Although AI needs resources like energy, innovative solutions help this challenge. Even if the global workloads of Datacenter in 2020 were about nine times what they were in 2010, for example, the demand for electricity from the data center increased only 10%.
This is partly because Microsoft works alone and with others, like AMD, Intel and Nvidia, to make its equipment more effective, from its personalized silicon series, Azure Maia and Cobaltto its unit of liquid cooling heat exchanger designed to effectively cool Large -scale AI systems.
In the coming years, new data centers supporting AI will be online and consume zero water for cooling and the company will expand its use of deleted liquid cooling systems such as Cold plates.
All this is part of a wider effort to make the infrastructure that AI is built on more efficient and durable in 2025.
As Microsoft helps build a more efficient AI infrastructure, it will continue to invest and use more Low carbon building materialsAs in carbon steel near zero, concrete alternatives and lamit-lami wood.
Microsoft will also continue to invest and use carbon -free energy sources such as wind, geothermal energy, nuclear and solar power. The company makes long -term investments to provide more carbon -free electricity on the networks where it works and continues to plead for the expansion of clean energy solutions in the world.
It is only a tranche of Microsoft’s planned infrastructure that will advance its objective of being a negative carbon company, positive in water, zero waste by 2030, explains Mark Russinovich, director of Azure technology, deputy director of information security and technical scholarship holders.
“In 2025 and beyond, we are increasingly going to a holistic vision of data centers, energy and resources, so that we can maximize the efficiency of our entire infrastructure,” explains Russinovich.