ABB: How to Stop Data Centres Overloading Electricity Grids AI is reshaping the data centre landscape, but its most immediate impact is being felt on energy systems. As demand rises sharply and fluctuates more unpredictably, maintaining grid stability is becoming a central concern. As operators race to deploy new AI capacity, focus is shifting beyond servers and racks to the strength and responsiveness of the underlying power infrastructure. In response, ABB has launched a pre-engineered modular synchronous condenser package designed to support grid stability requirements while shortening deployment timelines. The development comes amid rapid global AI adoption, which is driving up the consumption of electricity and making load patterns far more volatile. These evolving demand profiles present challenges for voltage and frequency control, particularly as large-scale facilities draw power from already constrained grids. How AI is changing grid requirements The expansion of AI data centres is placing mounting pressure on electricity networks, which must remain stable despite sharp and sudden changes in load. Industry forecasts suggest data centre capacity could nearly quadruple by 2030, intensifying the need for grid reinforcement and smarter energy management. Unlike conventional workloads, AI-driven processes can cause rapid spikes and drops in electricity usage. This variability can strain grid infrastructure and make it more difficult for new developments to secure timely grid connections. ABB’s new solution is engineered to operate at the grid connection point, addressing these challenges directly. The system delivers instantaneous inertia alongside dynamic reactive power, helping to stabilise voltage and frequency in real time. In practice, this allows it to absorb sudden load changes and limit the risk of wider network disturbances. David Bjerharg, Business Line Manager at ABB, says: “As data centres become increasingly widespread and AI-driven demand increases, grid stability is becoming a fundamental requirement for ongoing expansion. “This solution enables