Data Centers

Modular data centers (MDCs) are prefabricated off site, use standardized plug-and-play components, and enable simultaneous site preparation and module construction. Testing and certification to UL 2755:2025, Outline of Investigation for Prefabricated Modular Data Center Systems and Related Modular Units, enables evaluation of both the underlying subsystems and the complete MDC as a single integrated product before the unit is shipped. Once in the field, MDCs with UL 2755 certification typically require only evaluation of proper installation. In contrast, MDCs without UL 2755 certification may be subject to additional evaluations by local code authorities depending on jurisdictional requirements. As a result, UL 2755 certification can help streamline deployment and reduce time and resource demands related to approval processes.

Key sustainability challenges include:

• Managing rising energy consumption and greenhouse gas (GHG) emissions
• Water usage and efficiency, especially in drought-prone areas
• Integrating renewable energy and circular practices
• Meeting evolving sustainability reporting requirements, including the EU Corporate Sustainability Reporting Directive (CSRD) and Energy Efficiency Directive (EED)
• Demonstrating the accuracy of environmental claims
• Clearly communicating environmental impact to stakeholders

Time-to-power refers to the duration required to establish and operationalize electrical infrastructure at a new data center site. It is a critical factor in site selection, often outweighing financial costs, because it directly impacts how quickly a data center can become operational, which is crucial for meeting business demands, controlling costs and maintaining a competitive advantage.

Hyperscale data centers are massive, centralized facilities operated by large technology companies to support global-scale workloads, especially AI and cloud computing. They prioritize scalability, rapid deployment and high-density computing.

Edge data centers are smaller, distributed facilities located closer to end users or devices, enabling low-latency processing for applications such as IoT, 5G and real-time analytics.

Enterprise data centers are typically owned and operated by individual organizations for their internal IT needs, often featuring legacy systems. As a result, enterprise data centers strongly focus on modernization, compliance and sustainability.

Air cooling uses chilled air circulated through racks and rooms. Although it is cost-effective, it’s less efficient for high-density AI clusters.

Liquid cooling (including direct-to-chip and immersion cooling) uses fluids to absorb and dissipate heat directly from components, offering superior thermal management, energy efficiency and reliability for high-performance workloads. Liquid cooling is increasingly specified for hyperscale and AI clusters.