Balcony Solar With a Battery: Does the Storage Actually Pay Back?

If you register a plug-in solar device under Germany's simplified procedure, you receive no payment for electricity fed into the grid. That is not fine print, it is the wording of the statute. § 8 Abs. 5a EEG 2023 refers to systems that are "der unentgeltlichen Abnahme zugeordnet" — assigned to uncompensated offtake (statute text at gesetze-im-internet.de). Uncompensated means what it says. The surplus is given away.

For the storage question this inverts the logic. On a rooftop array with a feed-in tariff, the battery competes against a real revenue stream — every stored kilowatt-hour has to beat the tariff it displaces. On a balcony system, the battery competes against zero. Every kilowatt-hour it rescues is worth the full retail price.

That is the strongest argument for buying a battery, and it appears on almost no product page. The reason it still often fails to be enough comes further down: the battery has running costs of its own.

The power limits of the simplified procedure sit in the same paragraph: up to 2 kilowatts of installed module power and up to 800 volt-amperes of inverter power in total. Those figures are unchanged. If you read a report claiming the 800-watt limit was lifted in 2026, check the source's date against the statute.

Here is the whole thing. Every row carries either a source with a date or the word assumption. There is no third category, and no figure appears without its unit.

The arithmetic is three steps, so you can repeat it with your own numbers:

1. Additional self-consumed energy = annual yield × (share with storage − share without storage)
2. Actually usable after losses = that value × round-trip efficiency
3. Annual benefit = usable energy × electricity price − standby draw × electricity price

Step three is the one manufacturer charts leave out.

What 5 watts of standby costs per year

A battery that stays ready around the clock draws power around the clock. 5 W × 8,760 h = 43.8 kWh per year — that is arithmetic derived from an assumed wattage, not a measurement. At 37.0 ct/kWh it comes to €16.21 per year that the battery costs before it has rescued a single kilowatt-hour.

What is sourced is the range, and the range is alarming. HTW Berlin's Stromspeicher-Inspektion 2026, published 23.03.2026 together with its spin-off aquu, compared twelve systems from ten manufacturers and measured standby consumption from 4 W to 64 W. Those are home storage systems in the 5 kW and 10 kW classes, not balcony units — the figures do not transfer directly, but they show how far apart a frugal and a wasteful device sit. HTW puts the annual cost advantage of its test winner over the least efficient system in the same field at roughly €200.

Run the sensitivity yourself: at 10 W instead of 5 W, annual self-consumption rises to 87.6 kWh and costs €32.41 per year. That nearly halves the battery's benefit. If the standby figure is not in the datasheet, it is the most important question to put to the manufacturer — more important than any capacity claim.

The same arithmetic, applied three times. Every non-price input is an assumption and flagged as one above; the electricity price and the feed-in payment are sourced.

Two things jump out.

First: the spread from 5.3 to 25.8 years is not produced by device quality. It is produced by household size and purchase price. If you draw a lot of power in the evening, the battery earns its keep. If you are home during the day and already consume the output directly, what you mostly buy with a battery is its standby draw.

Second: doubling the purchase price from €500 to €1,000 doubles the payback period, and no efficiency difference compensates for that. A device that converts 15 percentage points better saves roughly €13.60 a year in case B. A device that costs €500 less saves €500 immediately. In this category price beats technology, and not narrowly.

If you want to substitute your own annual yield for the assumption, HTW Berlin's Solar Storage Systems research group publishes the Stecker-Solar-Simulator, which computes yield, self-consumption share and payback for your location, orientation and load profile — small battery systems included.

On 19.03.2026, Stiftung Warentest published the first large test of balcony solar batteries, in issue 4/26. Five devices were tested: Anker Solix Solarbank 3 E2700 Pro, EcoFlow Stream Ultra, FoxESS Avocado Orbit M, Growatt Nexa 2000 and Zendure SolarFlow 800 Pro. Warentest puts the test field's price range at "zwischen rund 700 und 1 000 Euro".

The finding, verbatim: "Drei der fünf getesteten Akkus bewerten wir mit der Gesamt­note Mangelhaft, weil sie zu viele elektromagnetische Stör­signale aussenden" — three of five batteries were graded deficient for emitting too much electromagnetic interference. Overall grades on test.de range from Befriedigend to Mangelhaft, satisfactory down to deficient. Not one device reached "gut".

In practice, electromagnetic interference means flickering televisions, humming speakers, and Wi-Fi routers that keep dropping their connection. This is not a marginal finding — it is the reason for three of the five grades.

Which device carries which grade sits behind the paywall. Trade outlets that read the full test name the Anker Solix Solarbank 3 Pro as the only device rated satisfactory, with appgefahren.de quoting a grade of 3.3 on 19.03.2026. That is a secondary source, not a primary statement from Warentest, and we label it as such.

Also behind the paywall: the measured charge and discharge efficiency. Which is exactly why the model above carries an assumption there rather than a measurement.

The hardware has converged. All five tested devices are LiFePO4 batteries with an integrated inverter, 800 W output and comparable capacity. What separates them is software: when the battery charges, when it discharges, and how precisely it hits zero export.

That is the lever the payback table abstracts into "share with storage". A rule engine that merely estimates household load discharges too early or too late. One that measures actual draw at the meter point hits it. The difference between a 70 % and an 80 % self-consumption share is about €37 a year in case C — more than a meter costs.

Which is why several systems rely on an external energy meter rather than an estimate. The Shelly Pro 3EM 3CT63, dreiphasiger WLAN- und LAN-Energiezähler, 63 A measures three-phase at the meter point and is listed as a compatible measurement source by several storage manufacturers. Installation in a German consumer unit belongs in the hands of a qualified electrician. Before buying, check your battery's compatibility list for which meters it supports — the lists differ between manufacturers and sometimes between firmware versions.

The four devices below are from the test field of 19.03.2026 or are immediate sibling models. All technical figures are manufacturer specifications from the respective datasheets — not our own measurements and not a verdict from us. Prices move daily, which is why the payback table above uses two round scenarios instead of one day's price.

Anker Solix Solarbank 3 E2700 Pro

The most expensive device in the field, and per the secondary sources above the only one rated satisfactory. 2.68 kWh nominal, four MPP trackers, 1,200 W bidirectional charging, expandable to 16 kWh per the manufacturer. The Anker SOLIX Solarbank 3 E2700 Pro, Balkonkraftwerk-Speicher 2,68 kWh, 4 MPPT is therefore the option for case C households — high evening draw, larger array. In the €1,000 scenario it still needs 10.7 years even there.

EcoFlow Stream Ultra

1.92 kWh, four MPPTs, 2,800 W solar input, two AC connections totalling 2,300 W, expandable to 11.52 kWh per the manufacturer. Present in the Warentest field. The EcoFlow Stream Ultra Balkonkraftwerk-Speicher 1,92 kWh, 4 MPPT, 2800 W Solareingang sits in the middle of the price band and is the obvious pick if the array is meant to grow later.

Growatt Nexa 2000

An all-in-one unit with integrated inverter, 2,048 Wh, four MPP trackers, Wi-Fi. Also part of the Warentest field of 19.03.2026. The Growatt NEXA 2000 Wechselrichter-Speicher, All-in-One Balkonkraftwerk-Lösung mit WLAN is the most compact of the four and needs no separate inverter.

Zendure SolarFlow 800 Plus

⚠️ Not the tested device. Warentest examined the SolarFlow 800 Pro; this is the cheaper 800 Plus at 1,920 Wh. The grade from the test explicitly does not apply to this model. The Zendure SolarFlow 800 Plus Balkonkraftwerk-Speicher 1920 Wh, modular erweiterbar represents the lower price scenario: if you do not need the Anker's 2.68 kWh and its expandability, case B payback becomes 8.2 years instead of 16.4.

Three points that matter for the purchase decision and that are routinely reported wrongly:

The limits are unchanged. § 8 Abs. 5a EEG 2023 states 2 kW of module power and 800 VA of inverter power, applied cumulatively per grid connection point. If you read that the limit was abolished in 2026, check the report against the statute.

Schuko is permitted, but the standard does not cover batteries. With DIN VDE V 0126-95, announced in November 2025 and published with a December 2025 edition date, a standard Schuko plug is normatively provided for plug-in solar devices up to 960 Wp of module power and 800 VA of inverter power, provided the device carries the required protective measures — as pv magazine described on 14.11.2025. Two caveats: the "V" stands for Vornorm, a preliminary standard, and batteries are not within its scope. Combinations with storage carry their own requirements. A VDE standard is also not a law.

The battery is registered separately. In the Bundesnetzagentur's Marktstammdatenregister the battery is its own unit with its own entry, and is not captured automatically alongside the solar array.

Four things are deliberately absent from this article:

• No measured efficiency for the five Warentest devices. The figure sits behind the paywall. The 85 % in the model is an assumption, not a measurement.

• No standby figure for balcony batteries. HTW's 4 to 64 W refers to home storage in the 5 kW and 10 kW classes. The 5 W in the model is an assumption.

• No primary-source mapping of individual Warentest grades to individual models. The 3.3 for the Anker comes from appgefahren.de, not from test.de.

• A widely quoted percentage for actually retrievable capacity that we could not attach to a specific model. Publishing a figure like that without naming the model would be an allegation against a manufacturer.

For how nominal capacity in milliamp-hours becomes actually usable watt-hours, and why the gap is not a defect, see the conversion step by step.

Every calculation path is shown so you can repeat it with your own numbers. Test results and legal positions are as dated above.

The honest answer to "does the battery pay back" is: somewhere between 5.3 and 25.8 years, and the spread is created in your household, not in the device. Three numbers decide it — your evening consumption, the purchase price, and the standby draw. Round-trip efficiency, the number product pages most like to talk about, is the least important of the four.