Large-load flexibility is increasingly becoming a necessity—not just a best practice. That’s according to the experts from Google, CoreWeave, Scale Microgrids, and Luminary Strategies who took the stage at DERVOS last week.Â
What flexibility looks like: Data centers can carefully adjust their power consumption to avoid straining the grid. That could mean, for example, shifting certain training tasks to off-peak hours.
Google’s perspective: Right now, the company isn’t “interested in economically optimizing its energy costs," said Varun Mehra, Google's technical lead for data center flexibility. Translation: The company is chasing certain, not cheap, power. And to get online quickly, the company is willing to get flexible.
The outlook: The panelists were optimistic that data centers can make flexibility work. Why?Â
The solutions are there: Easy, available fixes can make flexibility workable for customers, said Angela Kent, energy & climate policy engagement lead at CoreWeave.Â
“It’s demand-driven,” said Victoria Fleming, chief business officer of Scale. Data center developers, not policymakers or regulators, are leading the charge. That’s out of necessity—and the urgency will only continue to grow.
Regulators are moving fast(ish). The pace of FERC docket decisions is "surpassing anything that I've ever seen,” Kent said.
The bottom line: Flexibility is no longer optional. Take ERCOT, where all-time peak load hit roughly 91 GW this summer. The grid’s large-load queue: more than 400 GW. In ERCOT and around the country, new transmission alone won’t cut it.