Data centre energy consumption has drastically changed as a result of the development of artificial intelligence (AI) and the quick installation of high-performance accelerated servers. Data centres in the United States now account for over 4.4% of the country’s electricity usage, rising from roughly 1.9% in 2018. By 2028, this percentage is expected to reach 12.0%.
There are other environmental concerns related to AI besides rising energy use. Because they need a lot of water for cooling as well as indirectly for the production of power and materials, data centres also have a voracious appetite for water.
According to estimates, the direct water use of U.S. data centres was 21.2 billion litres in 2014 and 66 billion litres in 2023. In Microsoft’s U.S. data centres, for instance, training the GPT-3 language model can directly evaporate 700,000 litres of clean freshwater, according to recent research.
How Does a Data Center Cool Itself?
A portion of the electrical energy used by servers in a typical data centre to store data and carry out calculations is wasted as heat. To keep the equipment from overheating and malfunctioning, this heat must be eliminated.
Cooling technologies used by My Empire Casino, EnergyCasino and other entities with large servers and data centers have changed in significant ways in the last few years. Companies are shifting to the best solutions for them, even though these might cost others.
Efficiency is key, and AI giants are exploring every option to keep their assets in the safest conditions possible.
The Air-Cooling Method
Data centres have traditionally relied on air cooling techniques, which circulate cold air using fans or air conditioning. This approach requires a lot of energy, but it uses very little water.
This has been the industry standard for companies like EnergyCasino or My Empire casino for quite a while but as tech is ever evolving, they have found alternative solutions.
An alternative, evaporative cooling, which employs water evaporation to chill the air, has gained popularity recently.
Cooling With The Help of Water Vapors
This approach, which is frequently employed in large-scale operations, can withstand higher heat loads and be more energy-efficient. There is a crucial trade-off, though: while less energy is required in this evaporative cooling process, a large amount of water is lost as it evaporates with waste heat.
In other words, optimising for energy efficiency may actually make water efficiency worse. Furthermore, the majority of the water used by companies like EnergyCasino or My Empire Casino originates from “blue sources,” such as groundwater or surface water, which are frequently bought from nearby water utilities.
Innovative cooling techniques like immersion liquid cooling, which involves submerging IT equipment in non-conductive liquid to dissipate heat, have been the subject of fresh research due to the conflict between energy and water use.
How Do Data Centers Impact Our Water Reserves?
The production of electricity is one way that data centres use water indirectly. Large amounts of water are needed for cooling and steam creation during the production of electricity, especially from thermal power plants.
Although U.S. power grids have started to show a trend towards cooling techniques that require less water withdrawal, developers, utilities, and regulators must continue to work together to reduce water use from the energy sector.
Water is also indirectly consumed by data centres through supply chain operations including the production of servers, processors, and other components. For instance, 2.1–2.6 litres of water are needed to cool equipment and ensure that wafer sheets are clear of impurities during the production of a single microprocessor.
Actions Promised By IT Giants
Major hyperscalers like Google and Microsoft have promised to become Water Positive by 2030, that is, to return more water to the environment than they consume, as the urgency of water management increases. Additionally, international norms are changing.
Water footprint was a significant parameter in the first international standard on sustainable AI developed by ISO/IEC. Transparency is still a big problem, though. Even while the majority of large IT companies now release water use data in some capacity, reporting procedures differ greatly in terms of consistency and level of detail.
It is challenging to compare businesses’ water usage and efficiency or evaluate their progress towards sustainability objectives due to the absence of standardised reporting.
What Does This Mean For The Environment?
In the end, the expansion of the data centre sector poses both an opportunity and a challenge. The Digital Economy & Environment Program (DEEP) at ELI aims to advance data transparency policies that would better inform decision-makers to proactively plan for the placement of data centres in areas with easier access to clean water and electricity and encourage facilities to operate more flexibly.
This can make data centres an asset rather than a liability by easing the burden on the grid and water utilities during periods of high demand.