Data Centers Drive Interest in Alternative Battery Technologies for Energy Storage Solutions

Data centers could bring alternative battery types into the mainstream, developers say

Artificial intelligence data centers have specific energy requirements, and industry experts suggest that this is fueling renewed interest in energy storage technologies that have not gained traction with utilities.

According to industry professionals, the significant power demands of data centers on short notice may pave the way for established yet underutilized energy storage technologies. For instance, XL Batteries recently announced a contract to pilot a 333-kW organic flow battery at a data center operated by Prometheus Hyperscale in Wyoming. Other energy storage companies are also making similar announcements regarding their battery technologies.

Eos Energy Enterprises, known for its zinc-based batteries, is in discussions with a major developer in the data center sector. While data center operators are keen on renewable energy—which typically necessitates some form of energy storage—the decision to move forward with XL Batteries’ flow battery pilot was primarily driven by the need for a non-flammable technology that can quickly respond to substantial shifts in power demand, as explained by XL Batteries CEO Tom Sisto.

As more data centers seek to secure large power supplies swiftly, this could open doors for mature energy storage technologies that have yet to be widely adopted. Although most projects remain small-scale pilots as operators test new solutions, XL Batteries already has numerous potential buyers lined up, anticipating that the Prometheus project, scheduled for completion in 2027, will meet expectations. Prometheus has already agreed to purchase up to two 12.5 MW units, one in 2028 and another in 2029, contingent on the pilot’s success. Sisto remarked, “Once this is shown to be working, we’re probably already sold out.”

### A Demand for Innovation

Data centers have various unique power requirements that may push them towards exploring new forms of energy storage. Sam Allen, mission-critical director for technical innovation at Burns & McDonnell, an engineering and construction firm involved in several data center projects, noted that traditional cloud-based data centers generally have steady, predictable loads. In contrast, AI-driven data centers often experience sudden, significant load fluctuations, leading to more complex power needs.

Data centers, like other large infrastructure projects, face permitting and interconnection challenges. This has led operators to seek independent power sources that are both flexible and reliable, especially in light of their inconsistent power demands, according to Tisha Scroggin-Wicker, decarbonization director at Burns & McDonnell.

While lithium-ion batteries remain the most commonly discussed solution among Burns & McDonnell’s data center clients, alternatives such as flow batteries are garnering increased attention due to their ability to handle a high number of charge-discharge cycles with minimal performance degradation over time. This makes them particularly suitable for applications where durability and cycling capability are critical, as well as for sites experiencing frequent load fluctuations. Nickel-zinc batteries are also gaining traction due to their high power density, improved safety profile, and environmental benefits. Additionally, supercapacitors may be appealing for their capacity to deliver large amounts of power over short periods.

“There is a clear appetite for innovation, and a desire to explore beyond any single technology, especially when it comes to future-proofing infrastructure and supporting broader energy goals,” Allen stated.

### Use-Case Driven Decisions

While data centers show interest in behind-the-meter energy storage that could connect to the grid and participate in virtual power plant programs, most newly constructed data centers have not pursued this route. Many of Burns & McDonnell’s clients that have installed onsite energy storage are pairing it with dedicated generation to expedite bringing the data center online compared to utility involvement.

Burns & McDonnell has observed data centers expressing interest in acquiring and operating existing energy storage projects, regardless of the technology used. This includes lithium-ion projects, flow batteries, compressed air energy storage, and other long-duration energy storage solutions. These technologies could enable data center operators to better manage higher power densities, integrate more renewable energy, and ensure a reliable, flexible power supply—all on a relatively short timeline compared to grid connections.

However, it remains uncertain whether any particular energy storage technology will dominate among data centers. Jason Abiecunas, executive vice president of business development at FlexGen, a software provider for energy storage systems, noted that data centers are interested in a diverse range of storage technologies. This diversity reflects the unique operational needs of data centers but requires FlexGen to adapt to a potentially more technology-agnostic landscape.

When it comes time to deploy an energy storage system, Abiecunas indicated that only a few new technologies are gaining traction. “What’s becoming increasingly clear is that decisions regarding battery chemistry are driven by specific use cases,” he explained. For instance, when energy storage is deployed to replace traditional generators or extend backup capabilities, there is a noticeable interest in long-duration chemistries.

Sisto believes data centers do not have a significant need for what the utility sector considers long-duration energy storage. He, like Abiecunas, has observed that interest is primarily driven by specific use cases. The cost of owning and operating an energy storage project decreases with more frequent cycling. He anticipates that storage technologies with a 36-48 hour range are most likely to resonate with data center operators, while seasonal or ultra-long energy storage technologies may not.

Currently, many of the options that interest data centers remain relatively untested in real-world scenarios. Most data center operators prefer not to be the first to implement unproven technology at scale. As a result, many prospective customers are hesitant, waiting to see how ongoing pilot-scale projects perform before committing to any agreements. This cautious approach could lead to a rush for the limited manufacturing capacity available for some of these technologies once one proves reliable, according to Sisto. “That’s the natural tension,” he commented. “Everyone would love to wait and see, but at that point, you may miss the opportunity.”