Back when data center workloads followed relatively stable patterns, offering a degree of predictability, UPS systems were designed to provide a reliable safety net when the unexpected happened.

Sitting quietly behind the scenes, batteries were rarely called upon but expected to perform when a power disruption threatened uptime.

But in the new era of higher power densities, faster load changes, and demand for dynamic energy management, today’s complex landscape is challenging long-standing power strategies that saw batteries as passive components, just waiting on standby for an emergency.

Now, instead of a simplistic insurance policy, batteries are quickly becoming critical advantages in modern data center operations, expected to support increasingly dynamic environments while balancing power delivery, lifecycle performance, and sustainability objectives in tandem.

Tong Xu, vice president of sales, EMEA at CSB, frames this new power landscape as an active energy system, with batteries acting as core components of operational strategy.

“This is a big shift,” says Xu. “Batteries used to be almost purely standby. Now, operators expect them to handle more frequent use – whether that’s for energy optimization, peak shaving, or grid interaction.”

As approaches to UPS design are reshaped, Xu explains why the industry isn’t just asking whether batteries can provide sufficient backup during an outage, but considering how these systems perform across a wider range of modern operational scenarios.

The end of predictable power

“AI and HPC have changed the power dynamic completely,” says Xu. “They need much higher power density, faster ramp rates, and the loads are far more dynamic than before. Traditional setups were very predictable, but now workloads can spike really quickly.”

For UPS systems, the challenge isn’t just about delivering more power, but delivering it instantly, consistently, and reliably. This new reality is forcing the industry to reconsider the role of batteries within the wider power architecture.

As AI clusters grow increasingly energy hungry and rack densities continue to increase, it’s all hands on deck. Battery systems should now be considered as part of a broader strategy for maintaining performance and resilience.

Adding further complexity to modern power architectures is the fact that even a brief disruption can have significant consequences in AI environments. While traditional workloads may have tolerated a short interruption or extended recovery period, AI systems offer no such breathing space.

As the pressure for always-on, reliable operations continues to rise, batteries will increasingly be pushed from the wings and into the fore as an essential tool in the data center arsenal.

Delivering more for less

Simultaneously, inside high-density AI environments, operators require greater rack density without proportionally increasing physical footprints. This makes each and every element of the power chain, including battery systems, part of the wider efficiency equation.

“The more kilowatts they can put in, the better,” says Xu. “That’s why kilowatts per square feet has become a more and more important metric.”

However, higher density doesn’t just mean increasing capacity. Battery solutions must also be optimized for the specific way power is required. In many AI deployments, operators are less focused on maximizing runtime and more interested in delivering significant amounts of power for shorter periods.

“We’re seeing a very clear move towards higher kilowatts per rack, especially with AI-related activity,” explains Xu. “But runtime is no longer about maximizing duration. It’s much more application-driven. Operators want higher power, but for a shorter duration.”

This transition places new expectations on battery technology. Instead of optimizing for a single discharge event in an emergency, batteries must now demonstrate strong cycling capability, resilience under partial charge conditions, and consistent performance over repeated usage.

For operators, this means battery selection has become a more complex engineering decision. Factors such as operating profile, energy strategy, and expected lifecycle behavior all need to be considered alongside traditional requirements like capacity and runtime.

Selecting the right technology for the job

As battery requirements become more diverse, the idea of a single ‘best’ technology no longer makes sense. Each individual facility has different priorities, and the right solution depends heavily on how the battery will be used.

Key considerations include required power and energy profiles, available footprint, expected lifetime, total cost of ownership, and cycling requirements. A battery designed for a short-duration, high-power application will naturally have different characteristics from one intended for longer backup periods or frequent energy management use.

According to Xu, the most effective approach begins with understanding specific operational objectives rather than leading with a specific technology.

“There’s a common misconception that all batteries are basically the same. People think a battery is just a battery. But it’s much more than that, especially when we talk about data centers.

“We don’t want to simply tell customers, ‘use this battery, it’s the best.’ It doesn’t work that way. It’s about having the right conversations to understand their background, profiles, and goals to then help them identify the right solution.”

This application-focused approach allows operators to balance competing priorities across power density, runtime, and longevity.

Yet while each of these considerations are valuable, attempting to maximize every factor simultaneously can lead to unnecessary complexity or compromise. Here, balance is key.

“In an ideal world, you’d achieve all three,” says Xu. “But actually, you need to select one priority. Start with what really matters most – power, runtime, or lifespan – and then design around that factor.”

AI-ready solutions

Recognizing the changing requirements of modern data centers, CSB’s range of battery solutions – including its Calor XHT, XPL, and XHRL series – are specially designed to address the variation in operational needs that characterize today’s energy landscape.

The XPL and XHRL series focus on balancing performance and longevity, providing high-rate discharge capability while supporting extended design life. These characteristics make them suitable for operators looking for dependable performance across longer operational periods.

The Calor XHT series, meanwhile, is designed specifically around higher power density and short-duration discharge requirements, addressing the growing demand for compact, high-performance solutions in AI and hyperscale environments.

“The Calor XHT is a real standout right now,” explains Xu. “It’s optimized for higher power density and short-duration discharge.”

The technology is also designed to support operation in higher-temperature environments. In increasingly interconnected data centers where AI workloads now steer not just power approaches but facility design and cooling strategies, this represents a distinct advantage.

“Data center temperatures have risen dramatically from previous levels,” says Xu. “Traditional VRLA batteries might not be able to handle the heat as effectively.”

Moving forward, battery systems must evolve alongside changing approaches to cooling and energy efficiency to maintain reliable performance under new and dynamic environmental conditions.

Reliable and sustainable at the core

Despite rapid transformation taking place across the industry, one requirement remains unchanged: reliability. No matter how advanced the technology becomes, the fundamental purpose of a UPS system is to protect operations.

“At the end of the day, the battery is what ensures uptime,” says Xu. “If it fails, the whole system is at risk. Consistency and predictability throughout the entire lifecycle is what operators are looking for.”

This focus on lifecycle performance is also changing how operators evaluate battery investments. Rather than looking only at initial cost, operators are increasingly considering the wider impact of maintenance requirements, replacement cycles, and potential operational disruption in the decision making process.

As well as delivering all-important reliability, understanding the full lifecycle of battery technologies is increasingly essential in reducing the environmental impact of data centers.

For CSB, this means looking beyond simple assumptions about materials and considering factors such as recyclability, resource use, and end-of-life management.

Lead-based batteries, for instance, benefit from a mature recycling ecosystem developed over decades. CSB’s established collection and processing infrastructure supports a circular economy approach by allowing key materials to be recovered and reused.

“The battery itself is highly recyclable,” adds Xu. “The casing is plastic and can be recycled, the AGM material can be recycled, and the lead – which makes up around 70 percent of the battery – can be recycled and reused.”

Xu believes education remains an important part of the sustainability discussion, particularly when it comes to comparing different battery chemistries:

“People need to understand the facts. Different chemistries have different advantages, and each solution comes with its own considerations when reaching end of life.”

Rather than focusing on a single sustainability metric, operators must consider the complete lifecycle impact of their chosen technology and how it aligns with their wider operational goals to strike the right balance.

Ahead of the curve

The next generation of data centers will continue to demand more from their supporting infrastructure. Higher rack densities, smarter energy systems, and increasingly complex workloads will require batteries that can deliver more than backup support.

“Looking forward, operators are looking for higher power density, better cycling capability, faster recharge, and smarter integration with energy systems,” says Xu.

Yet while the role of batteries is changing, the fundamental requirement remains constant. Reliability will continue to define success in mission-critical environments where downtime is not an option.

“One thing won’t change: reliability,” adds Xu. “Without reliability, you get dark data centers – and that’s something nobody wants.”

By understanding operational priorities and identifying the right technology for the job, batteries can move beyond their traditional role as emergency backup and transition into active contributors to performance, efficiency, and sustainability, representing a key strategic advantage in the AI era.