By Jason Guck, Delta Edge CI
Most 2027 energy budgets are being built right now, and most of them are being built the same way they were built last year. Someone pulls the trailing twelve months of spend, applies a percentage escalator that came from a utility newsletter or a broker email, and moves on to the next line item. The number is defensible in the sense that it can be explained in a sentence. It is also, in a meaningful share of cases, wrong by enough to matter.
The problem is not the arithmetic. It is that a single blended escalator assumes your energy cost is one thing moving at one rate. It is not. It is at least four things moving in different directions, and in 2027 two of them are moving against each other in a way that a blended number will hide.
The commodity is forecast down while the delivered price goes up
Start with the two published numbers that matter most for next year.
In its September 2026 Short-Term Energy Outlook, the U.S. Energy Information Administration forecasts the Henry Hub natural gas spot price at $3.43 per million Btu in 2026, falling to $3.28 in 2027. Over the same period, EIA forecasts the average residential retail electricity price rising from 18.2 cents per kilowatt-hour in 2026 to 18.6 cents in 2027, after 17.3 cents in 2025.
Read those together. The fuel that sets the marginal price of electricity across much of the country is forecast to get cheaper, and the delivered price of electricity is forecast to keep rising anyway.
That is the whole argument for building a budget from measures rather than averages. If your escalator is anchored to a commodity view, it will point the wrong way. The increase is not in the molecule. It is in delivery, capacity and the fixed charges that fund system buildout, and those components do not respond to a good year in the gas market.
There is a demand-side reason for that. EIA forecasts commercial sector electricity sales increasing 3.3% in 2026 and 2.7% in 2027, with data center development and manufacturing activity named as the drivers. Load growth of that size does not get served out of spare capacity. It gets served out of new transmission, new generation and capacity procurement, and those costs land on the delivery side of every bill in the affected territory, including yours, whether or not your own consumption grew at all.
Four lines, not one
A budget built from measures separates the bill into components that behave differently, then forecasts each one on its own logic.
1. Supply, in dollars per kilowatt-hour
This is the part a contract or a market view actually governs. If you are on a fixed supply agreement, this line is known with certainty through the contract end date, and the only forecasting question is what happens after it. If you are on utility default service or an index product, this is the line where a commodity view belongs, and it is the only one.
2. Delivery, in dollars per kilowatt-hour
This is set by tariff, and tariffs change through rate cases with published schedules. You do not have to guess. The filings are public, they are made months ahead, and they state the requested increase and the classes affected. A budget that treats delivery as an unknown is leaving free information on the table. I wrote about reading those notices in a piece on fall utility rate filings, and the practice is worth building into the budget cycle rather than the complaint cycle.
3. Demand, in dollars per kilowatt
This is the line most budgets handle worst, because it is not driven by how much energy you use. It is driven by your highest single interval of the month, typically fifteen minutes. Two facilities with identical annual consumption can carry demand charges that differ by thirty percent because one of them starts everything at once at six in the morning and the other staggers.
Forecasting this line as a percentage of last year’s total spend is what causes budgets to miss. Forecast it from the measure: pull twelve months of billed demand by month, identify the peak-setting months, and ask whether anything planned for 2027 changes them. A new production line, an added shift, a chiller replacement, a tenant fit-out. Each of those is a demand event before it is an energy event, and it shows up on the bill the month it happens rather than gradually.
4. Capacity and ancillary, in dollars per year
In capacity markets your charge for the coming year is set by your load during a handful of system peak hours in the prior summer. That means a meaningful part of your 2027 cost was determined by what your buildings were doing on a few August afternoons in 2026, and nothing you do in 2027 will change it. This line belongs in the budget as a known quantity, not an estimate, and the ability to move it belongs in a separate conversation about next summer. The mechanics of participating in that market, rather than only paying into it, are covered in our post on demand response.
A worked example
Take a portfolio spending $1.2 million a year on electricity, split roughly 55% supply, 25% delivery, 15% demand and 5% capacity and fixed charges. That split is typical for a multi-site commercial operator in a regulated territory, though it should be pulled from actual bills rather than assumed.
The blended approach applies, say, 4% to the whole thing and budgets $1,248,000.
The measure-built approach says: supply is fixed under contract through September 2027, so 55% of the spend rises 0% for three quarters and carries an open-market assumption for one. Delivery has a rate case on file requesting an increase to the relevant class, so that 25% gets the filed number rather than a guess. Demand is flat except at two sites adding a shift, where billed kilowatts are expected to rise, so that 15% gets a site-specific figure. Capacity is already set from last summer’s peaks and is a known line.
The two totals may land close to each other. That is not the point. The measure-built version tells you which quarter the exposure sits in, which two sites drive the variance, and which single decision, renewing supply early or managing next summer’s peaks, changes the number. The blended version tells you nothing you can act on, and when it misses, it offers no explanation beyond the escalator being wrong.
What to ask for before the budget closes
Three requests, none of which requires a project or a capital decision.
- Twelve months of billed demand in kilowatts by site and month, not just dollars. This is on every bill and is rarely in the spreadsheet.
- The end date and rate of every supply contract, with the renewal window marked on the calendar rather than discovered by a letter.
- Any rate case on file affecting your service classes, with the requested increase and the expected effective date.
An organization that has those three items can build a 2027 energy budget it can defend line by line, and more importantly can see where the controllable money sits. In most portfolios it sits in demand and capacity, which are the two lines a blended escalator makes invisible.
Energy is one of the few operating costs where the forecast and the intervention are the same exercise. Pulling the numbers apart to budget them properly also tells you where to work. If you want that analysis run against your own bills before the budget closes, that is the kind of work we do at Delta Edge CI.