SUPPLY CHAIN
“ Our data centres are about twice as efficient as a standard enterprise data centre . And , in fact , compared to five years ago , we produce about five times as much computing power for the same amount of electricity used . AI is a key way we do this ,” Elman comments .
“ Deep mind is an AI machine-learning specialist organisation within Google . A number of years ago , this started using neural networks to take thousands of data points and figure out optimisations that humans simply can ' t do , due to the amount of data and the variability within that .”
“ And using those data points , passing them through an AI efficiency model , we were able to improve the efficiency of our cooling system by 30 %,” Elman states .
For the emissions coming from the data centres , Google uses a carbon intelligence programme to predict how much computing power will be needed on a day- to-day basis and how much renewable energy they think will be on the grid .
“ Using those data sets , we can then decide where we want to use that computing power . We can shift it between different data centres , giving us complete flexibility to use that computing power where the grids are the cleanest in the world . It ' s a really unique system .”
Alongside this , AI is also being used by Google to give cities and citydwellers actionable insights and precise sustainability metrics .
“ One of the solutions that we make available for free to cities and regions around the world is called Environmental Insights Explorer . Insights are really key to developing carbon plans and then tracking progress against them ,” Elman explains . “ Cities account for around 70 % of carbon emissions , but those cities often lack the data and the tools they need to create those really robust plans and track