Data centers and databases can be used as the most powerful amount of power as Japan by 2030

It is expected that the growth of the data center, which is largely motivated by the IQ, will increase the demand for energy in the coming years and the report this week from the International Energy Agency. New estimates How much.

It is expected that the amount of electricity needed for data centers around the world will reach more than twice the TERAWATT in 2030, which is more consumed by the entire country in Japan today, according to the report.

popularity Gen AI tools Like Openai Chatgpt And Google’s twin It has increased in the past few years. These large linguistic models and relatives require a huge amount of computing energy, which works on the high -end Graphics processing units Such a manufactured by Nvidia. This not only needs a lot of electricity to work but also generates heat – which means that more energy is needed to keep it calm. All this adds quickly.

About half of the increasing demand for electricity in the United States is expected to be 2030 for data centers. Processing data is expected to need more electricity from manufacturing all “intensity goods” together-aluminum, steel, cement and chemicals.

“Artificial intelligence is one of the largest stories in the energy world today-but so far, politics and market makers were lacking tools to understand the widespread impacts,” IEA CEO said in a press statement.

The use of artificial intelligence energy worries policy makers

The expected demand for energy on artificial intelligence has been documented well and in the minds of policy makers and experts for years. Only this week, the Energy and Trade Committee in the House of Representatives was held hearing On how to provide sufficient energy for data centers where Congress members try to understand the issue.

Experts said the United States electricity system will need improvements and changes to meet the increasing demand after decades of relatively flat consumption. One of the fields of focus in the committee was to provide “Baseload Energy” – a fixed electricity supply throughout the day. This desire to consistency has pushed questions from the members about whether renewable energy sources such as solar energy and wind will work or whether the increasing demand should be met by new power plants that burn fossil fuels such as natural gas. Burning these fossil fuels is a major engine of climate change.

There is also an issue whether the current US energy network can handle the necessary demand and additions.

“Even without expecting an increase in the demand for electricity, the American energy network needs modernization investments,” Melissa Lot, a professor of climate school at Colombia University. “The latest expectations for the rapid increasing demand for energy make these investments more urgent and necessary.”

The timing of the availability of electricity, given the high demand at times such as heat and daily waves down Solar generationIt is a problem that can be dealt with in different ways. Lot pointed to energy efficiency efforts, such as Power star devicesAnd efforts to reduce the demand known as Virtual power plants. These can reduce energy demand and provide peaks and valleys throughout the day.

Can Amnesty International solve its energy problem?

The IEA report indicates that the IEA IQ, over time, helps to solve the problem of energy demand. If AI accelerates scientific improvements and technology, this may lead to the best panels and solar batteries, for example. “Energy innovation challenges are characterized by the types of problems that Amnesty International is good in a solution,” said the IEA report.

But another solution may be how to use the energy intelligence centers. A A report earlier this year Researchers at Duke University have suggested that artificial intelligence databases can easily be turned off and adapt to the needs of the electrical system, allowing network operators to absorb growth more easily.

The researchers wrote: “This analysis should not be interpreted as indicating that the United States can meet the requirements of near and medium electricity in the full term without building a new peak capacity or expanding the network.” “Instead, highlights that flexible pregnancy strategies can help take advantage of the current head hall to quickly combine the loads, reduce the cost of expansion of the capacity, and enable more focus on the highest value of investments in the electric power system.”