Demand Charge Management for C&I: The Overlooked Storage Savings | TAICO

TK E-CUBE L200 · Liquid-Cooled C&I Energy Storage System

TAICO Energy Insight

By TAICO Editorial · Updated 2026 · 9 min read

In 2026, the investment logic of behind-the-meter storage is being re-examined. For years, most commercial and industrial (C&I) storage business cases revolved around a single model: charge at valley prices, discharge at peak prices, and pocket the spread. Simple, intuitive, easy to drop into an investment spreadsheet.

Then spot-market trial settlement expanded, and industrial users trading directly no longer fit neatly inside fixed time-of-use (TOU) tariffs. The once-stable peak–valley spread is becoming a spot price curve that moves with supply and demand.

Peak–valley arbitrage still matters — but its certainty is falling. For many large industrial users, another chronically undervalued line item may be more stable and easier to calculate: the demand charge. This article shows how demand charge management works, who benefits, and why the real battleground is the control system — not the battery.

Part 1
Read Your Bill First: There Is a "Rental" Line in It

A large industrial user's electricity bill is made of two parts.

• Energy charge — the actual kilowatt-hours consumed, multiplied by the applicable tariff.
• Demand charge (basic charge) — more like a monthly rental for grid capacity. You occupy transformer and supply capacity, so you pay for that resource, whether or not you use every watt of it.

The demand charge is typically billed in one of two ways:

• By transformer capacity. Based on the size of transformer installed — e.g. Fujian 24 CNY/kVA·month, Jiangsu 30 CNY/kVA·month. A plant with a 1,000 kVA transformer pays roughly 24,000–30,000 CNY every month.
• By maximum demand. Based on the highest power actually drawn in the month, usually recorded as a 15-minute average. The unit price is typically higher — e.g. Fujian 36 CNY/kW·month, Jiangsu 40 CNY/kW·month.

There is a critical detail hidden in this rule: one 15-minute spike can determine the whole month's demand charge. A plant runs steadily at 500 kW, then one production line starts up and pushes the 15-minute average to 1,000 kW. If that spike is recorded, the month's demand charge may be billed at 1,000 kW.

Worse, some regions penalize exceeding the contracted demand: if actual maximum demand passes 105% of the contract value, the excess can be charged at a multiple. For facilities with volatile loads, this is not a minor issue.

Part 2
How Storage Cuts the Demand Charge

Demand management works by discharging fast when a spike starts to form, capping the load the grid sees. The mechanism is direct: the storage system monitors plant load in real time. The moment load is about to break a preset threshold, it discharges immediately to cover that short-term power — so the metered maximum demand drops.

The savings formula is simple:

The beauty of this revenue stream is its determinism. It does not depend on spot spreads, and it does not require daily cycling. If a facility has short load spikes and storage shaves them, the entire month's demand bill comes down. Occasionally, a single well-timed 15-minute discharge locks in a month of savings.

Part 3
Which Facilities Benefit Most

Demand management is not for everyone. It suits load profiles with three characteristics — high, spiky, short:

Typical candidates include mechanical manufacturing, stamping, chemicals, steel, cement, cold-chain logistics, data centers, injection molding, welding and mining — where large equipment startups, kiln switching, compressor banks and refrigeration ramps create short sharp peaks.

If a facility's load is flat, or sits persistently high, the shaving headroom is small. Trimming a platform-shaped load requires a very large storage system, and the economics deteriorate quickly. So before sizing a project, do not look at annual consumption — look at the 15-minute load curve: how high are the peaks, how long do they last, how many times per month, and are they predictable?

Part 4
Why Demand Management Is Steadier Than Arbitrage

Arbitrage revenue comes from a price spread. Wide spread, high revenue; narrow spread, low revenue. Under spot markets, the cheap and expensive periods are no longer fixed — prices move with weather, load, renewable output, plant outages and market rules.

Demand management revenue comes from a billing rule. As long as the facility is billed by maximum demand, the rate is known, and every kilowatt shaved translates into calculable savings.

The difference in certainty is fundamental: arbitrage is like trading every day — deciding when to charge and when to discharge; demand management is like lowering a monthly rent — control the maximum demand and the fixed expense falls. In regions where spot prices are unstable, spreads are compressed and load spikes are obvious, demand management is often the floor of the project's revenue.

Part 5
The Hard Part Is the Control System — Not the Battery

Demand management looks simple on paper. In practice it is genuinely hard, for four reasons:

1. Response speed. The grid records a 15-minute average. Storage must act before the spike is recorded — discharge too late and the shave fails, the demand is locked in.
2. Prediction. The system must anticipate whether load will breach the threshold. Line startups, impact loads and cooling ramps are not always daily-regular — the controller needs historical data, real-time power and production rhythm to judge.
3. Energy allocation. Capacity is finite: keep reserve for a demand spike while still serving arbitrage. Spend it all on arbitrage today, and an afternoon spike is missed.
4. Avoiding over-discharge. Discharge too early or too much wastes energy, damages subsequent arbitrage, and adds cycling wear.

This is why the competitive battleground of behind-the-meter storage is moving from hardware price to EMS control strategy and operating algorithms. Buying the battery is only step one; whether it saves money depends on how it is dispatched.

Part 6
Stacking Demand Management and Arbitrage

Maximizing storage revenue is not an either/or choice. The better approach optimizes demand management, peak–valley arbitrage, demand response and ancillary services together.

The priority rule: facilities with high demand rates and obvious spikes should protect demand management first — the revenue is more certain, and missing one spike can cost the entire month's basic charge. During periods of unusually wide spreads, participate in arbitrage — but the system must keep enough energy and power in reserve for a possible spike.

The ideal case: one discharge earns multiple revenues. Suppose at 15:00 the plant load is peaking and the market price is also high. Storage discharging at that moment both shaves the maximum demand and avoids high-price purchases — saving the demand charge and the energy charge in a single action.

Where local rules allow, behind-the-meter storage can also join demand response or ancillary services. But regions differ on charging/discharging windows, settlement methods and market eligibility — revenues cannot simply be added up; each must be modeled under local rules.

Part 7
What to Check Before You Invest

1. The billing method. If the facility is billed by transformer capacity, direct demand-management savings are weak — first evaluate whether switching to maximum-demand billing is possible, then use storage to control it.
2. At least one year of 15-minute load data. Not monthly or annual kWh. Look at which days and hours the maximum demand occurs, how long it lasts, and whether production plans or equipment starts cause it.
3. The demand rate level. Rates differ by region (Jiangsu vs. Fujian, for example) and project economics differ sharply. The higher the rate, the greater the shaving value.
4. Configuration vs. the spike shape. Short spikes suit high-power, short-duration storage; long high-load periods need more capacity — model the economics carefully.
5. EMS capability. A system that only charges and discharges on a fixed schedule cannot do demand management. The controller must monitor in real time, predict load, hold dynamic reserve and respond fast.

Conclusion
Stop Counting Only Arbitrage

Behind-the-meter storage can no longer be justified on peak–valley arbitrage alone. As spot markets advance, the certainty of spreads falls and revenue models must be re-estimated. For many large industrial users, demand charge management may be the more stable and more important income stream.

A project worth investing in should be evaluated in this order: look at the load curve first, then the demand rate, then the TOU spread. Storage that shaves spikes, holds maximum demand, and still captures arbitrage — that is storage that actually earns. It is exactly the brief TAICO's TK E-Cube L200 is built for.

Stable. Smart. Sustainable.