I modeled the economics of owning a Supercharger site (Supercharger for Business) across 86 utility tariffs — what actually decides whether a site pencils
Since Tesla opened Supercharger for Business, a property owner can buy and own a Supercharger site — Tesla's configurator quotes the hardware and sketches revenue. I spent months modeling the side the configurator doesn't show, using the filed utility tariff for each territory (86 schedules across 49 states + DC so far). Some things I learned that I haven't seen written up anywhere:
The utility territory matters more than the state. The same 8-stall site with the same traffic can differ by six figures a year in electricity cost depending on which side of a utility boundary the parcel sits. The spread in filed demand charges alone runs from about $3/kW-month (Kansas) to $34/kW-month (Casper, WY) — roughly $1,500/mo vs $17,000/mo for the same 500 kW site before a single kWh is sold.
Demand charges come in four different structures, not one. Subscription blocks (PG&E's BEV-2 style), plain per-kW, load-factor tiers (the Massachusetts EV rates), and capped-kW designs (Florida's GSLD-1EV). Two territories with similar headline rates can bill completely differently at Supercharger load shapes — DCFC sites have brutal peak-to-average ratios, which is exactly what demand charges punish.
The meter bills more kWh than cars receive. Conversion and cabling losses mean a site that dispenses 1M kWh/yr purchases roughly 1.14M at a typical 12% loss factor. Revenue rides the small number, electricity cost rides the big one — about $29k/yr of invisible cost at California-ish rates.
Some states tax every kWh sold on top of the power bill. Pennsylvania 1.72 cents/kWh, Georgia 2.8 cents, a 3-cent cluster (OK/MT/WI/NE). Utah taxes gross charging revenue — works out to about 5 cents/kWh at typical retail. None of this appears on a utility bill.
Rough scale, at defaults: an 8-stall V4 build lands around $1.26M all-in (hardware + install + contingency + sales tax, before any utility upgrade), and a decently-ramped site grosses roughly $300k/yr at $0.45/kWh — against which the utility bill, Tesla's $0.10/kWh network fee, rent, and the tax stack all bite. Whether what's left clears cost of capital is genuinely site-by-site — I've modeled territories where it pencils comfortably and neighboring ones where it never does.
Disclosure: I built the tool I used for this — ForgeAsset (https://forgeasset.com). The interactive map (every US Supercharger over territory-level effective cost) and the demand-charge lookup are free, no signup. The full site model is paid; that's how I fund the tariff research. Happy to answer questions about any specific territory or utility in the comments, and if you spot a filed rate I've got wrong, I genuinely want to know — the whole point is matching the filed sheets.