How companies can gain the advantage by reevaluating AI data center plans using location intelligence
Artificial Intelligence (AI) data center site selection is expected to be a major challenge as generative AI expands into many of today’s enterprise computing applications. The reason is simple.
“The U.S. is facing a massive surge in power demand from data centers and electrification of the economy. The tech sector’s build out of data centers to support AI and the adoption of electric vehicles alone is expected to add 290 terawatt hours of electricity demand by the end of the decade,” according to a report released by the consulting firm Rystad Energy.
Some industry analysts estimate global data center capacity to double by 2030. As a result, to find the optimal location for new AI data centers, companies must consider:
- Data center power consumption and energy sources
- Availability of broadband
- Climate resilience
- Proximity to electrical grid transmission facilities
There are also other tangential factors related to permitting which could take between three to five years depending on environmental impacts and local regulatory mandates.
Location Intelligence Growth Solutions
Taken together, these components comprise the Location Intelligence for AI Infrastructure: LI4AI, which provides the context that will factor into site selection and suitability decisions.
Energy Demand
According to the Electric Power Research Institute, data centers powering advanced AI models could account for up to 9.1% of the U.S.’s overall energy demand by the end of the decade. Today that consumption is 4%.
The reason for this 5.1% increase: cooling.
AI chips require 10x more cooling capacity than chips that currently run cloud data centers, according to the Wall Street Journal. Hyperscale facilities consume between 20 and 50 megawatts annually. By comparison, the U.S. Energy Information Administration reports that the total electricity consumption by electric vehicles in 2023 was 7.6 million megawatts hours, and that was an increase of 45% from 2022’s total.
As a result, data centers will compete with manufacturing facilities and homes, which often serve as offices for remote workers and charging stations for electric vehicles. In response to this increased demand, electricity costs could rise.
The increasing number of data centers will only compound these energy costs. In 2021, there were 314 new hyperscale sites in development globally. That number was projected to surpass 1,000 by the end of 2024. By 2026, hyperscale data centers are expected to grow by 20% when compared to 2021.
Broadband
Artificial intelligence comes with a very specific cost: more bandwidth. As AI is deployed to facilitate more video content, support Large Language Models (LLMs), and achieve greater analytical computation, more bandwidth is an indispensable consideration.
“High-speed internet and multiple fiber providers and routes are essential for minimizing latency and ensuring reliable data transfer,” noted Dominika Koncewicz, Senior Analyst, Cloud Services and SD-WAN when writing in DataCenter Knowledge Magazine. “For instance, Northern Virginia, known as ‘Data Center Alley,’ has become a top choice due to its dense network of fiber-optic connections.”
Sustainability
According to Koncewicz, companies are prioritizing locations with a low risk of natural disasters to ensure uninterrupted operations. Proximity to end-users is another critical factor in selecting data center locations, as it significantly impacts latency, service delivery, and overall user experience.
Data centers strategically positioned near key user bases, such as manufacturing regions or financial districts, can offer faster, more reliable access to data and services. In some cases, however, data center workloads are not sensitive to latency and therefore, building data center campuses in rural areas is more economical.
Grid Connectivity
Some markets will have insufficient power to serve data centers. But another consideration is the ability to transmit power to the right places, according to Marc Ganzi, CEO of DigitalBridge, a firm that invests in data centers.
As a result, some firms are building additional transmission lines as well as substations to serve the grid. However, it can take two to four years to connect new facilities to the grid while it only takes about 18 months to construct a new data center.
Here, a key tenant of real estate rings true: gaining the advantage is a game of information arbitrage.
All things being equal, companies that utilize data most efficiently will benefit the most. Said another way, the organization that acquires the best data at the right time will win. But getting the best data requires information about substation locations, connectivity to power plants, and even new battery storage locations.
Bringing together these four component factors will help to identify regions where data center locations are optimized for reliability and sustainability. Since these components consist of limitations based on geography, the use of location intelligent solutions will allow for an easier identification of the proximity relationships that will indicate the convergence of these key factors listed above. In fact, this is a task where geospatial information and technology have always excelled to help point opportunities for a competitive advantage.
Location Intelligence Growth Solutions
The CoreLogic statements and information in this blog post may not be reproduced or used in any form without express written permission. While all the CoreLogic statements and information are believed to be accurate, CoreLogic makes no representation or warranty as to the completeness or accuracy of the statements and information and assumes no responsibility whatsoever for the information and statements or any reliance thereon. CoreLogic® is the registered trademark of CoreLogic, Inc. and/or its subsidiaries.
