Generative AI tools, while innovative, come with significant energy costs that may soon impact consumers. Training Large Language Models (LLMs) requires immense power, with thousands of GPUs running continuously, leading to projections that data centers could account for 20% of global electricity use by 2030-2035. This increased demand strains the already vulnerable energy infrastructure, causing rising energy bills—up to 20% according to PJM Interconnection. As AI advancements like OpenAI’s GPT-5 approach, energy consumption from data centers is set to rise sharply, contributing to nearly half of the US’s electricity demand growth by 2030. Coupled with challenges such as diminishing reliability of older energy plants and restrictive policies on fossil fuels, the reliance on renewable resources becomes critical. However, recent political decisions threaten the development of solar and wind energy, complicating the situation further. Consumers may soon face ballooning energy costs driven by the insatiable appetite of AI.
Source link
Share
Read more