Artificial intelligence is revolutionizing the future, but it comes with significant environmental costs, primarily due to its reliance on data centers that consume excessive electricity and water, mostly from fossil fuels. As AI adoption rises, so do emissions, with data centers now using more energy than entire nations. A study by the University of California, Riverside, introduces Federated Carbon Intelligence (FCI), a system that aims to mitigate these environmental impacts while prolonging server life. FCI intelligently routes AI workloads by integrating real-time server health data with environmental metrics, cutting carbon dioxide emissions by up to 45% over five years and extending server operation by 1.6 years. This dual approach addresses the deficiencies in sustainability strategies by emphasizing smarter hardware management alongside renewable energy. The researchers aim to collaborate with cloud providers to implement FCI in real-world data centers, enhancing sustainability without requiring new equipment. The study is published in MRS Energy and Sustainability.
Source link
Share
Read more