When you underwrite a stabilized multifamily or commercial asset, you grow the expense line by a blended three percent and move on. Utilities are a pass-through, or close to it. You almost never pull the utility’s pending rate case before you size the deal. That’s the miss. Over the next twenty-four months, the single largest swing in your controllable operating expense won’t be payroll or even insurance — it will be a decision made in a public utility commission docket about who pays for the data-center boom: the data centers, or you.

The signal isn’t in the rent comps. It’s in rate-case filings and large-load tariffs — and it leads the expense effect by a year or more.

The Pattern

Utility commissions decide rate structure in public, on the record. Almost nobody underwriting real estate reads those dockets. They should.

Step 1 — Data-center load breaks the shared rate base. For decades, large industrial users and households shared one rate base, and the spread between them quietly subsidized everyone’s bill. Hyperscale data centers are now individually large enough to break that math — a single campus can draw more power than a small city, and the old cost-sharing logic stops holding.

Step 2 — States split into two camps, and they did it the same week. On June 5, New York’s legislature passed a first-in-the-nation one-year moratorium on new data centers drawing 20 megawatts or more. The day before, June 4, Portland General Electric filed to do the opposite — keep the data centers, but raise their rates 29% while cutting residential rates 1.3% and commercial 2.2%. Ban it, or price it. That is the entire debate, running live, in two states, in one week.

Step 3 — Where neither happens, the cost lands on existing ratepayers. PJM, the grid operator for thirteen states, saw its capacity price clear at the $329-per-megawatt-day cap — more than ten times the level of two years ago, driven by data-center load growth. That flows into household and commercial bills in the delivery year that began this month.

Step 4 — Your operating expense diverges by jurisdiction, not by asset. Two identical buildings — one in a “price the load” state, one in a “did nothing” state — now sit on structurally different utility-cost curves. The building isn’t different. The statehouse is.

Step 5 — Value follows expense. Every incremental dollar of un-reimbursed utility cost capitalizes against your value at your exit cap rate. A persistent rate-class shift isn’t a line-item annoyance; it’s a permanent haircut or a permanent moat, depending on which side of the docket you landed on.

The connection nobody is making: Data-center load forces a rate-structure decision, which sets your utility-expense trajectory, which capitalizes directly into asset value — and all of it is visible in commission filings twelve to twenty-four months before it hits your operating statement.

The Data

New York moratorium threshold: 20 MW peak load triggers the one-year permit freeze (passed June 5, 2026); renewable-sourcing requirements phase in for facilities of 5 MW and above

Portland General Electric filing (June 4, 2026): data centers +29%, residential −1.3%, commercial −2.2%, applied to loads above 20 MW

Pre-filing imbalance: PGE residential customers paid more than two times the per-kilowatt-hour rate data centers paid (Oregon Citizens’ Utility Board)

PJM capacity price: $329.17 per megawatt-day, capped — more than ten times the $28.92 of two years prior; peak-load forecast up more than 5,400 MW year-over-year

Estimated household bill impact: 1.5% to 5% increases across thirteen states, beginning the June 2026 delivery year

Policy velocity: more than 300 data-center bills filed across 30-plus states in roughly six weeks

The hidden variable: Utilities aren’t a pass-through anymore — they’re a policy variable. The leading indicator for your future operating expense isn’t last year’s bill or a three-percent escalator. It’s whether your jurisdiction’s commission has ruled on large-load cost allocation yet. By the time it shows up in your actuals, the value has already moved.

Why This Matters

If you’re underwriting in or near a data-center growth corridor, here’s the exposure:

A flat utility-expense escalator may be fiction. In a “did nothing” jurisdiction inside PJM or a comparable grid, your utility line can jump well past your blended assumption. Underwrite the rate case, not the trend line.

Jurisdiction is now a diligence item, not a footnote. Two assets with identical trailing-twelve operating statements can have divergent forward expenses purely on rate-class policy. “Which state priced the load” belongs in your investment committee memo.

Reimbursement structure decides who absorbs it. Triple-net and RUBS structures push some of this risk to tenants — but only up to the point where occupancy starts to suffer. A gross lease eats the increase whole.

Exit risk is now a regulatory risk. A buyer two years out will price your asset off a utility curve a commission is setting right now. You can see their cap-rate input before they can.

The people who get hurt are the ones who treated the power bill as a pass-through instead of the swing variable it just became.

The Signal to Watch

This bottleneck is public if you know which dockets to read.

A name surfaced while you read this — the partner underwriting a deal in a data-center corridor, or anyone modeling a flat utility line in a PJM-territory metro. They’re pricing an expense the commission is about to rewrite, and they don’t know it yet.

Send them this before their next IC meeting. You’re not doing us a favor — you’re handing them cover, and you become the one who flagged it before it showed up in their actuals. That’s the person people keep close. One name, one forward, right now — while it’s in front of you.

Cross-domain reads like this one — spanning energy policy, grid economics, and real estate — are what I do for clients. Bring me your market or portfolio question at cokas.io: describe it, and get a written scope back within 48 hours. No sales call.

Submit a request at cokas.io → https://cokas.io