Tesla predicts that the power demand in 2045 will double compared to the current level. As the energy consumption of AI servers surges threefold, power shortage may hinder the vigorous development of this industry.
In recent days, billionaire Tesla has frequently expressed his concerns about power supply, predicting that within two years, the shortage will shift from "shortage of silicon" to "shortage of electricity," which could hinder the development of AI.
On Tuesday local time, Tesla addressed energy industry executives at a meeting held by PG&E, one of the largest utility companies in the United States, stating:
It is indeed necessary to expedite the timetable for (power supply) projects and have a sense of urgency. My biggest concern is the lack of urgency.
Tesla also stated:
If power demand remains relatively stable (which has been the case in the United States for some time) and there are no significant changes, then the projects can be delayed for a period of time.
But in rapidly changing situations where power demand continues to increase, we must take faster action.
This is the third major energy event Tesla has attended in the past 12 months.
Last month, at an energy conference held in Austin, Tesla predicted that with the increase in electric vehicle consumption, power demand in 2045 will be twice as high as it is currently. Tesla forecasts that by 2030, half of the new cars sold globally will be electric vehicles.
He emphasized:
I cannot overemphasize this enough, we need more power. No matter how much power you think we need, the actual need is even greater.
Energy executives and consultants also believe that power demand will increase, but not as dramatically as Tesla predicts.
PG&E expects power demand to increase by 70% over the next 20 years, which the California-based company points out is unprecedented. Similarly, McKinsey predicts that by 2050, U.S. power demand will double.
PG&E CEO Patti Poppe seems to agree with Musk's warning but did not immediately overturn the company's forecast. She told Tesla:
We will definitely take note of this. I will be the last person to doubt your future predictions.
Tesla also warned that power shortages could have serious consequences, similar to the recent chip shortage that has affected the technology and automotive industries. And this time, the shortage of electricity could hinder the development of AI.
At an event earlier this month, Tesla stated:
My prediction is that we will shift from... the extreme shortage of silicon today to... a shortage of electricity within two years. That's roughly the trend of things.
A previous article by Wall Street CN also mentioned that computing power is the foundation of AI development, and the bottleneck of computing power ultimately lies in electricity. As the competition for AI arms race intensifies, the competition for electricity becomes increasingly important.
According to a recent report published by Digital Information World, the energy consumption generated by training AI models in data centers will be three times that of regular cloud work. It is estimated that by 2030, the power demand of U.S. data centers will grow at an annual rate of approximately 10%.
One of Tesla's solutions is to run power plants around the clock and store unused energy during peak hours in battery packs for future use, thereby optimizing the power grid more effectively. Tesla recently stated:
I'm not sure if it will double the power supply... but the total energy output will increase by at least 50% to 100%.