Google Turns into "Electric Tiger" as AI Data Centers Strain US Power Grid

Phoenix Network Technology News Beijing Time August 5, according to Reuters, as the energy-intensive AI's power demand exceeds power supply, Google Company announced on Monday that it has signed a contract with two US public utility companies to reduce its AI data center's power consumption when electricity demand surges.
Currently, large technology companies' AI data centers have already put tremendous pressure on the power grid in the United States. In some areas, these power demands have even exceeded existing total power supply capacity.
Power shortages have raised concerns about ordinary families and businesses facing skyrocketing electricity bills and risks of power outages. At the same time, this has also posed challenges to the tech industry's rapid expansion of AI business, asAI development urgently requires massive power support.
Google has reached an agreement with Indiana-Michigan Power Company and Tennessee Valley Authority. When the power company requests it, Google will reduce its data center's power consumption to create space for the grid. This is the first batch of formal agreements that Google has made with power companies based on demand response plans, which will temporarily reduce the running load of machine learning workloads.
Google stated in a blog post: "This approach can allow large-scale power loads like data centers to quickly enter the grid, reducing the need for building new power stations and transmission lines, and helping grid operators manage power systems more efficiently."
Demand response plans have traditionally been used in other high-energy industries, such as heavy industry manufacturing or cryptocurrency mining. In return, companies can typically receive subsidies or discounts on electricity bills. This type of demand response plan involving AI data center activities is relatively new, and the commercial arrangements between Google and power companies are not yet clear.
Although this type of agreement currently only applies to a small part of the power grid's power demand, as the US power supply becomes increasingly tight, such arrangements may become more prevalent. (Author/Araxi)