The end of AI is photovoltaics and energy storage

According to Polaris Energy Storage Network, NVIDIA CEO Jen-Hsun Huang pointed out in a public speech that the future development of AI is closely linked to state-of-the-art energy storage. Huang Renxun made it clear in his speech: "The end of AI is photovoltaics and energy storage! We can't just think about computing power. If we only think about computers, we need to burn the energy of 14 earths."

As early as February 27, someone was telling the story of "energy storage and AI" in a social video, citing the so-called "Huang Renxun's speech". At the same time, OpenAI founder Sam Altman also expressed similar views. Altman said, "Future AI technology depends on energy, and we need more photovoltaics and energy storage."

ChatGPT consumes over 500,000 kilowatt hours of electricity per day

The rapid development of artificial intelligence technology has brought huge computing power and energy consumption. According to the New Yorker magazine, ChatGPT consumes more than 500,000 kilowatt hours of electricity per day, equivalent to 17,000 American households, in order to respond to approximately 200 million requests from users.

Recently, at the Bosch Internet 2024 Conference, Musk was interviewed remotely by the CEO and Chairman of Bosch. Musk mentioned that the development speed of artificial intelligence is unprecedented. It seems that every six months, the computing power will increase by 10 times, far exceeding the rate of Moore's Law doubling every 18 months. He predicts that within two years, the year will change from a "silicon shortage" to a "power shortage", which may hinder the development of AI.

"Currently, the growth of computing power is facing a bottleneck. It is obvious that there will be a shortage of transformers next, followed by electricity. By 2025, we will not have enough electricity to run all the chips." Musk said.

In fact, Musk has been worried about power shortages for more than a day. He has publicly emphasized the urgency of solving the growing power demand many times before. He believes the project needs to be accelerated to keep up with growing power demand.

The Boston Consulting Group has analyzed that it is expected that around 2030, the power consumption of U.S. data centers will triple compared with 2022, equivalent to 7.5% of the total power demand, which will significantly increase social electricity consumption.