Solar Grid Capacity in South Africa: What Durban Pays

Your July electricity bill from eThekwini has arrived. It is 9% bigger than June’s. You didn’t buy a machine. You didn’t add a shift.

Nothing in your business changed. The bill changed anyway.

Here’s where it gets tangled. Someone may have told you the increase was 8.76%, back in April. That figure is real. It just isn’t yours.

That mix-up is a clue about something larger. Solar grid capacity in South Africa is now the slow part of a project, not the easy part. And the tariff clock doesn’t pause while you queue.

Your Increase Was 9% on 1 July. Not 8.76% on 1 April.

Two different suppliers. Two different dates. Two different numbers.

Megaflex, Nightsave Urban, Ruraflex and Businessrate are Eskom-direct tariffs. Eskom supplies those customers and bills them itself. Picture a packhouse on a rural feeder, outside the metro line.

Those customers moved on 1 April 2026. Their approved average increase was 8.76%.

eThekwini works differently. The city buys power from Eskom in bulk, then resells it to you. Eskom raised that bulk price by 9.01% on 1 July. eThekwini passed 9% on to end users, cut back from the 10.5% it first proposed.

If eThekwini bills you, you are not on Megaflex. Not in the Durban CBD. Not in Pinetown, Mobeni, Prospecton or Riverhorse Valley. Not in most of Umhlanga.

You’re on an eThekwini tariff. Its own rate structure. Its own time-of-use windows, where they apply. Its own increase date.

So here is the practical test. Any model built on Eskom’s published price list will be wrong for you. The rates differ. The fixed and demand charges differ.

A Durban payback built on Megaflex rates describes someone else’s business. Ask to see your own tariff in the model. If nobody can produce it, you’ve learned something.

Solar Grid Capacity in South Africa: What I Won’t Pretend to Know

Connecting a commercial solar system takes longer than it did five years ago. The reasons are built in, not passing.

Your network was built to move power one way. Substation to you. One direction, like a river.

Solar on your roof reverses that flow on some feeders, for part of the day. So the utility must check things it used to take for granted. Protection settings. Fault levels. Voltage.

That means process. An application. A technical review. A signed connection agreement.

Then a compliance pack. NRS 097-2-1 sets how your inverter must behave at the grid. SANS 10142-1 is the wiring code your job must meet. Sub-100 kW systems register with NERSA. Larger ones face more scrutiny.

None of that is unfair. All of it takes calendar time.

Now here’s what I won’t do. I won’t quote you a queue length.

You have seen the claims. “The average connection now takes X months.” They travel widely. I cannot trace them to a published source.

They also vary. By city. By feeder. By whether your substation has room to spare. By how complete your first application was.

Anyone quoting one national figure is guessing. Don’t build a project plan on a guess.

What I can tell you comes from doing the work. The files that move fastest go in complete. Single-line diagram. Inverter certificates. Protection settings. Load profile. First time.

The ones that stall go back and forth three times over missing paper. That part is entirely in your hands. It is the cheapest schedule risk you will ever remove.

Still picking a partner? Ask who carries the approval work, and who has lodged with your city before. Our guide to choosing a solar company in Durban has more.

What a Slow Start Actually Costs

I can’t quote a queue length. But I can price the other side of it. What is one month of drift worth?

This is a worked example, not a quote. The assumptions are listed so you can swap in your own.

Assumptions — Terawatt modelling inputs, not market facts:

• A Durban maker on an eThekwini supply. The bill was R250,000 a month excl VAT before 1 July. An illustrative figure, not an average.
• A daytime-heavy load. Single shift. Machines running while the sun is up.
• A right-sized rooftop array offsets 30% of the account. This is an assumption, and the one most sensitive to your real load shape.

Step 1 — the July increase. 9% of R250,000 is R22,500 a month. Over a year that is R22,500 × 12 = R270,000. Same kilowatt-hours as June. The new baseline is R272,500 a month.

Step 2 — what the array is worth. 30% of R272,500 is R81,750 a month avoided. That is R981,000 a year.

Step 3 — the cost of drift. Every month between “we should do this” and switch-on is R81,750 you don’t bank. A six-month slip is 6 × R81,750 = R490,500.

That money isn’t deferred. It is spent. On electricity. Permanently.

Now the caveat I’d rather state than bury. A flat 30% of the bill total is a simplification.

Your bill has two parts. Energy is how many litres of water you drew over the month. Demand is how wide the pipe had to be at your worst moment. You pay for both.

Solar cuts litres. It rarely narrows the pipe.

It does little for fixed charges. And it only trims a demand charge if it’s making power when your peak lands.

So a real study works off a line-by-line bill, not a share of the total. Two sites with the same roof can get very different answers. For a first pass on your own numbers, start with the savings calculator.

One honest disqualifier. If Eskom bills you direct, none of this is yours. Your increase landed in April. Your rates sit in Eskom’s own schedule. The sums above describe a different customer.

The Increase Worth Modelling Is Next April’s

This part got very little airtime. It’s also the strongest reason to start now.

NERSA got Eskom’s regulatory asset base wrong by R54.7 billion. That base is the value of Eskom’s assets, used to set what the utility may earn.

A High Court order forced public consultation. The regulator then chose to claw the error back through tariffs, over three years. R12 billion in 2026/27. R23 billion in 2027/28. R19.7 billion in the period after that.

Now the useful bit. Without that recovery, the 2026/27 increases would have been 5.36% for Eskom-direct customers and 6.19% for cities. Instead they were 8.76% and 9.01%.

So on the municipal side, R12 billion of it is the gap between 6.19% and 9.01%. That is 2.82 percentage points of your July increase.

Next year’s instalment is R23 billion. Roughly double.

I won’t predict a percentage. Too much rests on Eskom’s next revenue bid and what eThekwini does with it. But the direction is not in doubt.

The instalment you have just absorbed was the small one.

Which leads somewhere unusual for this trade. If you model your solar payback on today’s tariff, you are being cautious, not hopeful. Our breakdown of solar and tariff increases shows why a rising tariff shortens a payback instead of stretching it.

How to Move Now Without Overcommitting

Moving early doesn’t mean signing for the biggest system on offer. It means starting the slow parts while you still have choices.

1. Get a proper load profile first. Half-hourly data for a year if you can get it. Six bills at minimum. Every later choice rests on it, and it costs only a request to your meter provider.
2. Send a complete application, not a fast one. A complete first file beats an early rough one every time. The rework loop is where schedules die.
3. Size to your daytime load, not your roof. Building past an export limit you will never be allowed to use is the priciest mistake in commercial rooftop. Our note on sizing a system to your business sets out how we cap it.
4. Pick inverters that can take batteries later. You don’t have to buy storage now. Just don’t buy hardware that rules it out. The demand side of your bill needs its own fix.
5. Phase it on purpose. A smaller system running this summer earns while a bigger one is still in review. Phase two is far easier on a live connection.
6. Check the tax treatment before you model it. There is a faster write-off for renewable-energy and storage assets. The rate and who qualifies depend on the asset, its commissioning date and its use. Ask your tax advisor first.

Frequently Asked Questions

Did Durban businesses get the 8.76% Eskom increase in April?

No. The 8.76% applied to Eskom-direct customers from 1 April 2026. eThekwini-supplied businesses got 9% from 1 July 2026, cut back from a proposed 10.5%. If the city bills you, then July is your date and 9% is the number that applies.

Am I on Megaflex if my business is in Durban?

Almost certainly not. Megaflex is an Eskom-direct tariff. If eThekwini issues your bill, you are on a city tariff with its own rates and its own increase date. Check the name at the top of your account. That tells you which schedule applies.

How long does grid connection approval take in eThekwini?

There is no reliable single answer, and we won’t invent one. It turns on your feeder, your substation’s spare room, the size of your system and how complete your file is. Incomplete applications are the most common cause of delay, and the easiest to avoid.

Should I wait for panel prices to drop further?

Panels are a shrinking share of a commercial project. The rest is structure, cabling, protection, inverters, labour and compliance. Meanwhile the tariff you’re avoiding climbs each year, and next year’s NERSA error recovery is roughly double this year’s. Waiting usually costs more than it saves.

Does solar reduce my demand charges as well?

Only by accident. Solar cuts the energy you buy while it is making power. If your billed peak lands early in the morning or in the evening, an array won’t touch it. That’s a storage job. See our guide to battery storage and peak shaving.

Sources: Eskom, “Eskom implements NERSA decision for Financial Year 2027”; Engineering News, 10 March 2026, “Nersa approves higher electricity hikes for 2026/27 after R54.7bn error”; eThekwini’s approved 2026/27 tariffs. Rand figures in the worked example are Terawatt modelling inputs, not published rates.

Allen Meyer

Director of Engineering · Terawatt Energy

Writes the Insights blog when not on a site.

LinkedIn More from Allen Meyer