Recently, we launched Nyquist licensing programme, which brings together selected Reactive Technologies capabilities across measurement, data and intelligence, making them available for integration into future energy technologies and partner solutions.
XMU2 is one key component of Nyquist technology stack as the upcoming reference hardware combines high-fidelity sensing, precise time synchronisation and NVIDIA-powered edge computing to capture and interpret increasingly complex grid behaviour.
We recently piloted XMU2 on a live data centre connection and the results provide a detailed view of how modern loads are reshaping the grid.
Capturing over half a million 24-bit samples per second per channel, XMU2 analysed supraharmonic emissions near the data centre with exceptional time and frequency resolution.
These fast, high-frequency disturbances are produced by power-electronic equipment such as UPS systems, server power supplies and power converters and cooling infrastructure. The emissions fall within the 2-150 kHz supraharmonic range. XMU2 measures this range as continuous 24-bit waveform data at over half a million samples per second per channel, preserving phase, timing and transient structure while remaining GPS-time-aligned to sub-microsecond accuracy across sites.
Persistent emissions around 76 kHz are clearly visible in the full-range spectrogram captured near a data centre using Reactive Technologies’ XMU2 prototype.
A 76.3 kHz supraharmonic disturbance detected near a data center in the low-voltage network. Recurring 140 ms bursts introduce significant high-frequency distortion, illustrating the emerging challenges that modern power-electronic installations pose to power quality monitoring and grid visibility.
XMU2 is our next generation measurement device, combining the wideband capability with sub-microsecond GPS time stamping. This enables precise correlation of fast electromagnetic phenomena across locations at a level of fidelity that is difficult to achieve with conventional measurement approaches.
How XMU2 Helps Data Centres
Supraharmonics can contribute to additional losses, thermal stress and accelerated ageing of on-site transformers and electrical equipment. With this level of measurement fidelity and timing accuracy, issues that would otherwise remain hidden become clearly visible. This enables better design decisions, more informed filtering strategies and reduced long-term operational risk.
How XMU2 Helps Grid Operators
These emissions can propagate into distribution and transmission networks, where supraharmonics can contribute to asset stress, excite resonances and lead to long-term degradation in performance and asset life. Sub-microsecond-level time alignment allows operators to identify sources, interactions and propagation paths with much greater confidence.
XMU2 + NVIDIA GPU: richer, faster signals for early detection and decisive action
XMU2 + AI reflects our patented architectural approach to power system analytics. Intelligence is no longer centralised, it is distributed, embedded and executed at the point of measurement, where the richest and fastest signals originate.
This edge-native paradigm is designed to capture the true nature of modern grid dynamics, particularly in environments dominated by inverter-based resources, data centres and high-frequency phenomena.
By bringing computation directly into the grid, XMU2 + AI enables:
- Wideband supraharmonic measurement beyond traditional harmonics
- 512 kS/s per channel at 24-bit resolution to capture fast electromagnetic phenomena
- Sub-microsecond GPS time stamping for precise cross-site correlation
- GPU-accelerated edge analytics using NVIDIA technology
- Real-time interpretation of power grid measurements
- Immediate detection and contextualisation of localised disturbances and anomalies, including supraharmonic distortions
- Ultra-low-latency analytics, independent of centralised processing constraints
XMU2: A reference architecture for AI-powered grid intelligence
Beyond the device itself, XMU2 also demonstrates a reference architecture for AI-powered grid intelligence.
Through Nyquist, elements of this architecture, combining high-resolution sensing, NVIDIA GPU-accelerated edge computing and scalable cloud analytics can be adapted and embedded into a broader range of partner technologies, from advanced PMUs and WMUs to next-generation smart meters.
This provides a pathway to extend high-fidelity measurement and edge intelligence across the grid, enabling greater visibility into fast electromagnetic phenomena across the network.
If you’re exploring how to bring AI and high-frequency measurement together in your own solutions, we’d be interested in discussing potential collaboration.
