The Price Signal That Changed the Decision
EU electricity prices for energy-intensive industries in 2025 averaged more than twice US levels and nearly 50% above Chinese levels, according to IEA data. That gap, similar to 2024, moved energy procurement out of the operations manual and into the boardroom, and it now sets the baseline for industrial on-site generation economics. A German chemical plant and a Texas refinery face fundamentally different cost structures before any production decision is made. When the price of a megawatt-hour differs by that much, buying from the grid stops being the obvious default. On-site generation becomes a question of engineering and economics, not corporate sustainability messaging.
What On-Site Generation Actually Includes
On-site generation covers a range of assets that operate behind the meter. Rooftop solar converts available space into electricity, though output follows weather and daylight. Gas-fired cogeneration produces electricity and useful heat together, which suits facilities with steady thermal loads. Diesel or natural gas generators provide backup during outages. Battery storage shifts cheap grid power or on-site solar output into expensive hours. Each option targets a different part of the bill: energy volume, peak demand, or outage risk.
The choice is not a technology contest. A food plant that needs continuous steam may find cogeneration attractive because one fuel serves two purposes. A warehouse with high cooling demand and a large roof may prefer solar paired with storage. An electronics fab that cannot tolerate voltage dips may justify generators for reliability alone, even if their energy cost is high. Most industrial sites end up with a blend of assets, not a single machine. This variety is part of why distributed energy resources rarely look the same from one facility to the next.
For any behind-the-meter asset, utilization drives cost. A rooftop solar array sized to meet annual average demand may produce far more than the site can use on sunny weekends, unless storage is added. A cogeneration unit that runs only when steam demand is high may have a lower capacity factor than assumed. On-site generation economics are sensitive to these operating patterns. A feasibility study built on average prices can mislead. The analysis must model hourly profiles, not annual totals.
Energy, Demand, and Reliability Are Separate Value Streams
Industrial electricity bills rarely turn on energy consumption alone. Many large customers also pay for the highest demand they set during a billing period. A single 15-minute spike can raise that demand charge for months, depending on the tariff. On-site generation can reduce both energy and demand, but the value depends on which part dominates the bill. Solar cuts energy purchases while the sun shines. A battery or load control shaves the peak, which is often the more valuable service on a demand-heavy tariff. The details of how that peak is measured are explained in the article on industrial demand charges.
Reliability value is harder to quantify. A process interruption can cost far more than the electricity that was not used. Some operators run generators only for outages and accept the fuel cost because lost production is more expensive. Others build on-site generation as core supply and treat the grid as backup. That decision changes the entire economic case. It also changes maintenance, staffing, and fuel logistics in ways the energy-only calculation misses.
Why Negative Prices Change the Trade-Off
Wholesale prices in many markets increasingly fall below zero during hours of low demand and abundant renewable generation. IEA data shows negative prices became more common across many markets in 2025, with the Nordic region and California as exceptions where responsive demand and battery storage reduced them. Negative prices signal a lack of system flexibility. For an industrial buyer, they are only a signal. Whether a plant can turn them into value depends on how its retail tariff is structured. Where wholesale settlement is passed through, or where a site has a responsive-demand arrangement with its supplier, a factory that shifts load into negative-price hours can be paid to consume, or at least avoid the cost of buying. A factory on a flat tariff, or one that cannot shift load, still pays its network charges and sees little of that value.
That flexibility gap is the real constraint. Many industrial processes are continuous and cannot pause when prices turn negative. A steel furnace, a paper machine, or a chemical reactor does not stop for a few hours without cost and risk. The divide between what wholesale prices do and what an individual plant can respond to is one reason behind-the-meter assets attract attention. Storage lets a site buy at negative or low prices and discharge later. But the battery must be sized and cycled enough to recover its capital cost.
Where the Arithmetic Favours On-Site Generation
Regional price differences explain much of the variation in on-site generation uptake. IEA data puts average EU wholesale electricity prices in 2025 roughly twice those of the United States and India, and above Australia and Japan. The EU average rose around 10% year-on-year to about USD 95/MWh, in line with higher natural gas prices. In that environment, the avoided cost from producing one megawatt-hour on site is high enough to support more capital. Where grid power is cheap, the same asset may never repay its investment. Identical equipment can look attractive in one country and uneconomic in another.
Carbon costs add another layer. EU ETS prices rose 15% in 2025, averaging around EUR 75 per tonne of CO2. A gas-fired cogeneration plant still emits, but its emissions per unit of useful output depend on the fuel and the efficiency of the process. A rooftop solar installation emits nothing at generation. Where the grid price embeds carbon costs, low-carbon on-site generation becomes more attractive. This is not universal: it depends on the fossil share of the local grid and the emissions intensity of the on-site alternative.
When the Grid Remains the Better Option
On-site generation is not automatically cheaper. How network costs are recovered varies by market and by tariff: in some jurisdictions they are spread across a broad customer base, while in others a factory that generates its own power still pays for grid connection, standby service, and often backup capacity. A poorly utilized generator can cost more per megawatt-hour than the utility tariff it replaces. Large industrial customers also need engineering staff to operate and maintain plants, which adds fixed cost. For many sites, the grid remains the simplest and lowest-risk option.
The comparison must be made at the margin. Each additional kilowatt-hour generated on site avoids a purchase from the grid, but it may also reduce the value of a demand charge reduction under some tariff designs. Where the grid has spare capacity and low prices, there is little to avoid. Where grid connection is constrained or the local utility imposes high standby rates, the balance shifts. The right answer differs by facility, tariff, and region. No single formula works everywhere.
What to Watch Next
IEA projections indicate world electricity consumption is set to increase strongly in 2026, driven by industry, appliances, cooling, data centres and electrification. That rising demand points to more load on grids that already face congestion in some regions. For industrial customers, higher grid prices and stronger peak signals tend to improve the economics of behind-the-meter assets. The current direction is toward hybrid arrangements, where on-site generation covers a base or peak portion and the grid supplies the rest.
Some industrial sites are also becoming participants in virtual power plants, where flexible load and generation are aggregated and dispatched as a single resource. That participation can create a revenue stream from flexibility that changes the on-site generation case again. A generator or battery no longer has to be justified only by avoided grid purchases. It can earn from grid services when the site does not need it. The boundary between a factory’s energy plant and a market participant is blurring.
References
- IEA Electricity 2026 – Prices: data on industrial electricity prices, EU wholesale price, negative price frequency, and EU ETS carbon prices.
- IEA Electricity Mid-Year Update 2026: global electricity consumption projections and demand drivers.
