Energy is the one opex line you can't renegotiate after close

Positioning · why the energy number is deal-critical

Everyone argues about whether energy is 10% of cold-storage operating cost or 60%. Both are true in different buildings — and both miss the point. What a lender actually underwrites is the spread: two identical cold buildings, two identical tenants, and the market you bought in decides whether the power bill eats the margin.

The objection, met head-on

If you have been told power is a rounding error, here is where that comes from — and where it breaks.

GCCA / IARW survey	Electric power = 10% of operating cost (labour 46%, property 35%). This is the operator-facing average across a mixed estate. findinfood / GCCA
UnionMetric (MAI-authored)	Energy = "60% to 70% of operating expenses" for a dedicated cold facility. UnionMetric
apers.app	Energy = 10–18% of opex vs 2–3% for dry industrial — and it is the operating-margin overlay: PJM +48.9%, ERCOT +21% (projected +45% in 2026), compressing operator margins 200–400 bps structurally. apers (May 2026)

Reconciliation, not a party line. The 10% figure is a landlord-wide average; the 60–70% figure is a single-purpose cold building carrying its own refrigeration. ColdStore publishes both, sourced, because a benchmark that hides the disagreement is not one an appraiser can cite.

What our own model shows

Across the 39 markets we model, the same reference cold facility ranges from $2.12 per pallet·yr in Seattle, WA to $9.49 in Inland Empire (Chino), CA — a 4.5× spread. The building model is identical everywhere. The only inputs that move are local climate and the state tariff (here WA at 6.61¢/kWh vs CA at 21.53¢/kWh).

That is the whole argument: "energy is small" and "energy is the swing factor" are the same fact stated at different altitudes. A 4–5× swing between markets is not a rounding error — it is a cap-rate question.

The structural squeeze

Energy cost is not a static line item. Power repriced hard in the two largest grids — PJM +48.9% and ERCOT +21%, with a projected +45% in 2026 — while the cap-rate premium cold storage used to earn over dry industrial has compressed to 0–50 bps (cold 5.0–5.5% vs dry 4.5–5.0%). The margin that used to buffer a power shock is gone, so the shock now lands on the valuation. apers (May 2026), JLL via Retail Asia

A financing brokerage says the quiet part out loud: "cold storage also has underwriting wrinkles that do not apply to dry industrial: specialized refrigeration equipment, higher build costs, tenant credit concentration…" Janover. Refrigeration is the wrinkle — and its energy draw is the part that varies most by market.

What to do with this

Underwriting a market? Start with the free 39-market table to see the band before you commit.
Underwriting a building? The paid report puts your actual building and tariff against the band, with sensitivity ranges.
Defending a number? Every figure cites to a dated public source — see the cite & data page.

Figures quoted above are from the linked third-party sources (GCCA/findinfood, UnionMetric, apers, Janover, JLL) and are reproduced with attribution. The market range is our own model over the 39 markets in the dataset. Not investment advice. Questions: hello@thermoneural.com.