Khaberni - As Generative AI heralds the opening of digital future horizons, environmental experts and scientists have issued a stark warning that the massive and rapid expansion of generative artificial intelligence is placing an unprecedented burden on the planet's resources, threatening to intensify the global climate crisis and water sustainability, and depleting the energy of 1.5 million households. A deep analysis reveals that each interaction with artificial intelligence systems like ChatGPT requires a worrying energy and water footprint, raising a crucial question: Can we push the limits of technology without pushing the planet towards environmental turmoil?
Concerns begin with the immense energy consumption necessary to operate complex artificial intelligence models and the massive data centers that support them. According to researchers, a single query on the latest versions of ChatGPT-5 may require between 18 to 20 watt-hours. When this figure is multiplied by the daily volume of interactions (estimated at 2.5 billion queries per day), the energy consumption of generative artificial intelligence equals the consumption of 1.5 million American households!
It's no longer just speculation; the consumption of electricity by data centers in France has risen from 2% to 10% of the national grid in just three years, prompting an urgent question about the sustainability of this growth.
Water Cooling
It's not just about energy, but also the lifeblood: water. The need to cool servers that continuously heat up has led data centers to use water-based cooling systems. An average data center consumes up to 396,000 gallons of water daily, equivalent to the daily needs of 13,000 households.
The situation is particularly critical in the microchip industry, where TSMC operations in Taiwan consume about 150,000 tons of water daily, accounting for more than 10% of the island's total water consumption!
As water scarcity has led countries like the Netherlands to decide to stop building new data centers, confirming that water has become as vital a resource as electricity in this technology sector.
Metal Race
The global race for advanced Graphics Processing Units (GPUs) from companies like Nvidia is speeding up, necessitating the extraction of additional rare metals and building more servers.
The Transition Project warns that the metals used in AI chips are the same ones essential for renewable energy technologies, such as batteries and wind turbines. This creates a direct competition between technological growth and the needs of the green transition. In a desperate attempt to meet demand, major technology companies are turning to nuclear energy, at a time when global electricity production is already struggling to keep up with the massive demand.
