April 23, 2024

News Collective

Complete New Zealand News World

The five keys to reducing the environmental impact of data

The five keys to reducing the environmental impact of data

Schroders says locating data centers in cold climates reduces emissions

the artificial intelligence It promises to change the world we live in. From a technological point of view, there’s no doubt about that. But it is necessary to take into account the mark you will leave behind environment. The direct impact of artificial intelligence is reMainly related to the physical infrastructure it requires, which includes ce

the artificial intelligence It promises to change the world we live in. From a technological point of view, there’s no doubt about that. But it is necessary to take into account the mark you will leave behind environment. The direct impact of artificial intelligence is reIt is mainly related to the physical infrastructure it requires, which includes data centres, processors and other specialized computing equipment, according to a study by Samuel Thomas, a sustainable investment analyst at Schroders.

The life cycle of AI computing is divided into four phases: production, transportation, operations, and end-of-life phases. The most significant emissions – between 70% and 80% – come from the operational phase, and specifically, from two aspects, energy and water. Regarding the first, it is estimated that between 1% and 1.5% of global electricity demand comes from data centers, which equates to about 220-320 TWh. secondly, Water footprint The number of cores is determined by the amount of water consumed to generate electricity and the amount of water consumed for cooling.

The good news is that, according to the Schroders analyst, there are formulas to reduce the impact:

See also  Yellen says the Fed can lower inflation without causing a recession

1- Encouraging its use Renewable energy To operate data centers.

2-Using and adjusting Pre-trained modelsAI teams save time and significant amounts of energy generation models.

3-Practices Energy efficiency: One way is through server virtualization, which allows you to run multiple virtual servers on a single physical server.

4- Systems cooling Which improves efficiency and reduces costs.

5-location: If the centers are built in places with abundant renewable energy or in cooler climates, significant emissions savings could be achieved.