This study present the first integrated assessment of how combined pressures from AI expansion and climate-driven cooling demand may create electricity bottlenecks in U.S. power systems. The following contents are included in this repository to support our key findings:
To run the codes in this repository, the following Python and core packages must be installed (version is given for refenrence):
- Python 3.9.13
- numpy 1.21.5
- pandas 2.23
The above packages can be conveniently downloaded through open-source library.
- Reserve Margin Rate Calculation.py: This file contains the calculation process for calculating the reserve margin rates of the U.S. grid regions between 2025 to 2034.
- AI Distribution.csv: the original distribution of AI servers across U.S. regions.
- Baseload Growth Rate.xlsx: the applied baseload growth rate data
- Regional AI Capacity.xlsx: the AI capacity across U.S. regions under the original distribution
- nerc_2024_projection_by_region.xlsx: nerc long-term grid projection for U.S. regions
- nerc_actual_load_2024_by_region.csv: nerc reproted actual peak load of U.S. regions
- PUE Values: contain used PUE values
- Space Cooling Hourly Load: contain used hourly space cooling load values
The file Reserve Margin Rate Calculation.py can be used to run a base case for estimating reserve margin rates. After download the codes and data, simply replace the "FILE PATH" used in the code file with the install path of our data folder to run the simulation.
Please use the following citation when using the data, methods or results of this work:
Xiao, T., You, F., Electricity Bottlenecks Emerge from AI Surge and Climate Change Strain in the U.S. Submitted to Science.
This project is covered under the Apache 2.0 License.