By now the standout advantages of BESS for data center developers are known. Vendors like Energy Vault, Eolian, Fluence, Hithium, ZincFive and more are supplying their systems to data centers while on the other side of the coin AI hyperscalers like Alibaba, Tencent, ByteDance, Meta, Google, and Microsoft have been investing in BESS.
Speed-to-power and cheaper power per megawatt-hour (in comparison to gas) are the two biggest areas BESS wins at, and these are highlighted in a report by Volta Foundation in which the organization’s AI & Data Center Infrastructure committee examined where, how, and why BESS are attractive and unattractive for data centers.
“BESS earns its place alongside firm generation,” said Volta Foundation’s report, adding that flexibility is a battery’s main benefit for data centers and grid operators. Besides the areas where BESS wins, Volta Foundation singled out four areas for improvement (where batteries are competing but not yet winning) and two where it is losing opportunities to other technologies.
“BESS wins the jobs that reward speed and daily cycling: peak shaving against demand charges, and bridge power as the battery layer of a gas-or-engine hybrid,” said the report, pointing out that although batteries cannot solve every problem, they are uniquely positioned to solve some of the most urgent issues.
In peak shaving and demand-charge management, for example, BESS can discharge through the costly hours. However, its benefit here depends on a region’s rate structure. In areas with no demand charges, its strong economic advantage disappears.
BESS provides temporary, or ‘bridge’ power for data center assets while they wait in an interconnection queue – the speed-to-power piece. But standalone batteries rarely carry a bridge alone, and Volta’s report identifies fuel cells as a competitor for bridge power jobs.
Meanwhile, BESS competes where software tools and power electronics have caught up – i.e., curtailment flexibility, uninterruptible power supply (UPS) ride-through, grid services, and sub-second power quality.
Losing out and problems faced
Big batteries remain uncompetitive in the jobs that demand sustained energy, long-duration backup, and continuous prime power. Long-duration backup jobs beyond four hours go to diesel and continuous prime power is still primarily served by gas turbines and fuel cells.
Volta Foundation noted that the use case for batteries for data centers could receive a boost as regulators continue to push for flexibility. Similarly, changing power architectures inside the data center and gas turbine supply chain shortages are also pushing operators towards batteries.
The next few years will see more data center deployment globally, and Volta Foundation’s AI & Data Center Infrastructure committee is calling for more collaboration between developers, battery and integration companies and grid operators to define a load profile the industry can design for.
As the report pointed out, the lack of load visibility is one of the biggest stumbling blocks for power suppliers trying to meet data centers’ demands, and batteries could provide value more effectively with a greater understanding of data centers’ power needs.
There is also no clearly defined way to integrate batteries into a data center, and that lack of certainty presents a risk for investors. Problems arising during the deployment stage tend to appear mostly at commissioning and during the first two years of operation. Integration, controls, and dispatch governance are the biggest culprits rather than cell chemistry.
As for future developments, Volta’s report said the next two to three years will decide whether batteries are remembered as a bridge the industry crossed or a fixture it built.
“If curtailment-based interconnection becomes the norm and duration economics keep improving, then installed BESS also earns a standing operational role in shaping load, shaving peaks, and selling grid services for the life of the facility,” the report stated. “However, if queues clear faster than expected and gas supply loosens, more of it reverts to idle backup,” the authors warned.
In any case, however, batteries provide an arbitrage-economics-independent value by serving as a facility’s power management layer. But how much of this value is realized depends on both industry and regulators’ next steps.
Hyperscalers such as Google and xAI seem to prefer vertically integrating their storage, and if this pattern continues and the data center operator becomes the biggest buyer for grid-scale batteries, this could have consequences for how systems are specified and financed. The United States will be BESS’s biggest data center market and it is rewriting the rules for how these loads connect.
Volta Foundation concluded its report with a series of recommendations addressed to the stakeholders involved. For data center developers and hyperscalers, it recommended that they design hybrid architectures from day one and specify the battery for daily duty so that backup capability arrives as a byproduct. Finally, because speed decides a project’s viability, hyperscalers should rank power options on time-to-power alongside levelized cost.
Grid operators and policymakers should standardize curtailment-based interconnection. They should clarify dispatch rates and market participation for batteries serving large loads and standardize large load tariffs to avoid confusion and delays.
Lastly, battery and integration companies should increase emphasis up the stack and develop products based on load archetype. For example, some loads call for high-power cells while others call for energy-dense cells.
The post How competitive is BESS for data centers, really? appeared first on Energy Storage.
