The Growing Demand for Data Centers: A Supply Chain Challenge
The digital age is upon us, and with it comes an insatiable demand for data centers. Recent reports indicate that the demand for new data centers has significantly outstripped the available stock. Developers are grappling with challenges in sourcing vital inputs, particularly power and other key components necessary for construction and operation. As a result, vacancy rates have plummeted to a mere 3%, with over 80% of new centers in major markets already pre-leased. This surge in demand is driving up lease rates, as highlighted in a recent report by Cushman & Wakefield.
Shifting Focus to Peripheral Markets
As the need for data centers grows, so does the interest in large-scale power availability and ample land. This shift has led hyperscalers and operators to explore historically peripheral markets, such as rural Georgia, North Carolina, Pennsylvania, Texas, Minnesota, and the Dakotas. These areas, once considered outliers, are now becoming attractive options for data center development, particularly for those with less stringent latency requirements for artificial intelligence (AI) applications. Bo Bond, executive managing director at Cushman & Wakefield, notes that this trend is reshaping the landscape of data center locations.
Innovative Power Solutions
To meet the increasing demand for power, some operators are collaborating with power companies to establish substations and transmission lines. Others are exploring alternative energy sources, working directly with third-party energy generation developers. This includes partnerships with companies specializing in wind, solar, battery storage, natural gas, and geothermal energy. The report anticipates a surge in operators seeking large-acreage sites with high-power capacity, particularly in rural markets. This proactive approach to securing power is essential for accommodating the long-term needs of data center operations.
The Role of AI in Data Center Development
AI is becoming a pivotal factor in the development of both hyperscale and colocation data centers. The report emphasizes that AI training facilities, which focus on teaching AI models to learn from data, are primarily located in large, latency-tolerant rural markets. Conversely, AI inference facilities, where AI models quickly process information, are strategically positioned near major cloud regions. This distinction highlights the varying requirements for data center locations based on the specific functions they serve in the AI ecosystem.
The Emergence of New Markets
As hyperscale growth accelerates, both in established and emerging markets, the report predicts further developments in newly emerging data center markets. Providers and investors are increasingly prioritizing power transmission capabilities and hyperscaler deployment. This trend is reshaping the traditional understanding of data center markets, with utility provider service areas becoming more influential than conventional metropolitan development patterns.
Leading Data Center Markets
Currently, Virginia stands as the world’s largest data center market, boasting 13.2 gigawatts (GW) of active or in-development capacity. Following Virginia are Atlanta (3.8 GW), Phoenix (2.9 GW), Chicago (2.4 GW), Dallas (2.3 GW), and Silicon Valley (1.3 GW). Other notable areas include Portland and Eastern Oregon (1.9 GW) and Columbus, OH (1.8 GW), driven by both hyperscale self-build projects and colocation deployments. This distribution of data center capacity underscores the evolving dynamics of the industry and the importance of strategic location selection.
Redefining Development Areas
The landscape of data center development is being redefined by the availability of utility services rather than traditional metropolitan factors. As the demand for data centers continues to rise, the focus on power availability and strategic location will play a crucial role in shaping the future of the industry. The interplay between technological advancements, energy solutions, and market dynamics will determine how and where data centers are built in the coming years.