37. savetovanje CIGRE Srbija (2025) SIGURNOST, STABILNOST, POUZDANOST I RESILIENCE ELEKTROENERGETSKOG SISTEMA MULTISEKTORSKO POVEZIVANJE U ENERGETICI I PRIVREDI – D2-05
AUTOR(I) / AUTHOR(S): Jelena Apostolović, Miljana Stojanović
DOI: 10.46793/CIGRE37.D2.05
SAŽETAK / ABSTRACT:
The rapid development of artificial intelligence (AI) significantly increases the demand for computing resources, directly influencing the rise in power consumption in data centers. The growing energy demands of data centers require more efficient cooling systems, which further increase water and power consumption and place additional strain on the environmental system. This trend presents challenges in terms of sustainability and reducing the environmental footprint (carbon, water, and infrastructure-related). Optimizing energy systems and implementing advanced technologies are key to mitigating these negative impacts.
The integration of smart resource management systems can contribute to better control of energy and water consumption, while more efficient infrastructure design enables the reduction of losses and increased reliability. Engineering solutions should focus on balancing the expansion of AI capacity with environmental preservation. The adoption of renewable energy sources and innovative cooling methods is becoming increasingly important. The future of data centers depends on the development of sustainable and energy-efficient models, with the strategic priority of ensuring stable performance while minimizing environmental consequences. This paper will examine the comparisons and differences between traditional data centers and AI data centers, the impact of increasing AI demands on power consumption, and their effect on the environmental footprint. Additionally, the advantages and „challenges“ of AI data centers will be analysed through various solutions.
KLJUČNE REČI / KEYWORDS:
AI, Data Center, power consumption, environmental footprint
PROJEKAT / ACKNOWLEDGEMENT:
LITERATURA / REFERENCES:
- E. Strubell, A. Ganesh, A. McCallum, „Energy and Policy Considerations for Deep Learning in NLP“, Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pp. 3645–3650, Florence, Italy, 2019.
- Google DeepMind (2017). Machine learning can boost the value of wind energy. Wired
- E. Masanet, A. Shehabi, N. Lei, S. Smith, J. Koomey, „Recalibrating global data center energy-use estimates“, Science, Vol. 367, No. 6481, pp. 984–986, 2020.
- Uptime Institute, „Global Data Center Survey 2023: Five Trends Shaping the Future“, Uptime Institute Report, 2023.
- U.S. Department of Energy (2021). Data Center Energy Efficiency & Sustainability Best Practices. Energy.gov
- IEA (International Energy Agency) (2023). Data Centres and Energy Demand: Global Trends & Future Scenarios. IEA Report
- ASHRAE (2022). Guide on Data Center Cooling. American Society of Heating, Refrigerating and Air-Conditioning Engineers.
- Tesla AI Infrastructure Team (2023). Dojo Supercomputer: AI Training at Scale. Tesla Research Publication
- Jones, N. (2018). The AI revolution and its environmental cost. Nature, 563(7729), 182-185. DOI: 10.1038/d41586-018-07286-2