DataBank Senior Director of Sustainability Jenny Gerson recently contributed to The AI Journal, addressing widespread concerns about AI’s exponential power requirements. While projections suggest AI data center power demand could grow 30-fold by 2035, with facilities scaling from 10-50MW to 100-500MW or more, Gerson argues the industry has the tools to manage this growth without triggering infrastructure collapse.
“The data center industry isn’t necessarily facing a fundamental power generation problem. The U.S. grid already maintains substantial reserves, with operators typically utilizing only about 50% of available generating capacity by design.”
— Jenny Gerson, Senior Director of Sustainability, DataBank
The article examines how data center operators are implementing energy storage systems, virtual power plant networks, and demand response strategies to create grid flexibility. A 2025 Duke University study found that if data centers can be flexible for just 22 hours per year, the U.S. grid can accommodate 76GW of new capacity without building new power plants.
Today, transmission infrastructure—not power generation—is the real bottleneck, with utilities facing multiyear backlogs and interconnection studies spanning four to eight years. She calls for coordinated action: utilities must accelerate approvals and transmission upgrades while data center operators continue deploying behind-the-meter solutions, strategic facility placement, and clean energy adoption [https://www.databank.com/resources/blogs/decarbonizing-data-centers/].
Read Jenny Gerson’s full analysis on AI power demands, its impact on the environment grid flexibility, and transmission challenges at The AI Journal.
