Data center developers and operators are facing increasing challenges in securing power for their facilities. With the growing demand for energy-intensive operations like AI, the need for innovative solutions has never been more pressing. One approach gaining traction is the adoption of behind-the-meter configurations, where power generation and consumption take place on the same site.
According to a report by Boston Consulting Group, the data center power shortage could reach over 45 GW, with thousands of projects seeking grid interconnection. As AI technology continues to drive up data center power consumption, the industry is expected to consume 7.5% of U.S. electricity in the next five years.
During a recent presentation at Data Center World, AFCOM program chair Bill Kleyman highlighted the critical issue of power constraints facing AI data centers. The AFCOM 2025 State of the Data Center report revealed that 62% of data centers are exploring on-site power generation to improve energy efficiency and resilience, with 19% already implementing behind-the-meter power solutions.
The behind-the-meter approach involves integrating renewable energy assets directly into data center facilities. This strategy helps alleviate grid congestion, reduce transmission losses, minimize environmental impact, accelerate deployment timelines, and enhance facility reliability by reducing vulnerability to outages.
While interest in renewable energy remains high among data centers, the limited availability of such resources has led many to turn to alternative solutions like natural gas generation and small modular reactors (SMRs) for reliable power supply.
Industry analysts predict a surge in AI workloads in the coming years, driving up power demand significantly. Vlad Galabov, research director at Omdia, forecasts that AI will account for over 50% of global data center capacity and more than 70% of revenue opportunities by 2030, requiring around 35 GW of self-generated data center power.
To meet the growing power demands of AI factories, developers are increasingly looking towards behind-the-meter and Bring Your Own Power (BYOP) solutions. Galabov anticipates increased investment and partnerships between hyperscalers, colocation providers, and AI-ready facility developers to address the power needs of the industry.
As data center capital expenditure is expected to reach $1 trillion globally by 2030, investments in physical infrastructure for power will play a crucial role in supporting the industry’s growth.
Many next-generation data center developers are exploring locations with abundant natural gas resources to meet their power requirements. With the favorable economics of natural gas extraction, low prices, and an extensive pipeline network, natural gas is becoming a preferred energy source for data centers.
The race to deliver AI services at a rapid pace has highlighted the importance of timely power provision. Builders are facing challenges in sourcing high-performance chips, electrical equipment, and power generation technologies like gas turbines and SMRs. Companies like Applied Digital are leveraging stranded power sources, such as wind farms in North Dakota, to quickly energize their operations and stay ahead in the AI race.
In conclusion, the data center industry is undergoing a significant transformation in response to the escalating power demands driven by AI technology. By embracing innovative approaches like behind-the-meter configurations and leveraging diverse energy sources like natural gas, data centers are positioning themselves to meet the challenges of the future and sustainably power the digital economy.
