Challenges with clean energy in the realm of Generative AI
During a visit to Sam Altman’s residence in San Francisco, a journalist noticed two pairs of pink high-tops on a shelf. One pair represented Mr. Altman’s machine-learning startup, OpenAI, while the other showcased Helion, a nuclear-fusion company he supports. Mr. Altman’s passion for these technologies goes beyond mere fashion statements. He envisions a future where the costs of intelligence and energy decrease in harmony, describing it as a “long and beautiful exponential curve.”
However, the reality is far from idyllic. Conversations with utility providers and data-center operators reveal a pressing energy dilemma that intersects with the AI revolution, electrification efforts, and climate change mitigation. The energy-intensive nature of “generative” AI, like OpenAI’s ChatGPT, poses a significant challenge to an already strained global energy system. The availability of sufficient clean energy to meet growing demands remains uncertain.
On the surface, the solution appears straightforward. Data centers, utilized by tech giants such as Alphabet, Amazon, and Microsoft for cloud services, have historically represented a small fraction of global energy consumption. These companies have made strides in enhancing energy efficiency and investing in clean energy initiatives to offset their environmental impact. In the United States, electricity providers serving these tech giants are eager to collaborate, having faced stagnant electricity demand for years and seeking new energy sources…
2024-04-11 09:42:06
Link from www.economist.com