The era of generative artificial intelligence (AI) is changing the world, both figuratively and literally. The energy and water consumption of the major technology companies, the main developers of this technology, as well as their carbon emissions, have skyrocketed in recent years. And projections show that the trend will not change. Although no company officially says that this increase is due to the emergence of AI, the numbers show a significant jump in 2022, the year in which OpenAI launched ChatGPT and inaugurated the generative AI race.
Seven of the 10 largest companies in the world By market capitalization, the largest companies in the world are technology companies, which gives an idea of the importance of the sector. Large industries have large resource needs. Nevertheless, the data show a significant jump. Shaolei Ren, associate professor of electrical and computer engineering at the University of California, Riverside and an AI sustainability specialist, believes that it is not a risky inference to infer that AI is responsible for this escalation in pollution and resource consumption. The increase in the last two years, he argues, has been very large and coincides in time with a strong investment in generative AI and other AI-related services.
More energy
The latest figures available from Google and Microsoft, the main developers of this technology, reflect strong growth for the second consecutive year in the three key figures. Google, responsible for the Gemini model, has just reported in its environmental memory an increase of 16.2% in energy consumption in 2023 compared to the previous year. For its part, Microsoft, owner of Copilot and which has lent its infrastructure to OpenAI to develop all versions of ChatGPT and the Dall-E image generator, has registered a growth of 28.7%, as reflected in its annual sustainability reportThe company founded by Bill Gates has doubled its energy needs between 2020 and 2023, going from 11.2 million megawatt-hours (MWh) to 24 million MWh. The same has happened at Google, with a 67% increase in this period.
GPU processors, the ones used to train AI models, are much more powerful than the CPUs that have been predominant in data centers until now, and therefore consume more energy (up to 10 times more). Training large language models requires tens of thousands of GPUs operating day and night for weeks or months. The most advanced models are periodically retrained to incorporate updated data and every time a user types a new language, the GPUs are retrained to ensure that the data is updated. prompt (an order) on your mobile or computer, the response is computed in a data center. All this activity has stretched energy demand, to the point that some companies, aware that the trend will continue to rise for some time, are considering developing small nuclear power plants to ensure a sufficient and stable supply.
More water
The data centers where AI (and all digital activity) is operated are large industrial warehouses populated with rows and rows of racksseveral processors arranged in the form of a cabinet or refrigerator. All these processors and servers, which store our data and run online programs, work day and night. This activity emits a lot of heat; if the temperature is not controlled, the equipment can break down.
Data centre cooling uses water, which is sprayed to cool the environment. Water consumption has also seen increases of 13.8% and 21% respectively in 2023, figures similar to those of the previous year. Microsoft, for example, has reported using almost 13 billion litres of water. More than half of this volume (about 8 billion litres) was evaporated or consumed, so it could not be reused. Google, on the other hand, needed less water, about 8.6 billion litres, but only returned 26.6% of this amount to the system.
These figures, however, do not give a complete picture of the real consumption of AI developers. Companies only provide data on the water they use to cool their data centres, but do not include in their reports either the water used to generate the electricity they consume or the water used in the supply chain of the products (mainly in the production of chips and other hardware), as is the case, for example, with carbon emissions.
“Companies intentionally hide that data,” Ren says. “That’s why it’s very revealing that Apple accidentally said in its latest environmental performance report “Its indirect water consumption due to the supply chain accounts for 99% of its total water footprint.” Based on Apple’s direct water consumption data, Ren concludes, that would imply that Apple’s actual consumption in 2023 was at least 300 billion liters. “That volume of water is enough to irrigate 0.1% of the wheat harvested per year worldwide,” he illustrates.
More emissions
As for carbon emissions, Google’s have grown by 13% and Microsoft’s by 3.8% in the last year. The increase is 67% and 40%, respectively, if we look at the last four years.
According to Ren, most of the pollution emitted by these companies is related to their supply chain. “The main driver of the increase in global carbon emissions is the emissions associated with the manufacturing of AI chips and the construction of data centers,” he explains.
While the energy efficiency of the hardware used to develop and run AI has increased in recent years and will continue to do so in the coming years, the researcher notes, “it is very unlikely that embodied carbon will decrease in the short term due to the increased demand for hardware of AI”.
The race for AI
From Google to Microsoft, from Meta to Amazon (which have not yet published their environmental reports for this year) to Apple, all the major technology companies are immersed in programmes to improve their carbon emissions records and reduce the amount of water used. The goal of many of them is to reach 2030 with a very low environmental footprint.
Ren and his colleagues explain in an article that has just been accepted in the journal Communication of the ACMa benchmark in the IT sector, provides some projections based on current consumption and industry trends. Global demand for AI will be responsible for the use of between 4.2 and 6.6 trillion litres by 2027, the equivalent of half of the water used each year in the United Kingdom. In the same year, energy demand for AI will be between 85 and 134 TWh. In comparative terms, global battery production in 2023 was around one terawatt per hour (1 TWh).
“If we only look at emissions from their direct energy and water consumption, they can achieve zero emissions and zero water consumption by 2030, maybe even sooner,” Ren concludes. “But if we look at their actual footprint, it is highly unlikely that they will achieve neutrality by 2030.”
You can follow THE COUNTRY Technology in Facebook and X or sign up here to receive our weekly newsletter.
