PG&E's BEV-2 rate, explained for charging-site hosts

PG&E territory holds more Superchargers than most states hold in total, and every one of them buys electricity under a rate design that barely resembles a normal commercial tariff. BEV-2 replaced the demand charge — the line that decides most DC fast-charging economics — with a subscription. That swap changes what a site pays, what it risks, and which inputs deserve attention before anyone commits capital.

This post describes the structure ForgeAsset models for PG&E sites: BEV-2-S, the secondary-voltage business EV rate. Nothing here is advice; it is a walk through what the tariff does and how the model prices it.

Who the rate is for#

PG&E's BEV rates exist specifically for separately-metered EV charging. BEV-1 covers small installations; BEV-2 covers sites above 100 kW — which means effectively every multi-stall DC fast-charging site. The "S" variant takes service at secondary voltage, the common case for a retail-adjacent site without its own primary-voltage infrastructure.

The subscription replaces the demand charge#

A conventional commercial tariff bills demand as a measurement: the single highest 15-minute draw of the month, at a filed dollars-per-kW rate. The host finds out what that cost after the fact.

BEV-2 inverts the direction. The host chooses a subscription level in advance, in 50 kW blocks — the modeled block price is $95.56 per block per month. A site that subscribes six blocks has bought 300 kW of capacity for the month at a known price, the way a data plan buys gigabytes.

Two consequences follow from the model's point of view:

• The cost is a chosen input, not a measured outcome. The model carries a subscribed-kW-per-stall assumption (150 kW per Tesla V4 stall by default) and prices blocks from it. There is no bad-month surprise built into the structure — a single 15-minute spike does not reprice the month.
• The choice itself becomes the risk. Subscribe too little and overage charges apply when draw exceeds the subscribed level; subscribe too much and the site pays every month for headroom it never uses. The tariff converts billing volatility into a sizing decision.

Three prices for energy, by clock and calendar#

Every kWh also pays an energy rate that depends on when it flows. BEV-2's day has three windows, and the modeled annualized rates are:

The shape is deliberate. The super-off-peak window sits in the middle of the day — the hours when California's grid is flooded with solar — so the cheapest electricity on this tariff is daytime electricity. That is the opposite of the overnight-cheap pattern most commercial TOU rates carry, and it happens to suit public fast charging, where daytime is when drivers show up.

The spread matters as much as the levels: peak energy costs roughly 2.8× super-off-peak. A site whose charging skews into the 4–9 p.m. window has a materially different cost stack than one whose traffic peaks at lunch. The model prices this through its time-of-use mix — 30% peak, 45% off-peak, 25% super-off-peak by default, and editable per scenario.

What the model does with all of it#

For a PG&E scenario the engine builds the monthly bill from four pieces: dispensed energy grossed up by the loss factor (the meter bills more kWh than cars receive), priced through the three TOU rates at the scenario's mix, plus the subscription blocks from the subscribed-kW assumption. Rate escalation then compounds the energy side year over year across the 15-year statement.

The same structure is why the cost-stack pillar treats electricity as two lines, not one: energy that scales with utilization, and capacity that does not. On BEV-2 the capacity line is fixed by choice; on a classic demand-charge tariff it is set by the worst 15 minutes of the month. The demand-charge lookup shows that contrast across every tariff in the library, and the utility ranking shows where PG&E's all-in cost lands against the other 85 modeled schedules — spoiler: the energy rates, not the subscription, are what place it near the expensive end.

The California stack around the tariff#

The tariff is one layer of the state's economics. The LCFS credit adds a modeled revenue line on every dispensed kWh, state incentives fund build-out, and the tax stack takes its own share — the California state guide walks the full picture, and the PG&E BEV-2-S page carries the modeled rates, sources, and the reference-site arithmetic in one place.

Rates cited are the model's current PG&E calibration, digit-verified against filed sheets; derivations live on the methodology page. To see the whole stack priced for a specific address, the wizard runs the full scenario.