Barclays analysts highlight growing energy demands for AI, with NVIDIA playing a crucial role

From Investing.com: 2024-07-14 05:01:02

Barclays analysts highlighted the energy demands associated with AI advancements and NVIDIA’s crucial role in the landscape. By 2030, data centers may consume over 9% of current U.S. electricity demand due to AI applications, with large language models growing 3.5 times larger annually. Despite improving energy efficiency in newer GPU generations, AI models’ complexity drives higher computational power needs.

The growing energy demand brought by AI technologies is projected to require 14.5 gigawatts of power, equivalent to about 110 terawatt-hours of energy to power around 8 million GPUs. With 70% of these GPUs anticipated to be in the U.S. by 2027, the nation faces over 10 gigawatts and 75 TWh of AI power demand in three years.

Barclays analysts foresee NVIDIA’s market cap pointing to a rise in AI power deployment, fueled by ongoing GPU development. With data centers relying heavily on grid electricity, managing peak power demands is critical for ensuring consistent operations. CEO of OpenAI, Sam Altman, highlighted the escalating energy needs of AI technologies at the Davos World Economic Forum.



Read more at Investing.com: NVIDIA consensus suggests a lot of imminent AI energy is needed: Barclays By Investing.com