Eaton, a leader in smart power management, has launched a pioneering feature to detect significant energy spikes—known as AI power bursts—in AI computing systems. This capability comes as a firmware update to Eaton’s Power Xpert Quality (PXQ) event analysis system, enabling edge-level detection of subsynchronous oscillations (SSO) in data centers. By delivering this as a firmware upgrade, Eaton provides operators with a powerful tool to protect critical infrastructure and boost resilience amid rising AI energy demands.
AI data centers face a growing challenge: sudden, unpredictable power surges caused by GPU-heavy AI workloads. These spikes can trigger SSO, a harmful electrical condition that endangers data center equipment and strains the electrical grid, which struggles with rapid load changes. If left unmanaged, SSO can cause costly damage such as transformer overheating and ferroresonance, leading to long outages and expensive repairs. Early detection and mitigation are crucial to prevent these risks.
Eaton’s PXQ platform is a versatile power quality meter used in switchgear, switchboards, and power distribution units. It features advanced edge analytics to identify power disturbances like sags, swells, and harmonics. With the new firmware upgrade, the PXQ can now also detect SSO events, giving operators the ability to act early and avoid long-term damage to data centers and the power grid.
This innovation is part of Eaton’s broader “grid-to-chip” approach, combining intelligent power distribution, backup systems, and digital tools to improve efficiency across AI facilities. Recent milestones include collaborating with NVIDIA to support 800 VDC power infrastructure for ultra-high-power racks and working with Siemens Energy to accelerate the design and delivery of integrated data centers.
Eaton will showcase its grid-to-chip solutions at Yotta 2025 in Las Vegas (September 8–10) at booth #319. Experts JP Buzzell and Peter De Bock, vice president of energy and cooling, will also present at the conference, sharing insights on addressing the increasing energy demands of AI data centers in today’s fast-evolving digital world.
