ChatGPT consumes over 500000 kilowatt hours of electricity per day, which is 17000 times the electricity consumption of American households. The daily electricity bill is 600000 yuan

The phenomenon of AI burning electricity is actually a warning given by Nvidia founder Huang Renxun and OpenAI founder Sam Ultraman.

According to a report from Financial Street, Huang Renxun stated in a public speech that the development of AI is actually closely related to energy storage. Huang Renxun said that people should not only focus on the computing power of AI, but also need to consider energy consumption more comprehensively.

Huang Renxun clearly warned in his speech that the end of AI lies in photovoltaics and energy storage. If we only consider computers, humans will need to burn up the energy of 14 Earths!

In addition, Ultraman actually put forward a similar viewpoint. Ultraman stated that the future AI industry will be highly dependent on energy, especially photovoltaic and energy storage technologies, and we need more photovoltaic and energy storage.

In addition to Huang Renxun and Ultraman, Musk has also warned that the United States will face power shortages in the future.

At the Bosch Connected World 2024 conference, Musk stated that the development speed of artificial intelligence is unprecedented, and computing power increases tenfold every six months.

Subsequently, Bosch CEO Stefan Hartung asked Musk, “How long will this high growth rate of computing power last?”?

Musk replied, it is obvious that this speed cannot continue indefinitely, otherwise the universe would not be able to accommodate it. Moreover, the growth of computing power has actually reached a bottleneck period.

Musk mentioned that he had predicted a shortage of chips a year ago. Subsequently, Musk once again made a new prophecy: the United States will experience a shortage of transformers, followed by a shortage of electricity. By next year, the United States will not have enough electricity to run all its chips.

It should be pointed out that in fact, the problem of electricity shortage in the United States is not unfounded by Musk. The current manufacturing scale in the United States is actually very limited, and transformers rely entirely on imports, with China and India being the main exporters of transformers.

At present, transformers in the United States are generally aging, and about 70% of transformers have been in use for over 25 years, while the service life of transformers is generally only 40 years. In addition, hurricane weather in the United States is becoming increasingly frequent, which can lead to damage to old transformers. Furthermore, the United States is currently developing new energy, but the installation of new energy has increased grid volatility, and transformers in the United States need to be replaced immediately.

According to data released by the US Energy Information Administration, electricity generation in the United States did not increase last year, but instead decreased by 0.9% year-on-year. Overall, there was no significant increase in new installed capacity.

Musk once said that the United States needs to triple its current electricity generation in order to maintain a balanced power process. Otherwise, the country’s electricity will not be able to meet the rapidly developing demand for AI, and there will be a nationwide electricity shortage.

In fact, OpenAI founder Ultraman already has his own plans in the field of electricity. He has invested in fusion enterprise Helion Energy as an individual, and last year Helion Energy signed a power purchase agreement with Microsoft that will only start in 2028.

From this, it can be seen that in fact