Austin, Texas – The revolution of artificial intelligence develops quickly in Texas, providing with it an increasing demand for a precious resource: water. While the AI promises transformative breakthroughs in the fields of manufacturing medicine, massive data centers supplying this technology require millions of gallons to cool their computers, a drain largely not listed on the already stressed water supplies.
Artificial intelligence is a thirsty technology. A recent study, for example, stressed that the creation of an email of 100 words using chatgpt can consume the equivalent of a bottle of water. With millions by taking advantage of AI daily and the boom starting just, the cumulative impact is amazing.
“This is a lot of water to consume,” said Mohammed Islam, assistant professor in the IT and engineering department at the University of Texas in Arlington. He described AI as “very thirsty”, noting that the training model for Chatgpt-4, for example, consumed of water equivalent to that used by 30,000 households every day for 100 days in a row.
An important part of the water used to cool these vast servers of servers evaporates. “… and we cannot recover it,” said Islam, stressing the permanent loss of this resource.
The challenge is not limited to the direct use of water for cooling. The enormous energy needs of AI still complicate the problem of water. Peter Voss, CEO of AIGO and an AGS pioneer (artificial general intelligence), underlined the necessary energy scale: “We heard that the island of three miles would be recommended only for Microsoft to feed the Microsoft data center.”
ALSO:“ It is devastating ” warns Austin PBS
However, even energy solutions and nuclear energy have a significant water footprint. “They consume a lot of water,” said Professor Islam. “So this means that you are going to get rid of the energy problem, but you will create a much more important problem in terms of water.”
Despite the intimidating scale of the problem, there is an increasing recognition of these challenges within the industry. Texas data centers are starting to explore and implement innovative solutions. Some pass the IT workloads to cooler night hours to reduce cooling requests, while others install rainwater harvesting systems.
Professor Islam refers to this evolving and more conscious approach as “calculation of water waste”. The principle is simple: “You do computer, but you must be wise on your water consumption.”
Although AI’s abilities are undeniably impressive – “it is incredible what it can do. It can do these summaries and offer a poem. You can have a conversation”, as your environmental cost described it is becoming more and more obvious.
The critical question remains: can technological innovation follow the rhythm of the climbing of AI requests, or the insatiable thirst for the computation power will exhaust the vital water resources of Texas? Perhaps, one day, the AI itself will help find a lasting response.
