Webinar: Data Centers, AI, and the Future of Power Systems

Gigawatt-scale AI data centers are entering interconnection queues across the country, but traditional grid planning tools weren't designed for these large loads. The questions from ISOs and utilities are getting harder. The timelines are getting longer. And the old approaches to load modeling and grid integration aren't working. 

Watch our technical discussion featuring Harvard, National Lab of the Rockies (NLR), and EPE engineers to discover what's changing – and what it means for your project. 

Watch the on-demand webinar here.

Panelists

• Le Xie, PhD, Gordon McKay Professor of Electrical Engineering, Harvard University 
• Kumaraguru Prabakar, PhD, Principal Engineer, National Laboratory of the Rockies (formerly National Renewable Energy Laboratory) 
• Billy Yancey, Vice President, Technical Services, Electric Power Engineers 
• Kevin Chen, PhD, Head of Advisory Services, Electric Power Engineers 

The Technical Challenge 

Traditional composite load models (CMLD) assume motor-driven loads with predictable ramp rates and power factor behavior. AI data centers are fundamentally different in several ways: 

• Power electronics-dominated loads rather than rotating machinery 
• Microsecond-scale ramps that challenge grid protection schemes 
• Gigawatt-scale facilities that represent significant portions of regional load 
• Closed-loop control systems that respond to IT workload scheduling, not just voltage/frequency 

ISOs and utilities are starting to require detailed modeling and validation that existing approaches can't provide. If you're developing data center infrastructure or conducting interconnection studies, you need to understand where the requirements are headed. 

What We Cover 

1. Load Modeling for Power Electronics-Based Data Centers: Why traditional CMLDs fail for GPU workloads, and what's needed instead.

2. Model Validation Without Stage Testing: Generators get MOD 25/26/27 stage tests. Data centers need a different approach. 

3. Data Center Flexibility and Grid Services: Bitcoin mining proved flexible loads work, but AI data centers are more complex. 

4. Infrastructure Constraints and Development Strategy: 7% annual demand growth meets multi-year transmission timelines. 

5. AI Tools in Grid Planning: Can AI help us plan for AI's power demand? Where does automation help and where does engineering judgment remain critical? 

Who Should Watch 

This webinar is relevant if you're: 

• Developing data center infrastructure (hyperscale, colocation, enterprise) 
• Conducting interconnection studies for large loads 
• Planning transmission or distribution systems accommodating data center growth 
• Evaluating sites for power-intensive facilities 
• Advising clients on data center power strategy 
• Managing generation projects competing for interconnection capacity 

 Watch the on-demand webinar here.