The expansion of AI data centers represents the most significant shift in energy demand that the Commonwealth of Pennsylvania has seen in decades. Following major investment announcements in 2025, including a landmark expansion by Amazon Web Services, the state's infrastructure is facing unprecedented pressures. For the first time in decades, the power grid is experiencing a sustained increase in electricity demand that is projected to grow for years to come.
The Surge in Electricity Demand
The scale of this growth is transforming regional energy planning. While data centers currently account for roughly 4% of national electricity usage, that figure is expected to reach 12% by 2030 with certain areas, like Pennsylvania, seeing even more growth.
- Regional Impact (PJM): Between 2025 and 2030, PJM Interconnection projects a load growth of 32 GW. An estimated 90% of that growth (30 GW) is attributed to data centers.
- Local Impact (PPL): PPL Electric Utilities anticipates 14.4 GW of load growth over the next decade. To put this into perspective, PPL’s current peak load is 7.5 GW.
As an energy exporter, with significant natural gas resources, Pennsylvania is uniquely positioned to meet the growing electricity demands of data centers. However, the Commonwealth has no recent historical reference for the significant energy uptick currently underway. The implications of this data center load growth on grid operations and expansion, regulations and policies, and economic and environmental impact on communities is complex and often unclear.
The Energy-Water Nexus
The footprint of an AI Data Center extends far beyond the server rack. High-performance computing equipment is needed to train and run AI models and algorithms. As this equipment operates it generates intense heat, requiring energy-intensive cooling systems that may strain local water supplies.
Data centers are already one of the top 10 water-consuming industries in the U.S. This water consumption is driven by three main factors:
- Cooling: Water is used to cool facilities and servers, maintaining optimal temperature for hardware to operate.
- Electricity Generation: A large amount of water is required to generate the electricity that data centers consume.
- Manufacturing: Water-intensive processes are involved in manufacturing equipment used in data centers.
Ultimately, solving the energy demands of AI requires a holistic nexus approach that accounts for the significant water intensity of both power generation and hardware cooling.
A Path Towards Innovation in AI Data Center Technologies
The ACES workshop brought together leaders from academia, industry, and government to identify emerging opportunities for innovation. By understanding the scope of the challenges, we can begin to alleviate the energy and water pressures from AI data centers through smarter technology and more resilient policy.