According to the report, Microsoft’s global water consumption increased by 34% between 2021 and 2022 as a result of significant investments in artificial intelligence.

In training GPT-3 at Microsoft data centers, it is estimated that 700,000 liters or about 185,000 gallons of fresh water were used. That’s enough water to fill the cooling tower of a nuclear reactor, according to Gizmodo, and the same amount used to produce 370 BMW cars or 320 Tesla cars, according to the study.

Using these calculations, it was determined that ChatGPT would need 500 ml of water for every 5 to 50 questions it answered.

In a very short period of time, generative AI assistants like ChatGPT, Bard, Dall-E and others have managed to become a part of not only the professional lives of many, but have also crept their way into their personal lives. And the touchy subject of whether AI will replace the need for humans still holds a prominent place today.

Small businesses developing advances in AI are also attracting significant investment from IT giants.

The company says the new 12nm-class DRAM consumes 23% less power than the previous generation. This offers IT companies a reduction in power consumption and carbon footprint in their servers and data centers. Samsung explains that such efficiency gains are made possible by a new material with a high dielectric constant that helps increase the capacity of the cells.

“With differentiated technology, Samsung’s industry-leading 12nm DDR5 DRAM delivers exceptional performance and energy efficiency. Our newest DRAM reflects our continued commitment to leading the DRAM market, not only with high-performance, high-capacity products that meet the demand of the large-scale computing market, but also by bringing to market next-generation solutions that support greater performance.”

commented the company.

Samsung’s 16Gb DDR5 DRAM memory chips operate at a maximum speed of 7.2 Gbps, which roughly means processing two 30GB UHD movies in one second. The manufacturer completed the compatibility evaluation with AMD last December. It is currently working with other technology companies to support an ever-growing list of next-generation computing applications and services.