We recently ran a poll on our KSTAR LinkedIn page asking data center professionals a direct question:

“What is the No.1 power threat at your data center?”

Some industry professionals cast their votes, and the results were striking. Two options tied at the top, while one registered zero votes — a result that tells its own story:

· Utility Outages – 43%

· High PUE & Energy Costs – 43%

· Scalability Bottlenecks – 14%

· Power Quality Issues – 0%

While utility outages have long been regarded as the primary threat to data center uptime, seeing energy efficiency rank equally, and scalability concerns appearing at all, reflects a meaningful shift in how operators are thinking about risk.

1. Why Utility Outages Remain a Top Concern

A power failure, however brief, can trigger service disruptions, data loss, equipment damage, SLA violations, and significant financial losses. As AI workloads scale up, the impact is amplified: GPU clusters run at near-maximum sustained utilization, making any interruption far more costly than it would be for traditional infrastructure.

This is why high-availability UPS systems, battery backup, and intelligent monitoring remain non-negotiable, and why the challenge has expanded beyond preventing blackouts to guaranteeing load stability as rack densities climb.

2. The Rising Threat: High PUE and Energy Costs

Perhaps the most significant finding is that High PUE & Energy Costs tied with utility outages — reflecting a structural shift in how risk is defined. Power is no longer just an availability issue. It is a profitability issue.

According to the IEA, data centers consumed approximately 415 TWh of electricity globally in 2024, with demand projected to grow substantially as AI infrastructure scales. Energy costs have become one of the largest and fastest-growing line items in a facility's operating budget, driving operators to focus on lowering PUE, improving power conversion efficiency, reducing cooling load, and maximizing infrastructure utilization.

A facility that remains online but runs inefficiently will face mounting financial pressure — and increasingly, regulatory pressure as sustainability requirements tighten across markets in Europe, the Middle East, and Southeast Asia.

What the 0% Result Tells Us

Power Quality Issues received zero votes — worth pausing on. It does not mean the risk is gone. Voltage sags, harmonics, and transients remain real threats. What it likely signals is that most professionals already regard power quality protection as a solved baseline, built into their UPS infrastructure, rather than an active concern. In that sense, 0% is a positive indicator.

3. Why These Two Challenges Are Connected

On the surface these seem like separate problems: one about reliability, one about economics. In practice, they share the same root. Aging UPS units, oversized but underloaded infrastructure, and poor power topology design all compound both risks at once: lower efficiency wastes energy, aging equipment raises reliability risk, poor architecture limits scalability, and excess heat drives up cooling demand.

Addressing one without the other creates a partial solution, which is why KSTAR's modular UPS and in-row cooling product lines are engineered to deliver high availability and high efficiency together, from initial deployment through capacity expansion.

4. What Our Poll Tells the Industry

Votes is a small sample, but the distribution maps closely to what we hear from customers across the Middle East, Southeast Asia, and Europe. Outage protection is the baseline expectation. Energy costs are fast becoming the differentiating concern. And Scalability Bottlenecks at 14% signals that the ability to expand capacity without a full infrastructure overhaul is moving up the priority list as AI demand accelerates.

The most competitive data centers in the AI era will not simply be the ones that stay online. They will be the ones that stay online efficiently, and scale without compromise. We will be diving deeper into the topics covered above in our upcoming content — what is the biggest challenge you are facing in design, deployment, or operations? What topics or insights would you like us to cover next?