For years we asked: “When will load shedding end?”
The better question is: “How did it begin?”
Because load shedding did not befall South Africa like a drought. It was built. Piece by piece. Through delay, looting, and arguments. And today in August 2026, we are still paying the bill.
1. HOW IT BEFELL US – THE CHRONICLE OF CHOICES
The origins of South Africa’s electricity crisis cannot be attributed to a single event or external shock.
Load shedding is the culmination of more than two decades of deferred decisions, institutional decay, and structural neglect. It did not befall the country by accident. It was produced, incrementally, through choices made in boardrooms and in parliament.
• First, the state stopped investing, and when it eventually did invest, it did so badly. In the late 1990s, government and Eskom determined, correctly at the time, that the country had surplus generation capacity. On that basis, the commissioning of new baseload power stations was postponed indefinitely. That assessment failed to account for sustained economic and population growth. While demand increased by approximately 3 to 4 percent each year, no significant new capacity was added for over a decade. By the late 2000s, demand had overtaken supply and the response was reactive. Medupi and Kusile were launched in haste. Both were scheduled for completion by 2012. Both were delivered roughly ten years late and at several hundred billion rand above their original budgets. More damaging than the delay and cost overruns were the design and construction defects that rendered the plants unreliable from the outset. This was the product of state capture, inflated procurement contracts, and the departure of skilled engineers from Eskom. The outcome was stark: the country built too little, and what it built was broken.
• Second, the institution responsible for electricity generation was systematically hollowed out. Between 2010 and 2018, Eskom ceased to function primarily as a public utility. It operated instead as a conduit for patronage and extraction. Maintenance budgets were raided. Coal supply agreements were inflated far above market rates. Governance structures were compromised. Experienced technical staff exited in large numbers and were not replaced with equivalent expertise. The effects of this were not immediately apparent because power infrastructure fails gradually and then suddenly. Deprived of preventative maintenance for years, the generation fleet deteriorated until catastrophic failure became inevitable. From 2018 onward, this manifested in frequent boiler tube leaks, unplanned unit trips, and extended outages. In effect, the country consumed the capital of the system that sustains it.
• Third, policymaking was paralyzed by ideological dispute at the precise moment that decisive action was required. While Eskom’s performance declined, government spent the better part of a decade contesting the role of private power producers. From 2012, independent companies, municipalities, and communities were ready to deploy solar and wind at scale. Yet regulatory licensing was capped first at 1MW and later at 10MW. Procurement rounds under the Renewable Energy Independent Power Producer Programme were repeatedly delayed. Rules governing grid access remained unclear. This was a deliberate choice of process over delivery. During the same period, countries such as Vietnam installed more than 20GW of solar capacity. South Africa produced paperwork. The central question became whether electricity should remain the exclusive domain of Eskom. The engineering question — where can we get power now — was sidelined.
Beneath these three failures lie four structural vulnerabilities that continue to undermine the system.
• The first is dependence on aging coal infrastructure. Approximately 80 percent of electricity is still generated by coal plants with an average age exceeding 40 years. These stations were not designed to operate continuously at high load factors this late in their lifecycle.
• The second is municipal fiscal collapse. Municipalities owe Eskom more than R80 billion. This debt deprives the utility of revenue required for maintenance and reinvestment, while weakening incentives for fiscal discipline.
• The third is large-scale theft and vandalism. The theft of copper cable and transformers, together with widespread illegal connections, results in the loss of gigawatts before electricity reaches consumers. It also erodes revenue and increases pressure on the remaining network.
• The fourth is the absence of an adequate reserve margin. A resilient grid requires roughly 15 percent of capacity to be held in reserve. For many years South Africa operated with less than 5 percent. In such a system, the loss of a small number of generating units immediately triggers extensive load shedding, often escalating to Stage 4.
Taken together, these factors explain how load shedding came to define daily life. It was not caused by weather or by external forces. It was the direct and foreseeable consequence of postponed investment, institutional looting, regulatory delay, and the neglect of basic engineering principles. To reverse it will require not new rhetoric, but a different sequence of choices: to build, to maintain, to enforce, and to allow power to be generated wherever it can be done reliably and at speed.
2. WHERE WE ARE IN A PERIOD OF MANAGED INSTABILITY
We are neither in Stage 6 as we were in 2023, nor have we reached a point of consistent supply. The present condition is better described as managed instability.
The load shedding schedule has been integrated into daily life. It is saved on phones, referenced in meetings, and built into household and business planning. Inverters have become standard installations. Generators are treated as capital expenditure rather than emergency purchases. This is widely referred to as progress because outages are less frequent and less severe. Yet the underlying structural deficit remains.
The relative improvement since 2023 can be explained by three developments.
• First, a significant portion of demand has migrated off the central grid. Rooftop solar installations, private generators, and deliberate demand reduction by commercial and industrial users have lowered the overall load that Eskom must meet at any given time. This shift was driven not by policy but by necessity, and it has had the incidental effect of easing pressure on the system.
• Second, renewable energy projects that were delayed for years have begun to reach commercial operation. Additional capacity from wind and solar farms procured under the Renewable Energy Independent Power Producer Programme is now feeding into the national grid. This has been complemented by a growing amount of battery storage, which provides short-term stability and reduces the need for immediate load shedding during peak periods.
• Third, Eskom has changed its maintenance approach. Instead of running units to failure, the utility is now taking plant offline in a planned manner for repairs and refurbishment. While this reduces available capacity in the short term, it is intended to improve reliability over the medium term and to reduce the frequency of unplanned breakdowns that previously triggered higher stages of load shedding.
Despite these gains, the fundamental weaknesses of the system have not been resolved. The generation fleet remains old, with most coal-fired stations operating well beyond their intended design life. Municipal debt to Eskom continues to grow and constrains the utility’s ability to invest in upkeep and infrastructure. The transmission and distribution network remains vulnerable to theft, vandalism, and underinvestment.
As a result, the system has not stabilized. It oscillates. There are periods of several weeks with little or no load shedding, followed by abrupt escalations to Stage 3 or Stage 4 when multiple units trip simultaneously. This pattern reflects a grid with no meaningful buffer.
In this sense, South Africa has moved from acute crisis to chronic condition. The emergency has receded, but the risk has not. We are managing the symptoms more effectively, while the underlying disease persists.
3. THE IMPACT: IT IS A REGRESSIVE TAX
Technically, load shedding is a form of emergency intervention. When electricity demand exceeds available supply, the frequency of the national grid begins to fall. If that decline is not arrested, the entire system can collapse into a blackout that may take weeks to restore. To prevent this, Eskom reduces load in predetermined stages, rotating outages across regions in order to preserve the integrity of the network as a whole. In engineering terms, it is the least damaging option available in a constrained system.
In human terms, however, load shedding functions as something else entirely. It operates as a tax, levied not through legislation but through lost hours, increased costs, and diminished dignity. And like most regressive taxes, its burden falls heaviest on those least able to absorb it.
For individuals and households, the consequences are immediate and material. Clinics lose refrigeration for vaccines and essential medicines. Learners lose study time after sunset because homes cannot be lit and devices cannot be charged. Elderly and disabled persons are confined in dark, cold, and sometimes unsafe conditions. When streetlights and traffic signals fail, the risk of crime and accidents increases. Beyond these specific harms, there is a broader erosion: the basic expectation that a modern state will provide power sufficient to cook, to wash, and to see after dark is no longer met as a matter of course.
For the economy, the impact is both direct and cumulative. Small enterprises such as spaza shops, panel beaters, and hair salons cannot migrate to remote work or operate on battery backup. Four hours without electricity is four hours without revenue. In manufacturing, machines must be shut down and restarted, which damages equipment and wastes production time. Across sectors, firms divert capital from expansion and employment toward generators, inverters, and fuel. Those additional costs are ultimately passed to consumers in the form of higher prices for food, transport, and data. At a macro level, unreliable power deters investment. No investor commits capital to long-term industrial projects in an environment where electricity supply cannot be guaranteed.
For society, the effect is corrosive and divisive. The response to outages has created two parallel systems of energy access. Those with means install solar, batteries, and generators. Those without means rely on candles, paraffin, and interrupted supply. This divergence hardens inequality and produces two grids: one that functions because it can pay its way out, and one that remains dependent on a failing public system. More fundamentally, repeated failure undermines the social contract. When the state cannot perform its most basic function of keeping the lights on, public confidence in its capacity to deliver any other essential service declines in tandem..
4. THE UNCOMPROMISING TRUTH AND THE WAY FORWARD
This crisis was not imposed upon us by forces beyond our control. It was produced by a sequence of choices made over two decades, and it will only be resolved by a different sequence of choices made consistently from this point forward. The work of repair has begun, but it is proceeding slowly, and it will remain slow because of the nature of the infrastructure involved.
Any assertion that load shedding will end within weeks is not credible. Equally, any claim that it will persist indefinitely is not supported by current evidence. The more accurate assessment lies between these positions. The frequency and severity of outages will diminish gradually, and unevenly, over several years. This is a function of lead times. New generation capacity, whether from renewables, gas, or storage, requires 3 to 5 years to plan, finance, and commission. Institutional credibility, once lost, requires even longer to restore.
Given that reality, four priorities must guide policy and implementation now.
• First, the pace of new build must increase substantially. Additional renewable energy, battery storage, and private generation must be brought onto the grid with urgency. The requirement here is for execution, not further policy announcements. Projects that are bankable and ready must be connected, and regulatory bottlenecks must be cleared.
• Second, the distribution system must be stabilized. Municipalities that are owed revenue must improve collection. Losses from theft, illegal connections, and vandalism must be reduced through enforcement and better infrastructure protection. A grid that leaks both electricity and revenue cannot be financially or technically sustainable, regardless of how much new generation is added upstream.
• Third, essential services must be shielded from the effects of load shedding. Hospitals, schools, and police stations should either be exempted from scheduled outages or provided with guaranteed backup supply. The protection of these institutions is not a matter of convenience. It is a matter of public safety and basic state function, and it should be treated as non-negotiable.
• Fourth, communication must be grounded in fact. The public is entitled to accurate data on plant performance, realistic timelines for new projects, and clear accountability when targets are missed. The cost of misinformation and denial exceeds the cost of the outages themselves, because it destroys the trust required for any long-term plan to succeed.
In short, the path out is known. It requires building at speed, collecting what is owed, protecting the vulnerable, and telling the truth. None of these are technically complex. All of them are politically difficult.
Technically we are not in Stage 4 or 6. The schedule may even say “no load shedding today”. But for millions of households and businesses the experience at the end of the day is the same – work cannot be done, appliances cannot run, and the day is lost.
This is because weak power and load shedding produce the same outcome through different mechanisms. Load shedding removes supply completely for a set period. Weak power leaves supply in place but strips it of its usefulness. A fridge that cannot get cold, a machine that cannot start, a student who cannot get light bright enough to read — the result is identical to an outage. The only difference is that with load shedding you are given a time. With weak power you are given uncertainty.
And the cost profile is similar, sometimes worse. With load shedding you budget for fuel, for backup, for lost hours. With chronic low voltage you pay twice: once for electricity that does not perform, and again for the damage it causes. Motors burn out. Circuit boards fail. Food spoils in a fridge that was “on” the whole time. There is no compensation, no rebate, and no stage to cite when you lodge a complaint.
At a system level both are symptoms of the same deficit. There is not enough firm capacity, the network is overloaded, and there is no reserve to hold voltage steady when demand rises. So instead of cutting you off, the system lets the voltage sag. It keeps the statistics looking better while pushing the failure down to the consumer.
In practice then, managed instability has two faces. One is scheduled darkness. The other is unscheduled weakness. Both break the basic promise of electricity: that when you switch something on, it will work.
Until voltage and frequency are stabilized across the network, we have not moved past load shedding. We have just renamed it. The lights may be on, but the country is still not powered.
CONCLUSION: SURVIVING THE NOISE
Load shedding cannot be understood solely as a technical failure of the electricity system. It is, more fundamentally, an examination of the capacity of the state to govern over time. The crisis poses a set of direct questions. Is it possible to plan and invest beyond the horizon of an electoral term? Is it possible to acknowledge failure publicly and to implement corrections without deflection? Is it possible to prioritize the unglamorous work of building infrastructure, maintaining assets, and collecting revenue, instead of pursuing political symbolism?
At present, the answer from society has been adaptation. Households run on inverters. Businesses run on diesel. Schedules are planned around outage timetables. Generators hum through the night and indicator lights blink on walls across the country. This ingenuity has allowed daily life to continue, and it should be recognized. But adaptation is not the same as progress. A country cannot achieve sustained economic growth, industrialization, or social development on the basis of coping mechanisms alone.
What is required is reliable power in both senses of the word. Literally, the grid must supply electricity that is predictable and sufficient. Politically, the state must demonstrate the authority and competence to deliver public goods consistently. One cannot exist without the other.
The condition for stable supply is therefore not technological. It is conceptual. Electricity must cease to be treated primarily as an instrument of political contestation and must instead be treated as an engineering problem. That means decisions based on cost, reliability, maintenance cycles, and load forecasts rather than on patronage, ideology, or short-term messaging.
Until that shift occurs, the current state will persist. We will not be experiencing freedom in any full sense. We will be managing the absence of it, one stage at a time. That is a tolerable condition for survival. It is not an acceptable condition for a country that intends to thrive.