Is solar still worth it for a business in Gauteng? What comes off the bill in 2026/27, and what stays

Yes: on the tariffs most Gauteng businesses buy on, a unit of grid electricity averages R2.02 to R4.88 per kWh across the year, excluding VAT, in 2026/27 (Table 1), and the daytime units that solar replaces on City Power’s three business tariffs average R2.81 to R4.88 (Table 2). A unit of rooftop solar costs about R1.25 under a power purchase agreement and about R0.97 over the life of a system you own. Against City Power’s tariffs that is a saving of 56 to 74% on every replaced unit under a PPA and 66 to 80% if you own the system (Table 2), and the gap widens with every July increase.

The limit is coverage, not price. A grid-tied system without batteries replaces only the units used while the sun is up, typically 30 to 50% of a site’s municipal kWh. The rest is consumed at night and on weekends, and only batteries can move daytime energy into those hours. Solar also leaves the fixed and demand charges untouched, so the bill after solar is smaller, not gone. The return on the units it does replace is still very good, and the numbers below are the 2026/27 schedules, not estimates.

What does a business kWh cost in Gauteng in 2026/27?

The energy charge alone, ex VAT, from our verified references for City Power, Ekurhuleni and Tshwane. City Power figures include its 6c/kWh network surcharge and 2% account surcharge; the other two metros have neither. Winter is June to August.

Table 1. Energy charge per kWh, 2026/27, ex VAT

The average column is the energy rate across the year for a flat load, the same kWh every hour, with nine summer and three winter months. On the time-of-use tariffs it is averaged over each metro’s clock, which on all three works out to 25 peak, 64 standard and 79 off-peak hours a week. Tshwane’s figures are the gazetted 2025/26 rates escalated by the promulgated 8.80%, pending the 2026/27 gazette. Ekurhuleni’s are the NERSA-approved schedule the city implemented on 1 July 2026 and has taken on review.

Table 1 and its average column cover the energy charge only. Fixed monthly charges and kVA demand charges sit on top of it, and solar does not change them. The full breakdowns are on the City Power, Ekurhuleni and Tshwane tariff pages.

What is a solar kWh worth, and what does one cost?

A solar unit used on site is worth the full energy rate of the hour it replaces, surcharges included. Rooftop systems generate between about 07:00 and 18:00, so that is a peak or standard rate, never an off-peak one.

The solar blended tariff is the average of those hourly rates weighted by when a rooftop system produces: what City Power charges, per kWh, for the units solar replaces. On the time-of-use tariff it is R2.81 across the year (R3.07 for a weekday-only site), above the R2.62 average unit in Table 1; on the flat tariffs it equals the flat rate. Solar itself costs about R1.25 per kWh under a PPA and about R0.97 over the life of an owned system.

Table 2. What City Power charges for the kWh solar replaces, against what solar costs per kWh, year one, ex VAT

Table 3. Owning the system: simple payback and year-one return per tariff, on our published budget figures, ex VAT

The gap widens every year: the grid rate is reset by NERSA (9.01% in July 2026, 8.84% approved for 2027), the PPA rate escalates on the terms in the agreement, and an owned system’s cost per unit is fixed the day it is built. A unit from your own roof is the cheapest unit a business in Gauteng can buy, and the only one whose price is not set by a regulator.

Three of our published Johannesburg projects show systems of this kind on real roofs: Kimbult Industrial Park at 547 kWp, Hillcrest Retail Centre at 145 kWp and Canero Office Park at 124 kWp.

Solar blended tariff: City Power’s 2026/27 energy rate, surcharges included, weighted by a Johannesburg rooftop’s hourly output from sun geometry across the seven-day clock, equal monthly production, nine summer and three winter months; energy charges only. PPA rate: the illustrative R1.25/kWh in the labelled example on our finance page; actual rates and escalation are set per site in the agreement. LCOE: our published budget figure of R7,850 per kWp ex VAT for a 200 kWp grid-tied system without batteries, 1,450 kWh per kWp in year one, 0.5% annual degradation, running costs of 3.75% of capital a year rising 6%, 25 years, costs and energy both discounted at 10%; R0.80 to R1.07 across the 100 to 500 kWp bands and lower running costs; about R0.80 after the Section 12B deduction at 27%. Payback and return: all solar energy consumed on site at the solar blended tariff, 1,450 kWh per kWp in year one, the published budget band for the representative size, running costs 3.75% of capital; payback = capital divided by the year-one saving less running costs; return = year-one saving divided by capital, before tax and before the Section 12B deduction. A design-stage illustration, not a quote.

Can I sell the extra solar power back to the grid?

Plan as if you cannot, and size the system for self-consumption. City Power currently requires zero export until its feed-in process exists. Tshwane pays its credit only after City approval. Ekurhuleni pays R1.05 to R1.53 per kWh to registered net consumers with a four-quadrant meter, well below the import rate. We count exported energy as zero in every business case.

How much of the bill can solar remove?

• Energy line: typically 30 to 50% of the site’s kWh without batteries, across the systems we operate. Daytime, weekday operations sit at the top of the range; 24-hour and evening-heavy operations at the bottom. Night and weekend units stay on the grid.
• Demand line: solar lowers it on many sites, typically by about 10% across our fleet (5 to 25%), and we never count it. One cloudy afternoon or a tripped inverter restores the peak, and City Power’s floor holds it for a year.
• Fixed lines: unchanged. In Johannesburg a prepaid business must move to conventional billing to register a generator, which adds monthly service and capacity charges.

The whole-bill saving is therefore always a smaller percentage than the kWh saving. A worked example: a 200 kWp system on City Power’s Large Power User demand tariff at a site using 60,000 kWh a month with 250 kVA billed demand, in a summer month. The system supplies about 24,167 kWh, all used on site, 40% of the units.

Table 4. A 200 kWp system on City Power’s LPU demand tariff: one summer month before and after, ex VAT

Solar took 40% of the units off the bill and the bill fell by 24.1% owned, or 14.0% under the PPA once the solar units are paid for, because the demand line and the night-time units did not move. The owned saving is what the capital in Table 3 buys; winter runs the same way at the higher winter rate.

Inputs are illustrative: 60,000 kWh a month, 250 kVA billed demand, production of 1,450 kWh per kWp a year spread evenly over twelve months, every solar unit consumed on site, the demand charge unchanged, and the PPA at the finance page’s illustrative R1.25/kWh. Rates are the 2026/27 schedule with surcharges; totals are computed from the unrounded rates.

Why is the bill still high after solar?

1. Night and weekend units are untouched.
2. The demand charge did not move, and Johannesburg’s 80% floor remembers last year’s peaks.
3. The 9.01% July increase landed on the remaining units and the fixed lines in the same year.
4. A prepaid account became a conventional one, with new service and capacity charges.
5. A reactive-energy line appeared: City Power bills reactive energy above 30% of the month’s kWh when power factor is below 0.96, and solar cuts kWh without cutting reactive draw. A correctly configured inverter compensates it; check the configuration before buying correction equipment.
6. The system is producing less than it should: dirt, shade, a failed string, a tripped inverter, or a zero-export setting discarding energy from a system sized beyond the daytime load. This one needs measurement against a weather-adjusted expectation, not explanation.

What do batteries change?

They move daytime energy into the evening and night, so the share of units solar replaces rises above 30 to 50%, and they can shave the demand peak dependably, which panels cannot. A stored unit costs more than a daytime solar unit, so a battery is sized for a specific job (a peak, a ride-through, or night supply) and priced against the line of the bill it targets.

Did the end of load shedding change the maths?

No. Eskom last implemented load shedding on 16 May 2025 and had run 441 days without it by 4 August 2026. A grid-tied system without batteries never generated during load shedding, because it must disconnect when the grid fails. The case was always the price of a daytime unit, and that price moved three ways in 2026:

• Eskom-direct customers paid 8.76% more from 1 April 2026, municipal customers 9.01% more from 1 July 2026, and NERSA has approved 8.83% (Eskom-direct) and 8.84% (municipal) from 1 April 2027. Stats SA’s producer price index put electricity at 8.3% year on year in July 2026 against 5.7% for manufactured goods.
• Fixed charges are growing by design. Eskom said on 31 August 2026 that it “will retain revenue through network charges, customer wheeling and revised tariff structures, which separate energy charges from the recovery of fixed costs.” Solar reaches the energy charge only.
• Equipment stopped getting cheaper. Chinese module prices rose more than 30% in early 2026 before settling.

What to do next

Gather twelve months of bills, or your interval meter data, and have the saving modelled against your actual tariff schedule, period by period. Ask for the kWh saving and the bill saving as two separate numbers, with the working.