Member briefing | Energy strategy, flexibility, and market developments
AI and cloud growth are changing the scale, shape, and urgency of electricity demand. For energy-intensive companies, the practical response is not panic: it is better forecasting, flexible load, and clearer procurement choices.
For years, many energy strategies were built around a familiar pattern: annual consumption, seasonal peaks, a few efficiency projects, and a supplier tender. AI data centres change that pattern. They bring large, dense, and relatively constant electricity demand into grids that were already adapting to electric vehicles, heat pumps, electrified production, and more variable renewable generation.
The important point for Scholt members is not that every company will run a data centre. It is that grid operators and suppliers now have to plan around loads that can be the size of small cities. That changes connection queues, local congestion, contract structures, and the way flexible capacity is valued. A company that understands its own load shape will be better prepared than one that only sees energy as a commodity invoice.
The current electricity discussion is less about whether global electricity demand is growing and more about where and when it is growing. IEA analysis points to strong demand growth in 2025 and 2026, with data centres and AI as visible drivers alongside manufacturing and broader electrification. That means local networks can feel pressure even when national averages look manageable.
For a member with sites across multiple regions, this matters. A warehouse, charging hub, production line, or cooling installation in a congested area may face different risks than the same load elsewhere. The commercial question becomes: can the site shift, buffer, curtail, or forecast demand in a way that makes it easier to connect and cheaper to operate?
Traditional procurement focuses on price per megawatt-hour, contract length, volume risk, and sustainability attributes. Those still matter, but they are no longer sufficient. If a grid area is constrained, flexible load can become a commercial advantage. The ability to move consumption away from expensive or stressed hours may reduce exposure to imbalance costs, network charges, and volatile wholesale prices.
This is why metering quality is becoming strategic. A business cannot sell or use flexibility if it cannot prove the baseline, the response, and the operational constraints. Interval data, meter associations, forecasts, and audit trails are not administrative detail anymore. They are the evidence layer for a smarter energy contract.
First, map large loads by site and by process. Identify what is critical, what can be shifted, what can be buffered with thermal storage or batteries, and what has a fixed production dependency. Second, check whether the metering setup reflects operational reality. If submetering, main meters, forecasts, and consumption reports are not aligned, flexibility conversations will stall.
Third, bring energy and operations teams into the same planning cycle. The question is not only whether a lower price is available. The question is what type of load the company wants to present to the market. A site that can provide reliable forecasts and controlled flexibility is a different customer than a site that only asks for more capacity during the same peak hours as everyone else.
AI data centres may look like a technology-sector story, but the knock-on effect is an energy-system story. More large loads increase the value of flexibility, transparency, and timely consumption data. For members, the near-term opportunity is to make energy planning more operational: connect meter readings, forecasts, documents, and site-level decisions so the business can react before grid pressure turns into cost pressure.
In the next procurement cycle, the most valuable question may be simple: what parts of our consumption are movable, measurable, and worth optimising? Companies that can answer that question will have more options as electricity demand continues to grow.