The company uses the DeepMind system to reduce PUE data center, which controls 120 variables in data centers to reduce energy consumption. The artificial intelligence system Google DeepMind has been used by the company to achieve significant reductions in energy consumption in some of their data centers.
The subsidiary of artificial intelligence Google revealed that savings translates into a 15% PUE data center improvement. The co-founder of DeepMind, Demis Hassabis, told Bloomberg that DeepMind took control of Google data centers parts to reduce energy consumption by manipulating cloud hosting server and computer equipment such as cooling systems, using their deep learning algorithms.
Deep Mind “controls about 120 variables in data centers. The fans and cooling systems, in addition to windows and other things,” Hassabis said. “They were very surprised.”
Google used 4,402,836 MWh of electricity in 2014, and a significant amount of this amount allocated to the operation of its data center network expansion, which the company fed from its ubiquitous enterprise search engines, to its Maps application, YouTube and of course, Google Cloud Platform.
The DeepMind system was able to reduce energy consumption in data centers determined by several percentage points, according Hassabis. “It’s not only a great savings in terms of cost, but also beneficial to the environment,” he said.
Other system improvements are already underway, not only to improve as time goes on, but because the company will put additional sensors in their data centers to fill the blind spots where artificial intelligence currently lacks information, Hassabis said.