The rapid expansion of artificial intelligence increases computing capacity requirements and creates new challenges for electrical grids. AI server workloads can rise or fall almost simultaneously, causing electricity consumption to swing by megawatts within seconds. Such abrupt power changes can strain supply systems and may drive the need for costly grid upgrades if not managed.
Supercapacitors: fast response for short peaks
Unlike batteries, supercapacitors store energy primarily through electrostatic mechanisms, enabling very fast charge and discharge. They tolerate a high number of charge/discharge cycles with limited performance degradation. While not ideal for long‑term energy storage, they are well suited to handling short‑duration, sudden power peaks (peak shaving) and damping rapid supply fluctuations.
An analogy is vehicle shock absorbers: just as suspension smooths road bumps, supercapacitors reduce abrupt changes in power demand.
Benefits for data centers
- Smoothing short power peaks can lower stress on electrical infrastructure and improve operational reliability.
- With appropriate system design, available electrical capacity can be used more efficiently, potentially reducing the need for expensive network upgrades.
- Compact supercapacitor‑based systems can free up floor space for computing equipment.
- Integration can help reduce overall energy use and heat generation, and the technology can be retrofitted into existing data‑center infrastructure with suitable engineering.
Siemens and Skeleton Technologies: investment and cooperation
Siemens has invested in Skeleton Technologies, a developer and manufacturer of supercapacitors, indicating a strategic interest in the technology. This cooperation aligns with Siemens’s goal of supporting data‑center owners and operators across the facility lifecycle.
Siemens offers a range of digital tools to support planning and decision‑making, including simulation software, digital twins, and elements of the Simcenter, Teamcenter and Building X portfolios. These solutions help designers and operators model power systems, optimize configurations, and make better‑informed investment decisions.
Integrated energy systems and AI‑driven cooling
Siemens supports resilient power architectures that combine microgrids, renewable generation, modular solutions, battery storage and intelligent control systems. These integrated approaches aim to make data centers more reliable, energy‑efficient and cost‑effective.
Part of the development roadmap includes AI‑assisted cooling optimization, which can adapt cooling systems to current operating conditions to improve efficiency and reduce energy consumption.
Additional partnerships and technical elements
Siemens is collaborating with several technology companies to modernize data‑center infrastructure:
-
Rittal: joint work on modular power units intended for placement near servers in the white space, reducing transmission losses when the power source is closer to the point of use.
-
NVIDIA and Fluence: together they have demonstrated a prefabricated, modular reference architecture to simplify and speed up data‑center deployment and expansion; coordinated central management and energy‑storage design can also boost operational resilience.
-
Reinhausen: a joint development effort on a solid‑state transformer for network conversion to support high‑performance servers. One objective is establishing a stable 800‑volt DC supply, which, combined with appropriate infrastructure, could enable more efficient power distribution for compute‑intensive systems.
Conclusion
Supercapacitors offer a rapid, high‑cycle option to mitigate short, intense power swings in AI data centers. Siemens’s investment in Skeleton Technologies and its collaborations with Rittal, NVIDIA, Fluence and Reinhausen—together with digital planning tools—are intended to create modular, resilient and more efficient power systems. These efforts aim to reduce grid strain, improve reliability and limit the need for costly infrastructure upgrades.
