Share this
by Jason Atwood, Laura Fletcher, and Greg Toothaker
Nearly every aspect of our digital lives funnels through the servers, storage systems, and networking apparatus housed in data centers. Whether you’re using an app on your phone, watching your favorite streaming service, sending an email, or accessing your company’s critical data, you and your devices interact with data centers thousands of times a day.
Let’s explore the current state of the US colocation data center market, the benefits of colocation data centers, and how they impact the power grid as a whole, affecting power traders, asset managers, and asset developers.
What Is a Colocation Data Center?
There are two broad categories of data centers: enterprise and colocation.
Enterprise data centers are private facilities owned and operated by a single company with massive amounts of data – the Googles, Amazons, Microsofts, Metas, and Apples of the world. They are dedicated solely to serving the needs of the owner.
Colocation data centers, also known as colos or multi-tenant data centers, are built by third-party developers who rent space (typically a rack, part of a rack, or a caged-off room) within the facility to multiple companies. Colocation customers vary in size but share the desire to locate their computing hardware and servers offsite. The business plugs in their equipment, and the colo provides the power, cooling, security, and necessary networking equipment.
The Current State of Colocation in the US
The US is the world’s largest colocation data center market, and demand for colo services in the country is seemingly insatiable. Valued at $12.4 billion in 2023, the market could more than double by 2029, reaching nearly $27 billion.
A recent report found that colo data centers in the US have just a 3% vacancy rate, despite the sector doubling its inventory in the past four years. Northern Virginia; Atlanta, Georgia; Dallas, Texas; Phoenix, Arizona; and Columbus, Ohio, are some of the hottest spots for data center development at the moment.
Banking and financial services, health care, media companies, and nearly every other segment of the economy are driving this voracious growth as companies move more data to the cloud and deploy new artificial intelligence-powered tools.
The Advantages of Colocation Data Centers
There are many reasons why businesses opt to house their data and computing equipment in a colo rather than in house. Here are just a few benefits of a colocation data center:
- Reliability: As mentioned above, the colocation provider provides reliable power, cooling, and high-speed network connections. Building out these systems in-house can be time-intensive and costly.
- Maintenance: Once companies build in-house data centers, highly skilled employees must maintain them. A colo has on-site IT professionals that monitor and manage servers, ensuring customers’ equipment is well cared for.
- Scalability: It’s much easier to scale your business by adding or subtracting rack space or bandwidth at a colo facility than to scale an on-premises (on-prem) facility.
- Security: Colocation data centers have strict protocols and systems to ensure data integrity. From video surveillance to fire suppression systems, colos are built to protect customers’ data.
- Risk Mitigation: Data centers have protocols for backing up data, which can minimize the impact of data breaches and natural disasters.
- Proximity: Companies can spread their data across multiple colos, ensuring computer systems are located near physical offices. This is particularly appealing to national and multi-national companies.
- Compliance: Regulations and laws around data center power use are evolving. Colos assume the burden of compliance with new requirements for their customers.
- Cost: Building, maintaining, and updating an on-prem data center is expensive. Colocation offers a cost-effective way for customers to support their computing equipment.
The Impact of Colocation Data Centers on Energy Traders and Power Grid Operators
The explosive growth of the colocation market has brought grid operators and energy traders both challenges and opportunities, particularly when it comes to power. Data centers, especially those that house the infrastructure driving artificial intelligence (AI) applications, require massive amounts of power.
AI-ready data centers have seen average power densities double in the past two years from 8 kW per rack to 17 kW. By 2027, AI workloads could demand rack densities of 30 kW. This jump was driven in part by the popularity of generative AI tools such as ChatGPT, which require more energy than traditional web searches.
To understand the scope, think of it this way: a typical data center requires roughly 200 MW of electricity – the equivalent of 40,000 households and one-fifth the capacity of a typical nuclear power plant. AI-ready facilities of 500 MW to more than 1,000 MW (1 GW) have already been announced.
The growing demand from data centers impacts not only a region’s existing power grid infrastructure but also the price of electricity. Tools like Yes Energy’s Infrastructure Insights and EnCompass help energy traders and utilities understand when and where these new data centers will be built, providing critical information for market forecasts and infrastructure planning.
Seeing the Impact of Data Centers on the Power Grid
Tool 1: Infrastructure Insights
Yes Energy’s Infrastructure Insights tracks colocation data center construction across the country, capturing the who, what, when, where, and why of each project, whether it’s announced, in development, or under construction.
As information is publicly available, the system provides details on the load value or size of the data center as well as the number of buildings.
Users can filter and visualize information on an easy-to-understand map, enabling them to identify trends, including where the next hot markets are emerging (as seen in the figure below).
Source: Yes Energy’s Infrastructure Insights
When selecting a particular data center from the map, users can see the project name, ownership, the reported in-service date, power capacity, project description, and more (as shown in the map below).
Source: Yes Energy’s Infrastructure Insights
Users can also filter Infrastructure Insights data to identify the top colo companies in the US and the estimated number of projects they’re developing.
Tool 2: EnCompass
EnCompass is Yes Energy’s powerful planning model that helps users forecast market prices, identify transmission congestion and constraints, forecast generator operations and renewable curtailment, and make optimal power supply decisions.
Utilities and energy traders use EnCompass to model the impact of a new data center on the price of electricity and the existing grid infrastructure. The model flags overloads and congestion at the transmission line, power plant, substation, or bus level, informing generation siting decisions for utilities and developers, as well as pricing strategies for traders.
That same information can also assist data center developers in their site selection process. By modeling the proposed load at multiple locations, developers can identify the best site for their new colocation data center. The system prioritizes the factors and base assumptions most important to the user, whether it’s the lowest price of electricity, reliability of the electricity supply, or something else.
In addition to analyzing existing infrastructure, EnCompass enables users to visualize the impact of new power system infrastructure, like a new transmission line or power plant, on a project.
When it solves, the EnCompass model always provides users with the locational marginal price (LMP), including the congestion and loss components used to determine the LMP. The platform also provides users with ancillary service prices.
Next Steps
Check out the Yes Energy Glossary for more information on the terms in this blog post.
Ready to see the impact of data centers near your trading or operating location? Contact Yes Energy to learn how Infrastructure Insights and EnCompass can optimize your simulations and enable better data-backed decisions.
About the author: Jason Atwood has experience in operations and engineering, generation and transmission planning, energy trading support, and market design. His work spans several energy sectors, including investor-owned utility, independent system operator, electric cooperative, and independent power producer. He is helping Yes Energy clients understand how our EnCompass solution can meet their needs.
About the author: Laura Fletcher is on the Yes Energy product team as an associate product manager. Prior to joining the team, Laura studied environmental engineering at Georgia Tech. She started working with energy data as a college intern and she has worked on various consulting projects, annual market forecasts, client relations, and database management.
