Tesla Gives Up $1,044 a Year for 27 kWh It Already Owns. A Megapack Holds the Same for 5.2 Years of It.

Abstract

Tesla will install two Powerwall units in a Texas home at no installation cost, lease them at $122 a month in the first year with a 3% annual escalator, and hand back an $87 monthly credit to households that enroll in its Tesla Electric Backup plan. The customer pays $35 plus tax. In return Tesla operates the battery — storing energy "during low-cost times" and selling it "back to the grid" during high-cost ones — while holding a minimum 20% backup reserve. Texas retail-choice areas only, two units, no solar.

The credit is the price of a virtual power plant, and it is payable in forgone revenue rather than cash: $87 × 12 = $1,044 a year, 71% of the year-one lease, on hardware Tesla owns either way. The money does not buy batteries. It buys the right to dispatch them.

There are three honest ways to price that right, and they do not agree. Per unit of energy, $1,044 against 27 kWh of nameplate is $38.67 per kWh-year, or $48.33 against the 21.6 kWh left after the reserve. Per unit of power, the same $1,044 is somewhere between $45 and $90 per kW-year, because Powerwall 3's output is grid-code configurable and Tesla publishes no setting for this programme. Per battery it is $522 a year, which is 3.5 times what Sunrun pays a CalReady household per battery — and roughly half what Sunrun's Storage Rewards tariff would pay for the same nameplate kilowatts. Which number you believe decides whether this is generous or cheap.

Two tests bound it. Against building the capacity, $200–285 per kWh of Megapack means the rental costs more than the asset after 5.2 to 7.4 years on nameplate, or 4.1 to 5.9 on dispatchable energy. Against buying the service from the market, Modo Energy's ERCOT benchmark puts the merchant battery fleet near $30 per kW-year for 2025 and a $29 pace in 2026 — so at every plausible grid-code setting Tesla is paying 1.5 to 3.0 times what a grid-scale battery earns in the same market for the same job.

That gap is the finding, and the escalator is the answer to it. The lease rises 3% a year; the credit prints flat at $87. Tesla has bought dispatch rights at a fixed nominal price against a rising revenue line, in a market where the value of those rights fell 52% year over year. By year ten the credit is 54.7% of the lease instead of 71%, and worth $800 in year-one dollars. The $87 is not a trade in electrons. It is customer acquisition that decays on schedule.

Table of contents

1. What Tesla published, and what it costs
2. Pricing a dispatch right three ways
3. The spread the battery has to earn
4. Test one: what the same storage costs to build
5. Test two: what the market pays for the same service
6. The collision: a fixed price against a falling curve
7. What the $87 is actually buying
8. What would prove this wrong
9. Implications

1. What Tesla published, and what it costs

Tesla's support page, read 14 September 2026, states the offer in one sentence: "The lease for two Powerwall units is $122/month in the first year with a 3% annual escalator. If you're enrolled in the Tesla Electric Backup plan, you may qualify for an $87 monthly credit. When applied to your lease payment, it would reduce the amount you owe each month in the first year to $35 + tax."

The mechanics are on the same page. Tesla, not the homeowner, runs the battery: "Tesla Electric will operate the battery system to store energy during low-cost times and then sell that energy back to the grid or use it to power your home during high-cost times." A minimum 20% backup reserve is held at all times, and the household keeps full access during outages. Eligibility is narrow — a Texas area with retail electric choice, exactly two Powerwall units, no solar panels, and continuous enrollment in good standing. The retailer is Tesla Energy Ventures, LLC, REP #10296.

The plan's rate card is published too, and it is the only place Tesla puts a number on the energy itself.

Every region pays the same 8.0 ¢/kWh for energy and receives 13.1 to 14.5 ¢/kWh for exports, with delivery charges passed through separately. The spread between those two columns — 5.1 to 6.5 ¢/kWh — is the only margin Tesla discloses anywhere in the programme.

2. Pricing a dispatch right three ways

Start from the disclosed numbers and take one operation at a time.

Rung 0. $87 a month, credited for as long as the household stays enrolled. Two Powerwall units. A 20% minimum reserve. Powerwall's nominal battery energy is 13.5 kWh, from Tesla's own datasheet.

Rung 1. $87 × 12 = $1,044 a year, per household, forgone rather than paid. Tesla owns the hardware in both worlds and collects the lease in both; the $1,044 is the difference between a household that lets Tesla dispatch and one that does not.

Rung 2. 2 × 13.5 = 27 kWh of nameplate energy. The reserve leaves 27 × 80% = 21.6 kWh that Tesla may actually move.

Rung 3. Three ways to divide it, and this is where the answer stops being single-valued:

The per-kWh figures come straight off disclosed numbers and are as hard as anything here. The per-kW figure is not: Powerwall 3's nominal AC output is grid-code configurable at 5.8, 7.6, 10 or 11.5 kW, and Tesla publishes no setting for this programme, so two units are anywhere from 11.6 to 23 kW and the price of a kilowatt-year moves by a factor of two. Every per-kW number below carries that range rather than a point estimate.

3. The spread the battery has to earn

A capacity price means little without a duty cycle. Tesla discloses none — not cycles per day, not hours per dispatch, not how often it calls the fleet. So invert the question: what spread must Tesla capture per kilowatt-hour moved to cover the $1,044?

At one full cycle a day — 21.6 kWh × 365 = 7,884 kWh — Tesla needs 13.2 ¢/kWh, which is 1.7 times the entire 8.0 ¢ retail energy rate it charges the same household, and more than double the 5.1–6.5 ¢ sellback spread it publishes. At two cycles a day the requirement falls to 6.6 ¢, which lands inside that published spread. Below one cycle a day the case stops closing on arbitrage alone.

This is the cost model's sharpest constraint, and it cuts one way: the credit only pays for itself on electrons if Tesla is cycling hard or capturing scarcity. ERCOT obliges on the second — summer scarcity intervals clear orders of magnitude above the daily spread, and a handful of hours can carry a year. But a programme underwritten by scarcity hours is underwritten by weather, which is the least contractible input in the model.

4. Test one: what the same storage costs to build

Tesla sells the alternative to itself. A Megapack runs roughly $200–285 per kWh before project-specific site work and financing, the range we used in our Megapack and Bloom comparison. Set the recurring rental against the one-time build:

• On nameplate energy, $38.67 per kWh-year against $200–285 per kWh is 5.2 to 7.4 years before the rental has cost more than the asset.
• On dispatchable energy, $48.33 against the same range is 4.1 to 5.9 years.

The fleet version is the same sum at scale. Our 500 MW dispatch piece put roughly 69,000 Powerwall owners behind California's September event. Were every one of them on these terms, the credit would run $72.0 million a year and secure 1,863 MWh of nameplate capacity. Buying that capacity as Megapacks costs $373–531 million once. Crossover: 5.2 to 7.4 years, the same answer the per-household sum gives, which is the check that it holds.

Three things sit on the rental side of that comparison and none are priced here. The household already has an interconnection, a meter and a permit, and Megapack's $200–285 excludes exactly the site work those represent. The batteries are distributed behind load, where a grid battery is not. And at the end of it Tesla still owns every Powerwall. A five-year crossover on capex alone is a floor for the rental's fair price, not a verdict against it.

5. Test two: what the market pays for the same service

The competitor left out is the one carrying the finding, so here is the field — and the bases do not line up, which is the point.

Read per battery, Tesla is generous: $522 a year against Sunrun's $150 per battery for CalReady, a seasonal standby obligation covering 4–9 p.m. from May to October. Read per kilowatt, Tesla is thrifty: Sunrun's Storage Rewards tariff pays $110 per available kW-year, which on two Powerwall 3 units at the 7.6 kW setting would be $1,672 against Tesla's $1,044.

Both readings are correct and they measure different products. CalReady is a seasonal, solar-paired, few-hours-a-day standby; Tesla's is year-round, no-solar, and hands over operational control of the whole system. A standby right for 150 afternoons is not the same asset as unrestricted dispatch for 365 days, and pricing them on the same line is the error the per-battery number invites.

The cleaner benchmark is the merchant fleet doing the identical job in the identical market. Modo Energy's ERCOT benchmark puts battery revenues at about $30 per kW-year for 2025, a trailing-twelve-month $2.43 per kW-month as of March 2026, and a $29 per kW-year pace through the first five months of 2026. Against that:

• At the 5.8 kW setting, Tesla pays $90 per kW-year — 3.0× the 2025 benchmark.
• At 7.6 kW, $68.68 — 2.3×.
• At 10 kW, $52.20 — 1.7×.
• At 11.5 kW, $45.39 — 1.5×.

There is no configuration at which Tesla is paying less than a grid battery earns. At the most favourable reading it pays half again as much; at the least, three times.

6. The collision: a fixed price against a falling curve

Two curves, moving in opposite directions.

The first is the value of the thing Tesla is buying. ERCOT battery revenues ran $54.1 per kW year-to-date through November 2024 and $26.0 through November 2025 — down 52% — as the fleet passed 14 GW entering 2026 and 16 GW by the second quarter, and daily spreads compressed. Day-ahead spreads fell about 50% year over year by June 2026. More batteries chasing the same volatility is a curve with one direction.

The second is what Tesla pays for it, which does not move at all. The credit prints $87, flat, while the lease escalates 3% a year.

Held flat in nominal terms, the credit falls from 71.3% of the lease in year one to 54.7% by year ten and 40.7% by year twenty, and its purchasing power decays from $1,044 to $800 and then $595 in year-one dollars. Tesla has written a fixed-nominal, indefinite-duration call on household storage and indexed the offsetting revenue to inflation.

That is the structure that makes a price 1.5–3.0× the merchant rate survivable. Tesla is not betting that dispatch rights are worth $1,044 a year forever. It is betting that $1,044 stops being a large number — against a lease at $159 a month in year ten and $214 in year twenty — faster than the spreads compress.

7. What the $87 is actually buying

The merchant reading does not close, so the credit has to be buying something other than electrons, and the page says what.

It buys a retail electricity customer at a fixed 8.0 ¢/kWh, in a deregulated market where acquisition is otherwise paid to brokers. It buys a hardware lease that escalates for as long as the customer stays. It buys installation without a subsidy line — Tesla books no upfront installation cost and recovers it through a twenty-year annuity. And it buys dispatch rights as the byproduct of all three, which is why Tesla can pay above the merchant rate for them without the trade being irrational.

Read that way the $1,044 is not a capacity price at all; it is a discount on a bundle, and the dispatch rights are what Tesla takes instead of a higher margin. The test of that reading is what happens to the credit when the electricity contract renews — because the page is careful to say the plan renews, not that the credit holds: "Tesla will offer you a competitive rate for renewal… You will need to renew your Tesla Electric plan to continue receiving the bill credit." Availability is promised for the life of the lease. The amount is promised for the first year only.

8. What would prove this wrong

One disclosure settles it. If Tesla raises the credit above $87 at the first wave of plan renewals, the merchant reading is right and the company is paying up for capacity because the capacity is worth it. If the credit holds at $87 while the lease escalates past $130 — which is what the published terms imply — it is an acquisition cost decaying on schedule, and the reading here holds.

Three narrower ways to be wrong. The grid-code setting could be the 11.5 kW maximum, which drops the per-kW price to $45.39 and makes the gap to ERCOT's $30 a premium rather than a multiple. The duty cycle could exceed two full cycles a day, which brings the required spread inside the published 5.1–6.5 ¢ and makes the arbitrage close on its own. And Tesla could be monetising ancillary services rather than energy, which the page's language — "sell that energy back to the grid" — describes but does not exclude.

9. Implications

1. The consumer headline is a year-one price. $35 plus tax is true for twelve months. The published escalator puts the same household at $72.18 in year ten and $126.93 in year twenty, against a credit that never moves. Nothing on the page is hidden; nothing on the page adds it up either.
2. A virtual power plant priced per battery tells you almost nothing. The same $1,044 is generous or thrifty depending on whether the denominator is the unit, the kilowatt or the kilowatt-hour, and the three disagree by more than a factor of three. Any VPP comparison that does not name its denominator is decoration.
3. Tesla is paying above the merchant rate for distributed capacity, and can afford to. 1.5–3.0× ERCOT's benchmark is defensible only because the capacity is a byproduct of a lease and a retail contract. A pure-play VPP operator paying these rates would not clear.
4. The build-versus-rent crossover is about five years and is not a verdict. It ignores interconnection, siting and the residual value of hardware Tesla still owns. It does say that anyone modelling this as cheap capacity should stop at year five.
5. The programme's real constraint is renewal, not economics. The credit's amount is committed for one year against a lease committed for twenty. Every figure above turns on what Tesla does at the first renewal.

Primary sources worth reading

• Powerwall Lease With Tesla Electric — the lease price, escalator, credit, eligibility, the operating language and the six-region rate card.
• Powerwall 3 datasheet — 13.5 kWh nominal energy and the grid-code-configurable AC output.
• Modo Energy's ERCOT BESS benchmark — monthly and trailing-twelve-month revenue per kW for the merchant fleet.
• Sunrun CalReady — the per-battery seasonal structure this is measured against.

Every lease term, credit, eligibility rule, operating description and published rate is from Tesla's Powerwall lease support page, read 14 September 2026; the page names no cycle count and no kilowatt-hour. The 13.5 kWh rating and the configurable AC output are from Tesla's Powerwall 3 datasheet. ERCOT fleet revenues are Modo Energy's published benchmark. Sunrun's CalReady and Storage Rewards figures are the company's own, press-reported and not verified here. Megapack pricing is the $200–285/kWh range used in our earlier comparison, itself indicative rather than quoted. The 69,000-owner California figure is Tesla's, from the September dispatch event. Every annualisation, break-even, required-spread and real-terms figure is ours, and the per-kilowatt figures depend on a grid-code setting Tesla does not publish.