2025.08 EJ DCFlex Workstreams

Not all data centers are alike. While a growing number of data centers are built with AI applications in mind, others provide cloud computing services. The varying purposes of data centers impacts their electricity demands and potential flexibility. "There are big differences in how they operate and what they can and can't do with compute load inside the data center," Porter said. For example, data centers supporting cloud computing products like Google Translate, Google Maps, and YouTube all have differing flexibility characteristics—characteristics that also differ from AI data centers. This workstream investigates, validates, and demonstrates new data center designs that enable flexible operations to support the grid when needed.

Grid-Informed Flexible Data Center Designs

A big challenge to leveraging the potential value of flexible data center operations is the lack of confidence many utilities have in the reliability of demand-side resources. "It's sometimes difficult for our utility partners to feel like they can rely on demand-side resources alongside supply-side resources," Kobor said. "This is the moment where we really need to find the models that get us past that logjam." This DCFlex workstream seeks to bolster understanding and confidence in the flexibility characteristics of data center loads. But it's also focused on developing utility programs that encourage and fully leverage flexibility. This can happen through incentives and by utilizing beneficial market structures. Importantly, utilities also understand that incentives and programs must address equity and affordability concerns.

Transformational Utility Programs

The third DCFlex workstream is focused on developing the tools and algorithms utilities need to confidently integrate flexibility into their interconnection, operations, and forecasting activities. This will include the creation of data center forecast maps and granular load forecasting methodologies. As data center development scales, it will be especially important for utilities and data centers to adopt standardized communication protocols. "Data centers don't want to figure out how to communicate and take signals from all these different utilities in a different way," Porter said. "Part of what we're driving with DCFlex is to develop standard protocols that then can work with all large loads going forward. That gets to be much more seamless, and we can make flexibility for large point loads more accessible to more types of companies and better support the grid."

Operational Flexibility Framework

Speed is paramount to data center development. Closing the gap between when a data center needs electricity and a utility's ability to provide it is at the heart of DCFlex. In some cases, that will mean devising interim solutions that allow data centers to begin operating while utilities expand their capacity to meet demand. This workstream examines both front-of-meter and behind-the-meter solutions to meet the near-term electricity needs of data centers. "What can we do with bridge solutions to meet data center needs in a timely manner?" Porter said. "Particularly in places they may need to build for business purposes, but the transmission capacity may not be there, and it may not be there for quite some time. How do we bridge that gap so that those companies can get up and running and help grow the economy?"

Data Center-Informed Energy Supply