Data centers have been identified as job-creation catalysts. These projections can seem promising for rural areas, especially those focused on developing their local economies and strengthening their workforce. However, there are sub-currents to those expectations: Data centers create jobs during construction but need vastly fewer jobs to maintain operations. At the same time, data centers may create conditions that favor a high-tech corridor for other new jobs. Finally, the broadband and power infrastructure required to support data centers, especially those housing AI training centers, may be attractive to other businesses that rely on robust, high-speed internet connections.
By way of introduction, AI engines are "trained" on large data sets, often from internet sources that include online news, social media, and commercial and academic content; AI tools are often described as "scraping" the internet for data. These AI engines, and the data centers in which they are housed, consume massive amounts of power and bandwidth for intensive computational needs. High-speed, low-latency symmetrical broadband that can support high-speed, simultaneous data transfers in both directions is a critical part of this process. Data center networks typically rely on 400 Gbps per end device, but an increase to 800 Gbps is anticipated as necessary to meet new market developments. Additionally, AI racks can demand 50kW to 100kW or more and require substantial cooling solutions, including liquid-based cooling when air-based systems are insufficient. AI is expected to account for 70% of data center capacity demand by 2030, reflecting a 33% CGAR growth between 2023 and 2030. "Hyperscalers" such as Amazon Web Services, Google Cloud, Microsoft Azure and Baidu are fueling most of this growth, and 43% of new data center facilities are expected to be dedicated to AI workloads.
But according to the Wall Street Journal, once up and running, data centers require a fraction of the jobs that were invested in building the facility. Other analysts agree, citing a surge in jobs to support construction and configuration but relatively smaller, concentrated job clusters once the centers are up and running. Nevertheless, that does not mean that data centers don't bring good.
The same energy needs that data centers require can be leveraged to support other industries with similar demands, including chemical, paper, and food and beverage processing. Likewise, lightning speed bandwidth is attractive to firms focused on advanced research and development as well as cybersecurity. Virginia's Data Center Alley, just down the road from NTCA in Ashburn, Virginia, is credited with creating more than three jobs for each long-term job in the data center.
And all of this is independent of the general benefits seen from engagement in rural spaces with broadband that would anyway be needed to support data centers. A study from Oklahoma State University found increased "business births" across eight states in areas with 100 Mbps. Another study from faculty at University of Illinois and University of Wisconsin-Madison found positive impacts of broadband on women-led start-ups in remote rural areas. And, a study supported by NTCA and other industry partners found substantially higher business growth, GDP increases, self-employment gains and higher per-capita income in rural areas with high broadband adoption rates.
The relationship between data centers and rural job growth is nuanced and reflects more than only initial construction and lasting follow-on jobs. Broadband and power infrastructure needs can be catalysts for ongoing economic development efforts. Accordingly, communities studying data centers can expand the scope of their analysis to consider broader community or even regional impacts. A strategy that pursues not only the data center but other industries that rely on similar needs can position communities competitively as they develop economic development strategies for the future.