According to the International Energy Agency (IEA) report “Key Questions on Energy and Artificial Intelligence”, these investments are expected to increase by 75 percent this year.
The report said that the total investments of just five companies operating in the technology sector exceeded investments in global oil and natural gas production.
On the other hand, although data center projects are increasing rapidly in many regions, it has been noted that not all projects are likely to be implemented.
According to the IEA's satellite-based monitoring data, the capacity of data centers, particularly those focused on artificial intelligence, has more than tripled in the last 18 months.
While the rapid development of artificial intelligence capabilities is having transformative impacts on economic growth, innovation and competitiveness, it is also driving significant changes in existing industries and workforce structure.
Energy efficiency increases through artificial intelligence
According to the report, the energy efficiency of artificial intelligence has improved rapidly in recent years.
Thanks to the further development of software and hardware, the energy consumption per artificial intelligence task decreases significantly every year.
It is stated that simple text-based queries using artificial intelligence use less power than a television running for the same period of time.
According to the IEA, if all traditional Internet searches were replaced by simple queries using artificial intelligence, annual electricity consumption would remain below 4 terawatt hours, which would be less than 1 percent of total data center consumption.
However, more complex AI applications such as video production, reasoning and autonomous tasks have higher energy consumption.
It is stated that such applications can consume hundreds or even thousands of times more energy per query compared to simple text operations.
The report found that the main factors determining the energy needs of artificial intelligence are productivity improvement, widespread use and the development of more energy-intensive applications.
The demand for electricity will increase rapidly
According to the IEA, data center electricity consumption was 485 terawatt hours last year. This consumption is expected to double to around 950 terawatt hours by 2030.
It is estimated that this increase corresponds to around 3 percent of global electricity demand.
Electricity consumption by artificial intelligence-focused data centers is expected to triple during this period and rapidly increase its share of total consumption.
The report said that the elimination of bottlenecks in energy equipment and chip production, as well as the widespread use of energy-intensive areas of artificial intelligence, could lead to a larger increase in demand after 2030.
Bir yanıt yazın