A high-voltage electric transmission tower seen from below against a clear blue sky

In short: Several US utilities facing a wave of very large proposed data center and crypto-mining loads have created a separate tariff class for them, and the defining feature of that class is not the rate per kilowatt-hour. It is a minimum demand charge: a commitment to pay for a large share of contracted capacity whether or not the facility ever uses it, backed by collateral and an exit fee if you leave early. None of that shows up when two sites are compared on cents per kilowatt-hour alone.

Why Large Loads Get Their Own Tariff

A standard industrial or commercial tariff was written for customers whose demand is small relative to the system and whose risk of leaving, or of never actually using what they signed up for, is manageable within normal rate-case assumptions. A data center or mining campus can ask for tens or hundreds of megawatts in a single interconnection request, which is a different kind of exposure. If the utility builds the substation, feeder and transmission upgrades to serve that request and the customer under-builds, delays, or walks away, the cost of that stranded infrastructure does not disappear. Under a standard tariff it is typically spread across every other ratepayer on the system.

That is the stated rationale behind the large-load tariff classes a number of state utility commissions have approved or had proposed to them over the last few years: keep the commercial risk of a speculative mega-load with the customer that created it, rather than with the residential and small-commercial customers who did not ask for a new substation. The mechanism that does this work is the minimum demand charge, and it is worth understanding before a term sheet is in front of you, not after.

The Core Commitment: Minimum Demand, Not Metered Demand

Under a conventional tariff, the bill tracks what you actually draw. Under a large-load tariff, the bill is built around a contracted demand figure negotiated before the facility is energized, and the customer agrees to pay for a defined minimum share of that figure every month for the life of the contract, regardless of whether the facility's actual load is lower. If contracted demand is 100 megawatts and the minimum demand percentage in the agreement is high, a facility that ramps slowly, that never fills its planned IT load, or that curtails heavily for economic or grid reasons can still be billed as though it were running close to the number it originally asked for.

This is the same economic idea as a take-or-pay clause in a power purchase agreement, but the counterparty is different and that difference matters. A PPA take-or-pay obligation runs to an independent generator or developer under a bilaterally negotiated contract. A large-load tariff's minimum demand obligation runs to the regulated utility providing retail delivery service, is typically filed with and approved by the state public utility commission as a tariff or an individually negotiated special contract, and its terms are shaped as much by what the regulator will approve on behalf of other ratepayers as by what the two commercial parties would otherwise agree. The result can look similar on a spreadsheet. The negotiating leverage, appeals process and public-interest framing behind it are not the same conversation.

Contract Term and the Ramp Schedule

Large-load tariffs are generally written as multi-year commitments, long enough that the utility's investment in dedicated infrastructure can be depreciated against a reasonably certain revenue stream. A facility rarely reaches its full contracted demand on day one, so these agreements typically include a ramp schedule: a defined path from a lower minimum demand percentage in the early years to a higher one once the facility is expected to be fully built out. Missing the ramp schedule, not just the final number, is itself commonly a trigger for additional charges, because a slower ramp than planned still leaves the utility holding capacity it built on the strength of the original schedule.

This sits downstream of, and is a separate question from, how long the interconnection study and network upgrades take in the first place. The queue decides when you can be connected at all and who pays for the upgrades that connection requires. The large-load tariff decides what you owe, as a floor, every month after you are connected, for the full term of the contract. A project can clear the queue on schedule and still be caught by a ramp schedule written before construction delays were known.

Collateral and Credit Support

Because a minimum demand commitment is only as good as the customer's ability to pay it, utilities offering large-load tariffs typically require credit support: a letter of credit, a parent company guarantee, or a cash security deposit, sized against some multiple of expected monthly billing under the contract. This is the same instrument family used in power purchase agreement collateral, where developers commonly ask for six to eighteen months of expected payments as security, but again the counterparty and the approval path differ. A utility's collateral requirement is generally set out in the tariff or special contract itself, filed with the regulator, and tied to the customer's credit rating rather than negotiated purely bilaterally. A newly formed project entity with no credit history should expect to post more collateral, for longer, than an investment-grade parent guaranteeing the obligation directly.

Exit Fees and Early Termination

The mirror image of the minimum demand commitment is the exit fee. If a customer terminates before the end of the contract term, or reduces its contracted demand materially, the tariff or special contract generally provides for a payment calculated to recover the utility's unrecovered cost of the infrastructure it built to serve that specific load, net of whatever can be resold to the next customer or reabsorbed into general system use. The exact formula varies by utility and docket, but the principle is consistent: the customer that caused a dedicated investment to be made is the customer who pays for it if that investment is stranded, rather than spreading the loss across the rest of the rate base.

This is the clause worth modeling pessimistically rather than optimistically. A project that signs a long-term contracted demand figure to secure favorable terms, and then scales down after market conditions change, can face an exit fee sized against the gap between what it contracted for and what it actually used, not just against what it formally terminates.

Special Contracts and Confidential Terms

Not every large load is served under a published tariff schedule. Where a load is large or unusual enough, state regulators often allow the utility to negotiate an individually tailored agreement, generally called a special contract or a negotiated service agreement, rather than fitting the customer into a standard rate class. Special contracts still require regulatory approval and are typically filed with the commission, but the specific commercial terms inside them, pricing, ramp schedule, collateral, exit fee, are frequently treated as confidential and are not published in the same way a standard tariff schedule is. That confidentiality cuts both ways for a prospective customer: it means competitors cannot see exactly what you agreed to, but it also means you generally cannot see what the utility has agreed to with anyone else before you negotiate your own terms.

What This Should and Should Not Decide

None of this should be read as a reason to avoid large-load markets. The utilities adopting these tariffs are, in most cases, the same ones with capacity to actually serve a large new load on a reasonable timeline, which is itself the scarcer resource right now. The point is to price the full commitment rather than the headline rate. A tariff with a lower cents-per-kilowatt-hour figure but an aggressive minimum demand percentage and a short, steep ramp schedule can cost more over the contract term than one with a higher nominal rate and a gentler structure, and that comparison cannot be made from the rate schedule alone.

It is also worth keeping the large-load tariff conversation separate from the voltage and service-level decision and from how the facility itself is financed. They interact, since a longer minimum-demand commitment can make project financing easier by de-risking the revenue line a lender is underwriting against, but they are negotiated with different counterparties on different timelines and should not be assumed to move together automatically.

A Checklist Before You Sign

  • Get the actual minimum demand percentage and the ramp schedule in writing, not a verbal assurance that it is "flexible."
  • Model the bill at the contracted minimum, not at your planned actual load, for every year of the term.
  • Ask whether missing the ramp schedule itself triggers a charge, separate from missing the final contracted demand figure.
  • Confirm the exact collateral instrument required, who can draw on it, and under what conditions it is released or reduced over the term.
  • Get the exit fee formula, not just its existence, and run it against a scenario where your load comes in meaningfully below plan.
  • Establish whether you are being offered a standard large-load tariff or an individually negotiated special contract, since the regulatory approval path and confidentiality differ.
  • Separate this conversation from the interconnection queue and network-upgrade cost questions, which are usually negotiated earlier and with a different part of the utility.
  • Keep this negotiation on the same timeline as your state and local tax incentive discussions, since all of them are usually decided before the site is finalized, not after.
  • Have counsel confirm whether the commission docket approving the tariff or contract is public, and if so, read it.

Frequently Asked Questions

What is a large-load utility tariff?

It is a separate electric service tariff, or an individually negotiated special contract, that some US utilities have created specifically for very large new customers such as data centers and crypto-mining facilities. Its defining feature is a minimum demand charge: the customer pays for a contracted share of its agreed capacity every month regardless of actual usage, which is intended to protect the utility's other customers from paying for infrastructure built for a load that never fully materializes.

How is a minimum demand charge different from a normal demand charge?

A normal demand charge is based on metered peak usage during the billing period, so a facility that uses less generally pays less. A minimum demand charge sets a floor independent of actual usage: if contracted demand is set at a given figure and the minimum demand percentage is high, the customer is billed close to that figure even if the facility's real draw is well below it. The floor, not the meter, sets the bill in a quiet month.

Is a large-load tariff's collateral the same as power purchase agreement collateral?

They are the same type of instrument, typically a letter of credit, parent guarantee or cash deposit, but they run to different counterparties. Power purchase agreement collateral protects an independent generator or developer under a bilateral contract. Large-load tariff collateral protects the regulated utility providing retail delivery service, and its terms are usually set out in a tariff or special contract filed with the state public utility commission rather than negotiated purely between two private parties.

What happens if we need to leave the contract early?

Most large-load tariffs and special contracts include an exit fee, calculated to recover the utility's unrecovered cost of infrastructure built specifically to serve that load, net of any amount it can recoup by reselling or reabsorbing the capacity. The exact formula varies by utility and by the specific docket or contract, which is why it should be requested and modeled before signing rather than assumed to be minor.

Does a lower per-kilowatt-hour rate mean a cheaper contract?

Not necessarily. The minimum demand percentage, the ramp schedule and the contract term all affect total cost over the life of the agreement, and a lower headline rate attached to an aggressive minimum demand commitment and a short ramp can cost more than a higher rate attached to a gentler structure. The full commitment, not the rate schedule's first line, is what should be compared across sites.

Power First, Then the Contract Terms

Rax Data & Energy provides colocation and hosting in the United States and the UAE. Tell us your target load and timeline, and we will set out what a connection actually commits you to before you negotiate it.

Contact Us See Pricing