Latest News

Google has agreed to reduce the power used by AI data centers in order to relieve strain on the US grid during times of high demand

Google announced on Monday that it had signed agreements with two U.S. utilities to reduce the power consumed by its AI data centers during periods of high demand on the grid. AI is a power-intensive technology, and power supply cannot keep up.

The country's utilities have been inundated by requests for electricity to power Big Tech's AI-based data centers. In some areas, the demand has exceeded total available supply.

This power crunch has raised concerns over soaring bills and blackouts for homes and businesses.

The technology industry has also been hampered in its expansion of AI. This requires huge amounts of energy - and fast.

Google's agreement with Indiana Michigan Power, Tennessee Power Authority and other electric utilities would require the tech giant to reduce power consumption at its data centers in order to make room on the grid.

These are the first formal agreements Google has made in its demand-response program with utilities, to temporarily reduce its machine learning workloads. Machine learning is a subset artificial intelligence.

Google stated in a recent blog that the technology allows data centers and other large loads to be connected more quickly. It also reduces the need for new power and transmission plants and helps grid operators manage power grids more efficiently and effectively.

Other energy-intensive industries, such as heavy manufacturing and cryptocurrency mining, have used demand-response programmes. Businesses receive either a payment or a reduction in their power bill.

Details of the commercial agreements between Google and the utilities are not known.

Demand-response arrangements are currently only applicable to a small part of the demand on the grid. However, as the supply of electricity in the United States tightens, these agreements may become more common. (Reporting and editing by Bernadette baum)

(source: Reuters)