Updated 5 August 2025 at 10:38 IST
Google to Turn Down the Power of Its AI During Heat Waves
Google's data centers are what make many of its services work, like search, picture production, and smart assistant responses. But these brains get quite heated.
Google is doing something to help local power systems deal with the record-breaking heat waves that are hitting some parts of the US. The company has made deals with Indiana Michigan Power and the Tennessee Valley Authority to use less AI-related energy when the grid is in danger of going down or when demand is very high.
Google's data centers are what make many of its services work, like search, picture production, and smart assistant responses. But these brains get quite heated. In a literal sense. AI workloads need a lot of computational power, which uses a lot of electricity.
When summer temperatures go over 100°F (38°C), the power grid gets heated up. To control this, Google will now move non-essential AI tasks away from areas where the power system is already under stress to keep things from getting worse. This could imply stopping some processes or relocating them to data centers in other states where power is easier to get.
Google has tried to be more energy-efficient before. Google has consistently prioritised power efficiency and sustainability across its operations, particularly in its data centres and AI infrastructure. It developed custom Tensor Processing Units (TPUs) to run AI workloads more efficiently than traditional chips and has set an ambitious goal to operate entirely on carbon-free energy 24/7 by 2030.
To further reduce energy use, Google employs AI to optimise data centre cooling, achieving up to 40% energy savings in some cases, and maintains one of the industry’s lowest Power Usage Effectiveness (PUE) ratings at around 1.1. Additionally, it uses carbon-aware computing to shift workloads to times and locations where cleaner energy is more available.
In 2024, Google developed a new AI-powered tool to help cities tackle extreme heat by identifying their most vulnerable neighbourhoods. Using advanced object detection and models that analyse factors like the amount of green space and the reflectiveness of rooftops, the tool provides detailed insights at the neighbourhood level. This allows city planners and local governments to understand where heat is most intense and where cooling efforts, such as adding trees or reflective materials, would have the greatest impact.
Published By : Priya Pathak
Published On: 5 August 2025 at 10:38 IST