AI Is Running Google’s Data Center Cooling Systems

 

Tech companies spend millions of dollars each year to keep their data centers cool. The cooling systems needed for facilities of this nature consume large amounts of electricity, which can also have a harmful impact on the environment. A growing number of data centers are moving to green energy. In order to help protect the planet. However, no matter the source of the energy, it is expensive to use so much of it on a daily basis.

Google has been making a concerted effort to optimize their cooling systems. Trying to keep temperatures where they need to be for optimum equipment performance. While also using as little energy as possible. For years they have been gathering data and using it to determine what actions they should take to ensure they were running efficiently. Recently, they made the bold decision to stop letting their engineers manage the data center cooling systems. Instead, they turned over control to their artificial intelligence platform.

Trusting the AI To Do Its Job

Google’s DeepMind system is their advanced AI platform that is used for a wide range of different activities. In a blog post written by the DeepMind team they said, “Even minor improvements would provide significant energy savings and reduce Co2 emissions to help combat climate change.”

DeepMind has been taking snapshots of the data center environment every five minutes,. They are using that information to determine what the best actions would be to keep the facility cool. Google has granted DeepMind the ability to make its proposed changes without direct human involvement.

So far, the AI system has done an extraordinary job. In just a few months, Google reports that DeepMind has produced an energy savings of around 30%. It is also continuing to come up with new ‘ideas’ and technologies to help further save energy.

A Model for the Future

While it is still fairly early on in this process, the results speak for themselves. A 30% reduction in energy use will translate to millions of dollars in savings. Google will undoubtedly implement this system in all their data centers in the future.

According to the company, “In the long term, we think there’s potential to apply this technology in other industrial settings, and help tackle climate change on an even grander scale.”

While data centers are a huge consumer of electricity, they aren’t the only ones using the energy. Applying DeepMind to other industries such as manufacturing, large buildings, and possibly even homes could help to save billions of dollars each year. Dramatically reducing Co2 emissions. With such significant reductions in energy consumption, it may help to make green energy technologies more viable as a primary electrical generation system.

AI technology