Savings
Can a battery without solar panels save money on a time-of-use tariff in South Africa?
Sometimes, but less than most people expect. On Eskom Homeflex in 2026/27 a battery charged at off-peak (about 253c per kWh all-in) and used in the winter evening peak (about 909c) earns roughly R55 a weekday per 10 kWh, but in the nine summer months the gap shrinks to about 204c and the same battery earns about R15 a weekday. That works out to very roughly R6 000 to R7 000 a year, so a battery-only system usually takes longer to pay back than its warranty lasts. Add panels and the maths changes completely.
Updated 27 October 2026 · Illustrative ranges — not a formal quote, credit decision or legal advice.
This question comes up a lot now that load shedding has eased: people who bought an inverter and battery for backup ask whether the battery can earn its keep by itself, and others ask whether to skip panels altogether. The short answer is that tariff arbitrage alone is a weak business case on residential tariffs, and it only works at all on a time-of-use tariff.
Step one: you need a time-of-use tariff
A battery saves money only if the electricity you store is cheaper than the electricity you later avoid. On a flat or block tariff, such as Ekurhuleni’s flat Tariff B, City Power prepaid, or Cape Town Home User, a kWh at 02:00 costs the same as a kWh at 19:00. Charging a battery from the grid then just buys the same kWh, loses about one in ten of them to charging and discharging losses, and saves nothing. On those tariffs a battery saves money only when it stores solar energy.
Time-of-use tariffs that make the question worth asking include Eskom’s Homeflex (mandatory for Eskom-direct homes that want to export solar) and Cape Town’s opt-in Residential Time-of-Use tariff for homes with smart (AMI) meters. See our Cape Town TOU answer for that one.
The Homeflex numbers for 2026/27
These are Eskom’s 2026/27 booklet rates incl. VAT, with about 62c per kWh of legacy, network demand and ancillary charges added, because those apply to every kWh you import.
| Period | All-in cost per kWh | Gap versus off-peak |
|---|---|---|
| Winter peak (June–August) | ≈ 909c | ≈ 656c |
| Winter standard | ≈ 321c | ≈ 68c |
| Summer peak (September–May) | ≈ 457c | ≈ 204c |
| Summer standard | ≈ 308c | ≈ 55c |
| Off-peak, both seasons | ≈ 253c | — |
Only the peak gap is big enough to matter. Standard-hour gaps of 55c to 68c are wiped out by charging losses, so a battery-only system should charge in off-peak and discharge only in peak.
Worked example: 10 kWh battery, no panels (illustrative)
Assumptions, stated plainly so you can change them: the battery charges 10 kWh from the grid in off-peak each weekday, delivers about 9 kWh to the house during peak (about 90% round-trip efficiency), the house actually uses at least 9 kWh in peak hours, and there are about 22 weekdays a month. Weekends have little or no peak on Homeflex, so we count no saving there.
| Winter weekday | Summer weekday | |
|---|---|---|
| Cost to charge 10 kWh at ≈ 253c | ≈ R25.30 | ≈ R25.30 |
| Peak kWh avoided: 9 kWh | × 909c ≈ R81.80 | × 457c ≈ R41.10 |
| Net saving per weekday | ≈ R56 | ≈ R16 |
| Months × weekdays | 3 × 22 | 9 × 22 |
| Season total | ≈ R3 700 | ≈ R3 150 |
That is roughly R6 850 a year, and only if the home really draws 9 kWh in the peak windows every weekday. Eskom’s booklet says the evening peak is now three hours and the morning peak two hours; a typical home uses perhaps 1 to 3 kW in those hours, so many households cannot absorb a full 9 kWh of peak discharge. If yours uses 5 kWh in peak, the saving falls to roughly R3 800 a year.
Compare that with the cost
Our battery guide lists 2026 supply-only retail bands of about R25 000 to R46 000 for a mid wall pack of about 10 kWh, before the hybrid inverter, installation and Certificate of Compliance. A hybrid inverter swap alone is quoted at about R18 000 to R55 000 installed in our inverter replacement answer. Even at the low end, a new R45 000 to R60 000 battery-and-inverter system earning R4 000 to R7 000 a year takes roughly 7 to 15 years to pay back. Solar MD rates its SS4160 for 12 years or 84 MWh, and many imported packs carry 5 to 10 years. So for a new purchase, the arbitrage case alone rarely clears the warranty period.
When it does make sense
- You already own the battery. If you bought it for load shedding, the cost is sunk. Setting it to charge in off-peak and discharge in peak is free money, about R4 000 to R7 000 a year on Homeflex, as long as you keep a backup reserve you are comfortable with. See grid charging before outages for the settings.
- You add panels. Then the battery stores free midday solar instead of 253c grid power. The same 9 kWh delivered in a summer peak now avoids about R41 a weekday with no charging cost, and it also works on weekends. That is why our do I need a battery answer starts with panels.
- You cannot install panels. Flats, heritage roofs and heavily shaded sites sometimes have no choice. A smaller pack sized to your actual peak use, around 5 kWh, beats a 10 kWh pack that is never fully discharged.
Rules and settings to get right
- Permission. A grid-tied battery inverter is still an inverter connected to the network. Most distributors, Eskom included, expect an application and NRS 097-2-1 compliant equipment even if you never export. Confirm with your supplier before installing; our registration answer explains the process.
- Charge window. Programme grid charging for off-peak only. Charging in standard hours costs about 55c to 68c more per kWh and destroys the margin.
- Seasonal clocks. Homeflex peak times differ between seasons and changed this year. Update the inverter’s discharge windows on 1 June and 1 September.
- Reserve. Keep a minimum state of charge for outages and battery health, often 20% to 30%. Every kWh held in reserve is a kWh that does not earn arbitrage, so choose the number deliberately.
- Cycle wear. One full cycle a weekday is about 260 cycles a year. Check how your warranty counts cycles or energy throughput; our cycles and warranty answer shows how to read it.
The bottom line
A battery on its own is a backup device that can earn a modest return on a time-of-use tariff, mostly in winter. It is not a stand-alone savings product on flat tariffs at all. If savings are the goal, panels come first and the battery second. If you already own one, set it up properly and it will quietly take a few thousand rand a year off an Eskom Homeflex bill.
Related: Homeflex with solar · usable kWh · battery comparison · bill still high?
Quick questions
Is it worth charging my battery from Eskom at night?
Only on a time-of-use tariff and only if you discharge it in peak. On flat or block tariffs grid charging saves nothing and loses about 10% to charging losses.
How much can a 10 kWh battery save on Homeflex without solar?
Roughly R6 000 to R7 000 a year in the best case on 2026/27 rates, and closer to R4 000 if the home uses only about 5 kWh in peak hours.
Do I need permission for a battery with no panels?
Usually yes, because the battery inverter is connected to the grid. Confirm with Eskom or your municipality and use NRS 097-2-1 compliant equipment.
Sources & related
- Eskom Tariffs and Charges Booklet 2026/2027 (Homeflex tables)
- Eskom Distribution: 2026/2027 tariff increase and structural changes
- City of Cape Town 2026/27 Budget Annexure 6 (residential TOU)
- Solar News 24: Homeflex export and Gen-offset table