In conditions of global growth of artificial intelligence the main problem becomes the lack of electricity. According to representatives of the largest technological companies, including billionaire Elon Musk, such a problem already limits the development of advanced technologies. Previously, AI suffered from microcircuit deficiency, now the main barrier is power supply.
The head of Amazon, Andy Jesse, also confirmed that current energy reserves are not enough for the functioning of new generative services. In this regard, companies like Amazon, Microsoft, and Alphabet (Google’s maternal company) invest billions of dollars in the infrastructure and construction of data centers, the planning and construction of which takes years.
Joint capital costs of Microsoft, Amazon, and Alphabet for cloud technologies. Forecasts for 2024 and 2025.
However, the construction of such objects requires not only time, but also places where it is still possible to increase energy capacities. In addition, access to electricity is also important, since problems with the power grid can become a serious obstacle to the development of AI.
Executive Vice-President of the Schneider Electric data center, Pankaj Charma, noted that the demand for the data center has always been high, but now it has reached a historical maximum. It is expected that by 2030 significantly more capacities will be required than currently available.
In addition, increased interest in the data centers enhances concern about environmental impact. Many countries seek to switch to renewable energy sources and electrification in response to climate change. For example, Amazon is actively discussing this issue with American officials.
By forecasts from the International Energy Agency, the consumption of electricity with data centers will double by 2026, reaching 1000 terawatt hours, which is comparable to the annual consumption of Japan. Such indicators set before the world community the task of improving efficiency and regulation to contain energy consumption.
Some regions are already faced with problems in energy consumption with an increase in demand for AI. For example, the large American energy company Dominion Energy has temporarily suspended the connection of new data centers until it solves the problem of growing electricity consumption. In the light of such problems, organizations are looking