Revenge of Gaia: Earth’s Climate Crisis and the Fate of Humanity. New York: Basic Books, 2006.
Gordon, Whitson and Jason Cohen. "Power-Hungry PC? Here's How Much Electricity Your Computer Consumes." PC Magazine, 20 Sept. 2023, www.pcmag.com/how-to/power-hungry-pc-how-much-electricity-computer-consumes. Accessed 30 July 2024.
Rakov, Susan and Abigail Ham. "Fact File: Computing Is Using More Energy Than Ever." Frontier Group, 5 Oct. 2023, frontiergroup.org/resources/fact-file-computing-is-using-more-energy-than-ever/. Accessed 30 July 2024.
Sui, Daniel, and David Rejeski. “Environmental Impacts of the Emerging Digital Economy: The E-for-Environment E-Commerce?” Environmental Management 29, no. 2 (2002): 155–63.