Do not adjust your set: Energy volatility is here to stay

For too long most industrial enterprises have run energy and operations as two separate worlds. In these plants, production teams as they go about manufacturing product regard power as a fixed cost that is centrally negotiated once a year. This is an increasingly inefficient – and needlessly expensive – way to do business.  

Putting a wall between energy and operations might have made sense when energy was reliably available and priced within a predictable range. During this period – roughly from the mid-1980s until the start of this decade – there was relatively little to gain from managing it hour by hour, much less second by second. 

But that is no longer the case: energy is no longer the stable, always-available resource it once was.  

The scale of the change is striking. The rise of AI and data centres – one of the defining growth stories of our industry – is on course to more than double electricity demand from that sector by 2030, alongside the wider electrification of industry and transport. This is, in large part, a story of progress. But it also reshapes the operating environment. In early 2025, average electricity prices in New York City rose 132 percent year-on-year, driven not by war or sanctions but by weather, storage and tighter supply. In n the United States over 2,600 GW of renewable and storage capacity now sits in interconnection queues, waiting on a grid racing to keep pace with demand.  

Importantly, energy risks such as climate-driven disruption, demand growth, grid fragility, cyberattacks and gridlock are no longer singular issues that can be mitigated against individually, for example with clever hedging strategies or with added backup for grid failure – or even by putting in place contingency plans for supply. Instead, they are increasingly interconnected and able effectively to amplify one another. 

In this respect, the risks are like dominoes: a wildfire in one region melts transmission lines and cuts power to another.  AI workloads spike demand faster than grids can respond. A cyberattack on a substation knocks out cooling systems in a data centre. The challenge now isn't just managing impact but managing the interdependence of these events. Clearly, responding both rapidly and effectively to a multifaceted risk environment requires clear-headed organisational unity. But for too many plants this unity is made impossible to achieve by internal fragmentation of operations – for example, entirely different teams overseeing energy data and process data. As a result, critical decisions that depend upon both get reconciled several weeks after the fact, usually in a report.  

Think of a chemical plant when the power price spikes unexpectedly at two in the afternoon. The energy team might spot the spike immediately, but production runs on unchanged, because the plant floor won’t receive the signal in a form it can act upon. Consequently, the operation runs for hours at the wrong cost. Finance might be able quantify that loss beautifully a month later, but it cannot prevent it. 

This is the shift I now hear in almost every customer conversation. Companies no longer want simply to see their energy data. They want to be empowered to act on it while it is happening.