Data centers are among the fastest-growing sources of electricity demand, with some now consuming as much power as entire cities. But unlike factories or hospitals, data center workloads are flexible. WESTMAP.AI, a new tool developed by researchers at the University of Utah’s John and Marcia Price College of Engineering, models and quantifies what happens when data center flexibility supports the power grid rather than strains it.
The WESTMAP.AI tool models and visualizes how data centers interconnect and operate within the Western Interconnection, the power grid that serves over 80 million people across 11 U.S. states, two Canadian provinces, and parts of Mexico. Its key findings demonstrate the impact of three approaches data centers can begin to take immediately: temporal flexibility, spatial flexibility, and deploying on-site energy resources. The model behind WESTMAP.AI was presented as a paper at the 2026 HICSS Conference.
Depending on the implementation of these approaches, data centers could expect to save between $62 and $590 million a year.
The WESTMAP.AI tool is developed by the NSF WIRED Global Center, involving Masood Parvania, Roger P. Webb Endowed Professor in the department of Electrical & Computer Engineering and Director of the NSF WIRED Global Center, along with Mohammad Amin Mirzaei, a Senior Power System Research Engineer there.
WESTMAP.AI allows users to model different growth scenarios for data center construction.
“Our tool was not developed to study any specific project,” says Parvania, “but rather is a generic platform that allows unbiased impact and feasibility analysis of various data center growth scenarios from the perspective of power grid, based on proven scientific and engineering methods.”
Their study’s first approach, temporal flexibility, involves data centers reducing their workloads during times when power demand is highest.
“Not all computing has to happen right away,” Parvania says. “AI training, backups, and batch jobs can run at night or during off peak hours when power is cheapest. At scale, this smooths demand across the entire grid.”
The second approach, spatial flexibility, takes advantage of local variations in power demand and price. By coordinating with data centers elsewhere on the grid, more sensitive computations can be completed even when local power demand is high.
“When one region’s grid is stressed, workloads move to a data center with surplus power,” Mirzaei says. “Instead of shipping electricity across overloaded lines, you ship the computation. The electrons stay local.”
Finally, data centers can invest in on-site resources that extend the impact of temporal and spatial flexibility. The more power a data center can generate and store on its own, the less stress it puts on the grid.
With only temporal flexibility measures, the researchers’ model shows grid savings of about $171,000 per day, or $62 million per year.
When data centers add spatial coordination — shifting workloads to regions where power is cheaper and more abundant — combined savings increase to $468,000 per day, or $171 million per year.
With all three approaches in place, including on-site generation and battery storage, the estimated savings reach $1.62 million per day, or $590 million per year. That reduction in fossil-fuel generation translates to approximately 3.5 million fewer tonnes of CO2 per year — comparable to taking more than 770,000 cars off the road.
“The technology and the data are ready — what’s needed now is action,” Parvania says. “Grid operators and the data center industry need to work together to make flexibility the standard, not the exception.”
“WESTMAP.AI gives them the roadmap,” Mirzaei adds. “The savings are there. The emissions reductions are there. Every day we wait is a day we leave more than a million dollars on the table.”