Dear Readers,
Welcome to this Special Issue on Next-generation asset management: Monitoring, AI, and predictive maintenance.
Across the articles slated for this edition, there is a narrative shift. While topics range from interoperability standards to predictive analytics, the subtext is united by a call for engineering discipline over hype. The energy transition cannot be solved by plugging in new technologies and trusting algorithms to figure it out. Next-generation asset management draws resources to confront the hard physical and operational limits of our evolving grid infrastructure.
The data centre bottleneck and the BESS boom
The grid’s physical infrastructure is colliding with the digital world’s rapid expansion. Artificial intelligence, cloud computing, and data-intensive applications are driving unprecedented growth in data center capacity, introducing highly concentrated loads that outpace the long lead times required to build new transmission infrastructure. As Siemens Energy’s Greg Collison points out, data centres have become “grid critical infrastructure”. Bridging these capacity gaps requires flexible, fast-deployable substations and advanced transmission technologies to ensure stability.
Beyond base frequency load management, transient demands from data centres create unprecedented high frequency content that require more Battery Energy Storage Systems (BESS). When xAI’s Colossus supercluster engineers in Memphis realised the local grid lacked the stiffness for its GPU racks, they deployed a fleet of Tesla Megapacks as a “shock absorber” to shield the grid from fast, high-magnitude power transients. However, as Rafael Narezzi warns, these distributed, software-controlled assets create cyber-physical vulnerabilities. Because “speed of deployment without speed of security creates opportunity”, true next-generation asset management requires continuous, validated visibility to prevent these grid-active machines from becoming pathways for disruption.
Engineering discipline over hype
Relying on BESS and AI as quick fixes introduces vulnerabilities if not managed with engineering discipline. Hanane Oudli warns that treating BESS as a plug-and-play storage box rather than a grid-active machine is a fundamental mistake. Furthermore, Tony McGrail highlights the dangers of over-relying on “black box” AI, demonstrating that Large Language Models (LLMs) can confidently generate fabricated insights and spurious correlations. Both authors stress that in the context of next-generation asset management, AI monitoring and dashboards must be the last line of defence, not the first.
Engineering discipline also means confronting the physical and geopolitical realities of the grid. As outlined in my latest column, even universal, open standards like IEC 61850 are being quietly weaponised through bespoke national dialects and implementation guidelines, meaning true interoperability demands rigorous engineering one layer down at the implementation level. Furthermore, Nina Sasaki Støa-Aanensen reminds us that physical challenges of the transition remain paramount. Developing a reliable SF₆-free MVDC circuit breaker for multiterminal DC grids demands intensive physical testing and collaborative innovation, proving that new grid architectures require hard hardware solutions, not just digital overlays.
Harnessing AI for condition-based maintenance
Even though AI data centres source their own signature power quality problems, AI will be harnessed to implement Condition-Based Maintenance (CBM) and next-generation asset management.
Rather than AI replacing human experts, humans will deploy AI to filter massive datasets, flagging early-stage anomalies so experts could make informed decisions. Dr. Ki Yeoung Kweon underscores this, noting that platforms like Hyosung’s ARMOUR⁺ are critical for managing ageing infrastructure, though he stresses the engineering challenge of integrating these advanced analytics across legacy grid assets. We also see intelligence applied to asset design: Dr. Boya Zhang demonstrates how AI and generative models are actively being used to design the next generation of eco-friendly insulating gases, replacing slow trial-and-error with disciplined computational chemistry. Additionally, predictive maintenance in switchgear (Gaurav Joshi, Yasunori Ito) and Machine Learning in GIS (Bharat Nandula) illustrate the shift away from unreliable time-based maintenance toward continuous condition-based monitoring.
The energy transition…
is neither a technology problem nor a human problem: it is a systems problem. Next-generation asset management succeeds when we ground decisions in both physical constraints and engineering judgment.
The articles in this issue show how. The grid’s stability depends on it.
Douglas Maly
Editor-in-Chief
Switchgear Magazine
