South Africa's Paradox: Eskom Fixed the Grid, Then Demand Collapsed

TechCentral reported this week that South Africa generated less electricity in June 2026 than in any June in at least seven years, including June 2020 when the economy was shuttered under Covid restrictions.^[1] The paradox is stark: Eskom's coal-fired power stations produced 92.1 TWh in the first half of 2026, down 9.5 percent year on year, while the utility's Energy Availability Factor hit 82.04% on 26 July 2026, its best single day since 2017.^[1] Unplanned outages averaged 47.8% less than the same week a year earlier.^[1] The utility has, in other words, repaired its fleet into a collapsing customer base.

What happened is the mechanism every utility fears: a combination of price and reliability pushed customers to exit the grid. During the rolling blackouts of 2022 and 2023, South Africa's distributed solar capacity began climbing as businesses, mines, and households installed backup generation. Once Eskom stabilized the supply, those customers did not leave; they stayed offline, feeding their own panels and batteries into the demand. The result is what economists call a death spiral, but what grid operators recognize as the cost structure finally catching up with the generation fleet.

Eskom's own tariff increases drove much of this. The utility has raised electricity prices for end-users by more than 400% over the past decade, making rooftop solar attractive even before the blackouts began. By the time load-shedding ended, Eskom reported suspending it after 79 consecutive days as of 14 June 2024, then 93 days as of 28 June 2024, the damage to demand was permanent. Independent power producers and private distributed generation had already claimed market share; customers who could afford it had built solar installations sized to run independently of Eskom's network.^[1] Now, demand continues to decline even as supply reliability improves.

This is not a South African exception. It is the mirror that utilities across the world are beginning to face. Eskom's experience, fixing the system only to find the customer base had moved on, is the same path that grids in Australia, California, and parts of Europe are traveling. Australia, which has rooftop solar on roughly one in three freestanding homes, has watched its grid operator step up frequency support and voltage management spending as distributed generation grows; the grid is more stable than ever, but utility revenue from distributed customers has fallen sharply. The difference is that Australia built the infrastructure to absorb that transition; South Africa, still dependent on a centralized coal fleet, did not.

The mechanism at work in Eskom's collapse is regulatory permission for self-generation combined with tariff levels high enough to justify the upfront cost of solar. South Africa's net metering and feed-in rules allowed grid-connected rooftop systems to export surplus power back to the network, making the economics of distributed solar more attractive. Once the blackout crisis created the urgency, price and reliability aligned to push customers out. Eskom's recovery plan fixed the supply problem but did not fix the tariff structure that created the incentive to leave.

This is the hard lesson for utilities everywhere: a customer who has installed solar and batteries does not need your grid to be reliable. They need it not to exist as a dependency. Eskom fixed the fleet. It did not fix the tariff, the ownership structure, or the regulatory framework that would have kept customers inside the boundary. The electricity generation fell because the customers had already left.