Center for Technology and Innovation

Da⁠t⁠a Cen⁠t⁠ers Face an Uncer⁠t⁠a⁠i⁠n Fu⁠t⁠ure

By: Turner Loesel / 2024

Turner Loesel

POLICY ANALYST

Center for Technology and Innovation

2024

July 31, 2024

For nearly three decades, the U.S. has experienced gradual growth in energy demand, with demand increasing an average of only 0.3% every year. Such gradual increases are attributable to Americans adopting more energy-efficient technology like LED lighting, smart thermostats, and hybrid vehicles. However, this trend has taken a dramatic turn in recent years. The Federal Energy Regulatory Commission (FERC) announced that grid planners predict a 4.7% increase in nationwide electricity demand over the next five years— a figure they believe to be conservative. John Wilson and Zach Zimmerman of Grid Strategies explain that this surge is caused by “general economic growth, population growth, temperature trends, and electrification.” 

To address this spike in demand, policymakers need to implement forward-thinking strategies that will allow the country’s energy infrastructure to keep pace with technological advancements and our insatiable appetite for power. 

Data centers, the massive computer facilities forming our digital world’s backbone, are contributing to this surge in demand. Inside these centers are football field-sized banks of computer servers responsible for storing, processing, and disseminating information. These facilities are also vital to the continued advancement of artificial intelligence (AI) technologies. These power-hungry hubs consume 4% of the U.S. energy supply, and demand is projected to double to 9% by 2030. This trend is evident nationwide, with even disaster-prone states like Florida seeing significant growth, where 123 data centers have been built across the state to supplement escalating computational demands. The rise of computationally intensive AI models will make data centers critical components of American infrastructure over the coming years.  

To operate, data centers require vast amounts of almost constant power. While lightbulbs can be turned on or off as needed, data centers require a sustained energy supply. To remain operational, servers must also be cooled to avoid overheating. Cooling alone accounts for 40% of a data center’s energy consumption. A single large data center can require the same amount of electricity needed to power 750,000 homes. Highlighting the vast power data centers will need going forward, Goldman Sachs reported that “US utilities will need to invest around $50 billion in new generation capacity just to support data centers alone.” As the AI revolution continues, the report estimates that the demand for data center power will grow by 160% by 2030.

The expected strain on the energy grid has prompted pushback from several states. Georgia passed legislation banning new data centers around their beltline following a 211% increase in data center construction since 2023, although the Governor ultimately vetoed the legislation. Arizona, Illinois, and Arkansas have also passed legislation restricting or banning data center construction. Even at the municipal level, projects have been subjected to lengthy planning reviews, delaying projects.

However, these measures risk jeopardizing the substantial economic benefits that data centers bring to their communities. In Loudoun County, Virginia, data centers generated $528 million in tax revenue in 2023 – “two and a half times the tax revenue from motor vehicle sales.” These facilities create significant employment opportunities, both directly and indirectly. The U.S. Chamber of Commerce Technology Engagement Center reports that, on average, data centers employ 1,688 local workers, provide $77.7 million in wages, produce $243.5 million in output along the local economy’s supply chain, and generate $9.9 million in revenue for state and local governments.

Crafting legislation that makes it harder for data centers to exist could also jeopardize AI technology more broadly, denying Americans the benefits of the emerging technology. Raul Martynek, CEO of Data Bank, cautions that “without adequate data center infrastructure to support the HPC (high-performance computing) hardware required to train and operate AI models, AI adoption could be slowed.” A lagged adoption of AI could have significant consequences for American consumers, who may miss out on innovative products and services that improve their daily lives. Furthermore, it could undermine America’s global competitiveness by pushing innovators to rival markets like China.

Criticizing data center development may be low-hanging fruit for lawmakers who still see large technology companies as entities needing further regulation, but hampering data center development could have far-reaching consequences for our digital future and emergent technologies. These facilities are critical not only for supporting local economies but also for supporting the growth of AI. However, clamping down on data centers over concerns over energy use is not the answer. Instead, lawmakers need to find a way to balance supporting data center growth with ensuring that the country’s grid can keep pace with increasing energy demands.