Site navigation

Google uses DeepMind AI for Autonomous Data Centre Cooling

Theo Priestley


DeepMind controls the data centre cooling system.

An initiative that started in 2016 comes full circle as the artificial intelligence is now in charge of data centre cooling and industrial control.

While Microsoft looks to undersea data centre cooling and elsewhere in Europe countries are harnessing waste heat as an energy source in itself, Google is taking a different approach by putting its DeepMind artificial intelligence in control.

In 2016 Google set out to implement the AI system in several of its data centres as a way of reducing the amount of energy the server farms consume, projecting a “15% reduction on energy used every year by the entire data centre fleet.” according to DeepMind cofounder Mustafa Suleyman.

Fast forward to 2018 and they’ve achieved much more.

AI in the Data Centre Driving Seat

In a recent post, DeepMind explains that their original thinking was simple: “even minor improvements would provide significant energy savings and reduce CO2 emissions to help combat climate change.” they claimed.

“Now we’re taking this system to the next level: instead of human-implemented recommendations, our AI system is directly controlling data centre cooling, while remaining under the expert supervision of our data centre operators. This first-of-its-kind cloud-based control system is now safely delivering energy savings in multiple Google data centres.”

Every five minutes, the AI controller pulls a snapshot of the data centre cooling system from thousands of sensors and feeds it into the DeepMind neural network, which predict how different combinations of potential actions will affect future energy consumption. The AI system then identifies which actions will minimise the energy consumption while satisfying a robust set of safety constraints.

Those actions are sent back to the data centre, where the actions are verified by the local control system and then implemented.

The system also performs a number of other tasks, outlined in the post, such as: continuous monitoring, automatic failover, human override, and continuous comms.

Over nine months, Google claims that the AI control system performance increased from a 12% improvement (the initial launch of autonomous control) to around a 30% improvement.

Google sees applications beyond the data centre, saying that 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.”

Theo Priestley

Latest News

Artificial Intelligence News
Recruitment Technology
Funding Government
HealthTech MedTech
%d bloggers like this: