When a data center forces major grid expansion, the infrastructure costs that follow get recovered over 40 years or more — and if that data center exits, downsizes, or was never fully obligated to pay, those unrecovered costs flow to whoever is left on the system. For Indiana manufacturers, hospitals, universities, and large commercial operators locked into their utility territory, that means you. Whether your utility's tariff and your state's rate case process actually prevent that from happening is the central question — and right now, most of them don't.

This post is for C&I operators — plant managers, facility managers, COOs, energy managers — with five-, six-, or seven-figure electric bills who cannot simply walk away from their utility service territory. You'll know, by the end, which tariff safeguards actually exist, which line items on your bill concentrate your exposure, what FERC's June 2026 orders do and don't cover, and what you need to do before the next rate case to protect your operation.

What Utility Cost Allocation for Data Centers Actually Is

The electric grid is not free infrastructure. Every substation, transmission line, and interconnection upgrade has to be paid for — and the entity that caused the need for that investment is supposed to bear the cost. That is the principle. The reality is more complicated.

When a large new load like a data center requires significant grid expansion, the utility builds the infrastructure and then recovers those capital costs through rates over the life of the asset — typically 40 years or longer. If the customer that triggered the investment stays on the system and pays its allocated share for the full recovery period, existing customers see minimal rate impact. The math works.

The problem is what happens when it doesn't go that way. A data center that exits the territory early, downsizes below its contracted capacity, or was never structured with an enforceable long-term obligation creates a gap. The infrastructure was built. The cost is real. But the customer causing it is gone or paying less than it owes. The remaining costs don't disappear — they get socialized across the rest of the rate base. That includes you.

The specific line items where this exposure concentrates are capacity charges, transmission adders, and system benefit charges — the broadly socialized components of your bill, not the usage components. On your bill, these may appear as separate riders or rolled into a demand-based charge depending on your utility's tariff structure. The point is that these are the line items that don't scale neatly with how much electricity you actually consumed. They reflect system costs, and when system costs increase because of grid buildout for loads that don't fully pay, you absorb the difference.

Why This Exists on Paper vs. How It Works in Real Life

On paper, the cost causation principle is foundational to utility regulation: the customer or class that causes a cost should pay for it. Regulators designed cost-of-service studies and rate cases around this principle. It is not a controversial idea in regulatory theory.

In practice, three things break it down.

First, the scale of data center load growth is genuinely unprecedented. The U.S. is in the strongest four-year stretch of electricity demand growth since 2000, driven primarily by data centers. Dominion in Virginia reported 40 gigawatts of potential contracted data center demand in its service territory. Georgia Power's pipeline grew from 16 gigawatts to more than 34. Ohio utilities are forecasting up to 30,000 megawatts of data center peak demand against current peak demand of roughly 9,000. These numbers are not projections buried in planning documents — they flow directly into capital plans that land in your next rate case.

Second, tariff design hasn't kept pace. Most utility tariffs were not written with this scale of concentrated large-load growth in mind. The safeguards that would protect existing customers — minimum contract terms, collateral requirements, exit fees, minimum billing demand provisions — are inconsistently present across the country, and many territories have none of them.

Third, the information gap is enormous. Most C&I operators have no visibility into their utility's IRP pipeline, what large loads are under contract, what the capital commitments behind those loads look like, or what the tariff actually says about cost recovery if those loads exit. The rate case comes, rates go up, and the reasoning is buried in 10,000 pages of regulatory filings.

When Utility Cost Allocation Protections Actually Help Facilities Like Yours

The five safeguard categories that protect existing customers from data center cost cross-subsidies are documented. RMI reviewed 65 state-level tariffs and identified each. Here is what protection actually looks like — and where it exists today.

Minimum contract terms. Present in 22 of 65 tariffs reviewed. These align the large load's financial commitment with the cost recovery period of long-lived assets. Kentucky Power set a 20-year minimum for new loads of 150 megawatts or more. Dominion introduced a 14-year minimum. Georgia's Public Service Commission approved contracts up to 15 years for data centers in January 2025. The critical gap: even 20 years falls short of a 40-year asset lifespan, so remaining costs can still land on other customers when the contract ends. Minimum terms reduce the risk. They don't eliminate it.

Minimum monthly billing demand. Present in 24 of 65 tariffs. This provision ensures the large load pays even if it uses less than it contracted. Typical designs set the minimum at 80 to 90 percent of either contracted capacity or a rolling peak. The mechanism only protects existing customers if the demand charge rate actually reflects the cost of serving large loads — a design assumption that is worth verifying, not assumed.

Collateral requirements. Present in 37 of 65 tariffs. These require large new loads to post security — cash, letter of credit, or bond — against the infrastructure costs they're committing the utility to build. If the load exits, the collateral covers some portion of the stranded cost before it hits the rate base.

Exit fees. Present in 31 of 65 tariffs. These function as the back-end version of collateral — a financial penalty for leaving before cost recovery is complete. They create a financial disincentive for early departure and recover some stranded costs when a customer does leave.

Capacity reassignment. Present in only 12 of 65 tariffs. This allows a departing customer to transfer contracted capacity to a successor, limiting stranded costs for remaining customers. It is the least common safeguard and the most elegant solution to the stranded cost problem — because it keeps the infrastructure obligation whole rather than just penalizing departure.

If your utility's territory has most of these in place, your exposure is meaningfully reduced. If your territory is missing three or four of them, that is your rate case argument.

When Utility Cost Allocation Is a Serious Problem for Your Operation

The scenarios where your exposure is highest are not hypothetical. They are playing out now in specific geographies.

Your utility has significant large-load pipeline without enforceable tariff safeguards. Virginia, Georgia, and Ohio are the clearest examples at scale. If your utility has disclosed a multi-gigawatt data center pipeline in its IRP but lacks minimum contract terms and exit fees in its tariff, the cost exposure to existing ratepayers is structural, not theoretical.

Your state's rate case process doesn't give industrial customers meaningful intervention rights. Rate cases fight over costs already committed. If industrial customer groups are not organized and active in your state's PUC proceedings, the cost allocation decisions that affect your bill for decades get made without your input.

Your utility is not subject to FERC's June 2026 show cause orders. FERC's authority stops at wholesale. Vertically integrated utilities in the Southeast, Northwest, and desert Southwest outside CAISO face no federal mandate on large-load cost allocation reform. That is where significant data center development is proposed — including Arizona, Utah, and Georgia. If your utility operates outside the major RTOs, do not assume federal reform pressure is doing any protective work for you.

Your facility cannot absorb rate increases because margins are already compressed. This is the practical reality for most of the operators I work with. A manufacturer or a hospital cannot relocate because its utility decided to socialize 40 years of data center grid costs across the rate base. Relocation is the least available lever for most of you. That makes what happens in your IRP proceeding and your rate case extraordinarily consequential.

The wholesale price signal is already moving. A CSIS analysis found wholesale price increases of up to 267 percent in specific data-center-heavy zones under peak conditions. That is a local-peak figure, not a system-wide average — but it is directionally consistent with what happens when large-load growth outpaces grid capacity in a concentrated geography.

Vendor Pitches, Red Flags, and Questions That Smoke Out BS

The vendors most likely to appear in your building after you start asking about data center cost exposure are self-supply and microgrid vendors — and some of them will model this risk in ways that support their pitch more than they support your decision-making.

Here are questions worth asking before you act on any analysis that cites this risk as a reason to invest:

• What specific line items on my current bill are you attributing to cross-subsidy exposure, and what is your source for that allocation? If the vendor can't point to specific tariff sheet sections and rate case filings, the number is made up.
• Have you modeled both scenarios — with and without FERC's contested transmission charge treatment for behind-the-meter loads? FERC's February 2025 rejection of the Susquehanna–Amazon co-location interconnection agreement indicated that behind-the-meter co-located loads may still owe transmission and ancillary charges on a gross demand basis. That interpretation is under active litigation. If a self-supply proposal doesn't model both outcomes, the economics may be built on a legal assumption that doesn't hold.
• What is your assumption for the rate trajectory over the project payback period, and what data center pipeline data is that based on? A vendor projecting rate increases without tying that projection to your utility's specific IRP disclosures is guessing.
• What happens to the project economics if the large-load pipeline in your territory doesn't materialize at the scale currently disclosed? Utilities regularly update IRP projections. Contracted demand can be withdrawn or restructured. A project that only pencils out if the worst-case rate trajectory occurs is a project with undisclosed risk.

Questions to ask your utility or a regulatory consultant:

• Does your current tariff include minimum contract terms, collateral requirements, exit fees, minimum billing demand provisions, and capacity reassignment for large new loads?
• What large-load pipeline has your utility disclosed in its most recent IRP, and what is the current status of those contracts?
• Is there an active proceeding at your state PUC on large-load cost allocation — and which industrial customer groups have intervened or are eligible to intervene?

What You Can Do This Week

Pull your current bill and identify the socialized components. Look for capacity charges, transmission adders, and system benefit charges. These are the line items where cross-subsidy exposure concentrates. If you can't identify them by name on your bill, request a tariff sheet explanation from your utility's key accounts team — that is a standard request and they are required to provide it.

Find your utility's most recent IRP filing. In Indiana, NIPSCO and AES Indiana both file IRPs with the IURC. These are public documents. Look for the large-load pipeline section — the numbers are there. If you can't locate it, your utility's regulatory affairs team can direct you to the filing.

Benchmark your utility's tariff against the five safeguard categories. Minimum contract terms, minimum billing demand, collateral requirements, exit fees, and capacity reassignment. This is not a regulatory deep-dive — it is a checklist. A regulatory consultant or industrial customer group can run through it in a single call.

Find out if your state has an active proceeding on large-load cost allocation. North Carolina launched a proceeding in 2025 specifically on how to fairly integrate large electric load additions. Georgia requires regulatory approval before a data center over 100 megawatts can connect. These are the templates. If your state is not doing this yet, the industrial customer group that represents your sector is the place to start.

If self-supply is on the table, model transmission charge exposure under both FERC interpretations. The Susquehanna–Amazon litigation is not resolved. Get two numbers — one with transmission and ancillary charges applied on a gross demand basis, one without — before you commit to anything.

The Bottom Line on Utility Cost Allocation for Data Centers

If your utility's territory is absorbing large-load growth and your tariff lacks enforceable safeguards — minimum contract terms, collateral, exit fees, minimum billing demand, and capacity reassignment — you are exposed to decades of cost recovery for infrastructure you did not cause. The line items that carry that exposure are the broadly socialized components of your bill: capacity charges, transmission adders, and system benefit charges.

FERC's June 2026 show cause orders are meaningful pressure, but FERC's authority stops at wholesale. Your state public utility commission is the actual decision-maker on whether you absorb these costs. And CSIS scored the six major grid operators at roughly 20 percent sufficient against FERC's reform categories — this reform cycle is multiyear and piecemeal. You cannot wait for it to filter down.

The rate case and the IRP are where cost allocation gets decided for your facility. The window to act is before capital is committed, not after it is built and in the rate base.

Frequently Asked Questions: Utility Cost Allocation for Data Centers

Q: How do data center grid expansions raise electricity costs for Indiana manufacturers?

A: When a data center requires major grid infrastructure — transmission upgrades, substation builds, interconnection — those costs are recovered through rates over 40 years or more. If the data center exits early, downsizes, or was never obligated to pay for the full cost recovery period, the unrecovered costs are socialized across the remaining rate base. For Indiana manufacturers locked into their utility territory, that cost lands on capacity charges, transmission adders, and system benefit charges on their electric bill.

Q: What utility tariff safeguards protect C&I customers from data center cost cross-subsidies?

A: RMI identified five safeguard categories across 65 state tariffs: minimum contract terms (in 22 tariffs), minimum monthly billing demand (24 tariffs), collateral requirements (37 tariffs), exit fees (31 tariffs), and capacity reassignment provisions (12 tariffs). Territories with most or all of these in place provide meaningful protection. Territories missing three or four of them expose existing C&I customers to stranded infrastructure cost when large loads exit or downsize.

Q: What is a minimum billing demand provision and why does it matter to industrial customers?

A: A minimum billing demand provision requires a large load — like a data center — to pay for a minimum percentage of its contracted capacity even if it actually uses less. Typical designs set the floor at 80 to 90 percent of contracted or rolling-peak demand. For existing C&I customers, this matters because it ensures the customer causing the infrastructure cost continues paying toward it even during low-use periods, reducing the portion that would otherwise be socialized to the rest of the rate base.

Q: What should Indiana manufacturers do before their next utility rate case?

A: Pull your current bill and identify capacity charges, transmission adders, and system benefit charges — the broadly socialized line items where cross-subsidy exposure concentrates. Locate your utility's most recent IRP filing and check the disclosed large-load pipeline. Benchmark your utility's tariff against the five safeguard categories. And find out whether an active PUC proceeding on large-load cost allocation is open in Indiana, then identify which industrial customer groups are positioned to intervene.

Q: Does FERC's June 2026 large-load order protect my facility from absorbing data center grid costs?

A: Not directly. FERC issued show cause orders to major RTOs and ISOs on June 18, 2026, requiring that large loads bear the risk and cost of network upgrades — but FERC's authority stops at wholesale. Your state public utility commission determines how wholesale costs are sub-allocated to retail customers. FERC reform is also piecemeal: CSIS scored existing RTO tariffs at roughly 20 percent sufficient against FERC's reform categories. Additionally, vertically integrated utilities in the Southeast, Northwest, and desert Southwest outside CAISO are not subject to these orders at all.

Q: What bill line items expose Indiana C&I operators to data center cross-subsidy risk?

A: The highest-exposure line items are capacity charges, transmission adders, and system benefit charges — the broadly socialized components of your bill that don't scale directly with how much electricity your facility consumed. Usage charges (kilowatt-hours) are more directly cost-caused. The socialized components are where infrastructure costs for large new loads get distributed across the rate base when those loads don't fully pay for what they caused.

If this episode got you thinking about the rate case exposure your facility is carrying — or an energy or infrastructure decision you're evaluating in light of where utility costs are heading in Indiana — the TEG Energy Decision Blueprint is the starting point. We look at your bills, your interval data, and your operational realities, and we give you a full written opinion: whether the numbers actually work, what you may not have been told, and what the decision looks like once the modeling is done correctly. No obligation beyond that conversation.

For more context on the infrastructure decisions that interact with cost allocation risk, the posts on C&I microgrids for industrial facilities and utility standby charges for on-site generation are worth reading alongside this one — both are directly relevant if you're evaluating self-supply as a hedge against rising socialized costs.