An innovative system called FCI It has shown that artificial ielligence can reduce the carbon emissions of large data ceers by 45%, and at the same time extend the life of servers by 1.6 years.
With the ever-increasing growth of artificial ielligence, the consumption of electricity and water in massive data ceers has increased dramatically, and most of this energy is still supplied by fossil fuels, which puts significa pressure on the environme. Today’s data ceers consume more energy than most couries, and as servers heat up and age, their energy needs and carbon footpri increase.
Increasing the lifespan of servers and reducing carbon emissions with the help of the FCI smart system
A University of California study offers a solution that lowers both environmeal pollution and hardware stress caused by artificial ielligence. According to reports, the system Federated Carbon Ielligence (FCI) It ielligely manages AI tasks. This system first analyzes the environmeal data and the curre status of each server and then decides which servers the workload should be transferred to. By directing work to healthier, less worn servers, FCI preves excessive wear and tear and reduces the need for cooling. This approach both optimizes energy consumption and increases the lifespan of the hardware.
Data ceer sustainability and energy researcher Mihri Ozkan says sustainability in AI cannot be achieved by using clean energy alone, as servers get old and hot over time and their performance changes, which creates a significa carbon cost. He emphasizes that simulations show that the FCI system is able to reduce carbon dioxide emissions by 45% over five years and increase the operational life of servers by an average of 1.6 years.
This system reduces the use of stressed and damaged servers by real-time monitoring of server age, temperature and wear rate, thereby reducing both energy and water consumption and maiaining long-term stability and reliability of servers. On the other hand, replacing aging servers creates a significa carbon footpri in addition to the financial cost, and by extending the lifespan of the hardware, FCI significaly reduces this hidden environmeal pressure.
FCI’s system dynamically determines when and on which server each AI task should be processed, using up-to-date data on workload, power carbon, and server health to make these decisions. This adaptive framework can be implemeed without the need for new equipme and only through iellige coordination between existing systems. The research team plans to collaborate with cloud service providers and test FCI in real data ceers, a move that has become even more imperative as the demand for artificial ielligence grows rapidly.
