Every solar battery pitch leans on the same line: charge cheap, discharge expensive, pocket the difference. What it usually skips is the number that determines whether that math actually works — the size of the gap between your utility's peak and off-peak rate, in real cents per kWh, not a percentage. A 20-cent spread on 400 kWh of shifted monthly usage is worth roughly $80 a month. A 7-cent spread on the same usage is worth $28. Same behavior, same battery, wildly different payback — and almost nobody publishes the utility-by-utility numbers side by side.
Here's what the actual 2026 rate sheets say, and which utilities clear the bar.
Run your own numbers first. The Solar Math Worksheet is one fillable page: your rate structure, TOU spread, battery sizing goal, incentive stack, and payback estimate — the math this site keeps telling you to do, in order. Download the free PDF. No email required.
What "worth it" actually requires
A battery earning its keep on rate arbitrage alone needs three things to line up, not one:
A peak/off-peak spread wide enough to matter — generally north of 20¢/kWh, since installed battery costs now run $850–$1,700 per usable kWh and every cycle needs to earn its share of that back.
A peak window that actually overlaps your usage. SCE's 5–8 PM plan is useless to you if you're not home until 6:30. A wide spread on hours you were never drawing from the grid anyway saves nothing.
A tariff that holds still long enough to pay the battery off. Utilities revise TOU schedules almost every year — several of the rates below changed in just the last six months — and a 10-year Tesla Powerwall warranty assumes the rate structure you signed up for still exists in year 8.
Miss any of the three and the "arbitrage" is theoretical.
The rate spread map (2026 published rates)
| Utility | State | On-peak | Off-peak | Spread | Clears 20¢? | |---|---|---|---|---|---| | SCE (TOU-D-5-8PM) | CA | $0.74/kWh (summer) | $0.34/kWh | ~$0.40 | Yes | | Con Edison (Rate III) | NY | $0.38/kWh | $0.05/kWh | ~$0.33 | Yes | | SDG&E (TOU-DR1) | CA | $0.578/kWh | $0.270/kWh (super off-peak) | ~$0.31 | Yes | | PG&E (E-TOU-C) | CA | ~$0.59/kWh | ~$0.34/kWh | ~$0.25 | Yes | | SRP (TOU-E) | AZ | ~$0.34/kWh (summer) | ~$0.12/kWh | ~$0.22 | Yes | | PSE&G (new TOU, June 2026) | NJ | ~$0.31/kWh | ~$0.09/kWh | ~$0.22 | Yes | | Xcel Energy | CO | $0.213/kWh (summer) | $0.079/kWh | ~$0.13 | No | | National Grid (SC-1 VTOU) | MA/NY | $0.142/kWh (summer) | $0.059/kWh | ~$0.08 | No | | Potomac Edison | MD | ~$0.17/kWh | ~$0.10/kWh | ~$0.07 | No |
Rates pulled from each utility's own 2026 published tariff or rate-comparison page where cited; several are seasonal and the summer figure is shown since that's when spreads are widest. Confirm your specific rate code before assuming these apply — SCE and PG&E alone each run four or five different residential TOU schedules with different windows and different spreads.
The pattern: California, Arizona, and the New York/New Jersey corridor are where TOU arbitrage has real teeth right now. Colorado, Maryland, and much of New England are running spreads under 15¢ — a battery there is a backup-power purchase, not an arbitrage play, no matter what the sales pitch implies.
What the spread number doesn't tell you
A 40-cent SCE spread sounds like an easy yes. It isn't automatically one, for three reasons the rate sheet doesn't mention.
The federal battery tax credit is gone. The 30% residential credit under Section 25D expired December 31, 2025. Every payback calculation you find online that still nets out "30% off" is using a subsidy that no longer applies to anyone buying today. That alone moves a typical Powerwall 3 install from roughly $9,400 after-credit back up to its full $12,000–$16,500 range.
TOU-only payback is long — 15 to 21 years, per industry cost data published this year. A Tesla Powerwall 2 at $11,500 installed, earning only the $540/year a moderate arbitrage spread typically generates, takes about 21 years to break even — well past the 10-year warranty. High-spread utilities like SCE, SDG&E, and Con Edison shrink that considerably (arbitrage alone can hit $720–$960/year at 20¢+ spreads), but "shrink" still generally means 6–9 years, not the 3–4 years implied by marketing copy that quietly folds in a demand-response program most homeowners aren't enrolled in.
Demand-response stacking is what actually gets paybacks under 5 years — and it's regional. Massachusetts and Rhode Island's ConnectedSolutions program pays $225–$275 per kW discharged during summer events, adding $1,125–$1,375/year on top of arbitrage. California, Arizona, and New York don't have an equivalent program at that payout level. So a Massachusetts battery with a mediocre 8¢ TOU spread can out-earn a California battery relying on a 40¢ spread with no demand-response stacking — the opposite of what the spread table alone suggests.
Battery degradation eats the spread over time. Lithium batteries lose roughly 2–3% of capacity a year. A 13.5 kWh Powerwall is down to about 10.8–11.5 kWh of usable capacity by year 10 — meaning the arbitrage income that looked attractive in year one is smaller every year after, right as the battery approaches the edge of its warranty.
Checking your own utility
The published number to look for is the total energy charge difference between your plan's most expensive and least expensive tier — not the "average" rate, and not a percentage-savings claim. Pull your utility's actual current tariff sheet (search "[your utility] time of use rate schedule PDF") rather than relying on a third-party summary, since TOU windows and cent amounts get revised often enough that even this list may be stale within a year. If your spread lands under 15¢, a battery is a resilience purchase — backup power during outages — not a bill-reduction one, and it should be evaluated on that basis instead.
Sources: SCE, SDG&E, PG&E-affiliated CleanPowerSF, Con Edison, and Xcel Energy 2026 published rate schedules; NuWatt Energy 2026 battery cost and payback data; industry installed-cost comparisons (Tesla, FranklinWH, Enphase) current as of 2026.