Eaton ePDU G3 Advanced Metered Input 3-Phase 60A PDU

The Eaton ePDU G3 Advanced Metered Input 3-Phase 60A PDU is a next-generation power distribution unit engineered for mission-critical data center environments requiring real-time power visibility at the outlet level. Built on a three-phase architecture, this unit supports 60A input capacity and delivers dynamic load balancing across multiple high-power consumption devices typical in GPU-accelerated and AI computing deployments. The G3 platform incorporates advanced metering electronics that provide granular power consumption data, enabling operators to optimize cooling efficiency, prevent thermal hotspots, and enforce capacity governance across densely populated equipment racks.

Designed specifically for hyperscale and high-performance computing workloads, the ePDU G3 Advanced supports daisy-chaining of up to 32 units and integrates seamlessly with Eaton's Intelligent Power Manager software suite for centralized monitoring and alerting. Each outlet circuit is independently monitored for real-time voltage, current, power factor, and watt-hour consumption, allowing facility teams to track power draw at sub-circuit granularity. This visibility is critical for identifying equipment anomalies, predicting component failures, and optimizing rack-level power allocation in environments where individual GPUs or accelerators may draw 400+ watts continuously.

Launched in 2023-Q2, the G3 Advanced represents Eaton's response to the explosive growth in AI infrastructure density, where traditional metering at the PDU input level proves insufficient for workload management and capacity planning. The unit features hot-swap intelligent outlets, redundant power supplies, and Ethernet-based connectivity for both local and remote management via industry-standard protocols.

Ideal for

• GPU cluster power monitoring in AI training and inference facilities, enabling per-device power budgeting and thermal workload distribution across 48-96 high-power accelerators per rack
• Hyperscale data center capacity planning and chargeback operations, tracking power consumption by tenant, application, or business unit for accurate utility cost allocation
• High-density HPC (High-Performance Computing) rack deployments requiring real-time circuit-level alerting to prevent oversubscription and automatic load shedding during peak demand periods
• Machine learning operations (MLOps) infrastructure monitoring, correlating power draw patterns with job execution schedules and model training phases for operational insights
• Multi-tenant colocation environments where precise per-outlet metering ensures fair resource allocation and enables transparent billing based on actual consumption rather than provisioned capacity
• Disaster recovery site power validation and load testing, using detailed metering data to confirm adequate cooling capacity and identify bottlenecks before mission-critical workload migration

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.