Insights

From scrolling through social media to joining a virtual meeting, streaming a show or navigating with GPS, so much of modern life is dependent on data centers. This constant digital demand places continuous pressure on the infrastructure that powers our connected world. 

In today’s power-constrained environment, data centers and developers face significant challenges. Power architecture has become one of the biggest drivers of schedule risk, shaping everything from site selection to long-term system resilience. In addition, project delays can trigger significant opportunity costs and contractual impacts, making early clarity around design priorities a critical first step in these advanced facilities. 

Here are a few important factors to consider when planning and designing data center infrastructure in the current utility landscape.

Power Constraints Shape Data Center Outcomes

The data center industry is no longer primarily driven by servers or cooling capacity, but by when and how power can be secured. Utilities are increasingly unable to deliver the scale, timing or reliability that data centers need, forcing owners to rethink traditional grid-dependent approaches or behind the meter self-generation.

Power strategy directly impacts site selection, schedule risk and when the facility goes online, not just operating costs. Understanding upstream utility limitations is now essential to reducing risk and protecting project schedules. Discussions with utility providers about committed versus planned capacity and identifying offsite upgrades and dependencies that could affect critical path, and comparing them to the cost and complexity of behind the meter generation, are essential to understanding the true energy availability of the project site.

Key Takeaway: Owners must evaluate power architecture top to bottom as a strategic business decision that directly affects timing, cost and long-term operational success. 

Redundancy Alone Doesn’t Ensure Reduced Risk

Redundancy models improve reliability, but more redundancy does not always mean less risk in real world operation. Two common models illustrate the tradeoffs:

• N+1 provides one additional unit (or module) beyond what is required to carry the critical load. For example, one extra generator, UPS module or cooling equipment.
• 2N provides two independent systems or paths, each capable of carrying the full critical load. Depending on the design and operating model, these may be implemented as active/active or active/standby.

While 2N systems can offer more reliability, they can increase complexity and cost with the introduction of additional equipment, controls and potential failure points if they are not commissioned thoroughly, maintained rigorously and operated with disciplined procedures. 2N also doesn’t guarantee smooth transitions during outages and maintenance or protect against misalignment between design assumptions and real-world operations. 

The goal is not necessarily maximum redundancy. It is to identify which redundancy topology is the best fit for the facility’s operating model considering staffing, procedures, maintenance philosophy and tolerance for complexity.

Key Takeaway: Reliability targets must align with operational expectations, not just specification language or industry assumptions. 

Design for Restart, Not Just Steady State

Uptime is not only about staying online. It is also about how predictably you can recover when something goes wrong. Many systems are only designed to perform well once stable, but not all systems are engineered and tested for restart after a partial or complete outage. 

Complex architectures with multiple redundant paths can struggle during restart due to sequencing, timing, control-system dependencies or load-step behavior. Restart issues often reveal themselves during real events when time and pressure expose gaps that do not appear in design documents.

Questions data center owners should ask include:

• Has this system been designed, commissioned and validated for credible restart scenarios?
• Are restart sequences documented and practiced?
• Does the operations team have clear, actionable guidance to execute a restart confidently?

Key Takeaway: True resilience requires data center power systems to be engineered and validated for restart, not just operation under ideal design conditions. 

Prioritize Resilience and Restartability

For modern data centers, rethinking power architecture is no longer optional. Power constraints, operational complexity and accelerating digital demand are reshaping how data centers must be planned and delivered. When owners prioritize utility constraints into early decisions, align redundancy with operational realities and optimize for restart capability, they reduce risk and develop capacity that communities and businesses increasingly rely on. 

For more information about B&N’s data center expertise, contact Ryan O’Keefe, PE.