How many #datacenters are spread across the country?

Key Concepts

• Data Center Transparency: The difficulty in obtaining comprehensive public information about data centers due to corporate secrecy and trade secret exemptions.
• Air Quality Permits: A crucial regulatory mechanism used to indirectly identify and gather data on data centers, specifically concerning their backup generators.
• Generator Capacity as a Proxy: Using the permitted capacity of backup generators to extrapolate the power needs and scale of data centers.
• LLC Ownership and Disclosure: The requirement for large tech companies to disclose their subsidiary LLCs, enabling the identification of data center ownership.
• Comprehensive Data Center Tally: The creation of the most extensive count of U.S. data centers to date, based on air permit data.

The Challenge of Data Center Disclosure

The transcript highlights the significant challenge in obtaining a definitive public directory or official map of data centers. There is no single regulatory body or government agency responsible for this information. Companies are reluctant to disclose such details, leading to difficulties in acquiring records. This difficulty manifests as redacted records and requests being denied under the grounds of trade secrets.

Leveraging Air Quality Permits for Data Acquisition

A key insight revealed is that most data centers require backup generators to ensure continuous operation in case of grid failure. The installation of these generators necessitates applying for an air quality permit. This regulatory requirement becomes the primary avenue for researchers to gather information about data centers. The methodology involved filing public record requests for air permits issued to data centers for their backup generators in every state.

Extracting Information from Air Permits

These air permits provide critical details, including:

• Generator Capacity: This data allows for the extrapolation of the power needs of the data center.
• Ownership Clues: Permits can offer hints about the entity that owns the data center.

Case Study: Columbus, Ohio Data Center

An example is provided of a rapidly expanding data center near Columbus, Ohio, with at least 164 emergency generators permitted. The air permit for this facility was initially applied for by an LLC named Mellin Enterprises LLC. However, this LLC claimed a trade secrets exemption, leading to the redaction of most of the requested information.

Unmasking Ownership Through LLC Disclosure

The transcript explains that major tech giants are obligated to disclose any companies, including LLCs, that they own. By meticulously examining official records, researchers were able to uncover the true ownership of the Ohio data center. It was revealed that Mellan Enterprises LLC was not the actual owner; the facility is owned by Google.

Building a Comprehensive Data Center Map

This investigative process was replicated hundreds of times to construct a comprehensive map of America's data center industry. The map represents the most extensive tally to date, with each dot signifying a data center large enough to require a permit for its backup generators. These facilities are either already built or have been approved for construction by the end of 2024.

Explosive Growth of the Data Center Industry

The research identified a total of 2,240 data centers. This figure represents a substantial increase, nearly quadrupling the number of data centers recorded in 2010.

Conclusion and Key Takeaways

The transcript underscores the opacity surrounding the data center industry and the innovative methods required to gain insights. By utilizing air quality permits for backup generators as a proxy, researchers have managed to compile the most comprehensive count of U.S. data centers to date. This data reveals a significant and rapid expansion of the industry, with the number of facilities nearly quadrupling since 2010. The case study of the Ohio data center demonstrates how corporate structures and trade secret claims can be navigated to uncover ownership by major tech companies. The core takeaway is that while direct disclosure is limited, indirect regulatory pathways can be effectively employed to map and understand the burgeoning data center landscape.