What is it about?
AI data centers need steady clean power. We show how pairing a data center with a small modular reactor and planning at day day-ahead level, matching flexible jobs to reactor output and low-carbon grid energy, can cut fossil use by about 40% without hurting performance.
Featured Image
Photo by Alexandre Viard on Unsplash
Why is it important?
This work is the first to show how small modular nuclear reactors can be practically coordinated with data centers. While data centers and nuclear plants have been studied separately, we demonstrate how to jointly schedule both computing workloads and flexible reactor output in a simple day-ahead plan. This is timely because demand from AI and cloud computing is skyrocketing while grids are struggling to decarbonize. Our approach offers a realistic way for data centers to directly tap into clean, reliable nuclear energy, potentially transforming how digital infrastructure is powered while cutting carbon emissions at scale.
Perspectives
Writing this article was particularly meaningful for me because it allowed me to bring together two fields I care about, sustainable computing and clean energy. The rapid rise of AI and data centers can feel at odds with climate goals, but this work shows that with creative planning, digital growth and decarbonization can go hand in hand. I hope readers see this as not just a technical solution, but as a hopeful example of how innovation in one sector can directly support global sustainability.
Yijie Yang
Hong Kong Polytechnic University
Read the Original
This page is a summary of: Uncertainty-Aware Day-Ahead Datacenter Workload Planning with Load-Following Small Modular Reactors, ACM SIGEnergy Energy Informatics Review, July 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3757892.3757905.
You can read the full text:
Contributors
The following have contributed to this page
