The first electric lights in SA were lit up by the Kimberley municipality in the 1880s, and in the early days of electricity the power generators were the municipalities and the burgeoning mining companies.
Even after Eskom was founded in 1923, it was the Victoria Falls and Transvaal Power Company that continued to supply many of the mines, and it was only after 1948 that SA's power system was nationalised, and ultimately centralised.
The prospect of mining companies and other large power users generating their own power harks back, in a way, to the decentralised power grid of those days.
It's a prospect that has been given impetus at long last by President Cyril Ramaphosa's surprise announcement last week that the government will liberalise the licensing regime for "embedded generation" projects of up to 100MW.
It could radically transform SA's power paradigm - and has far-reaching implications over time for the national grid, and for Eskom and municipalities.
Load-shedding has been the big driver of the growing clamour over the past couple of years for the government to make it easier for companies to generate their own power.
Last year saw a record number of load-shedding hours, even in an ultra-weak economy, and this year looks set to beat that record. SA's unreliable (and increasingly unaffordable) power supply is widely acknowledged to be the biggest constraint on investment and growth.
With Eskom battling to keep the lights on and the power system forecast to be short of at least 4,000MW over the next five years, the utility's own CEO, André de Ruyter, has supported calls to free up companies to generate their own electricity.
Though mineral resources & energy minister Gwede Mantashe long resisted the move, it's surely no coincidence that he finally reluctantly went along with Ramaphosa in a week when load-shedding again went to stage 4.
If combating load-shedding has been one driver of the push to self-generation, combating climate change has been another. Investors are putting companies under ever more pressure to move to cleaner energy. And with the EU looking to impose carbon border taxes on imports from carbon-intensive countries, South African exporters - from miners to car manufacturers - could be in trouble if they can't start shifting away from fossil fuel-fired energy.
All of which means there is potentially plenty of private sector investment ready to go into plenty of projects, large and small. A survey by Meridien Economics indicated that lifting the threshold below which companies would not have to apply for a licence to 50MW had the potential to unlock 5,000MW of additional power capacity over the next four to five years, attracting R75bn-R85bn in investment and creating 16,000 direct jobs.
Meridien's Grové Steyn, who is on the president's Economic Advisory Council and has also been advising business on embedded generation, hasn't modelled the impact of a 100MW threshold, but the potential is clearly larger.
But it's as yet unclear how far the government and the energy regulator will go to do what's needed to unlock the full potential of embedded generation.
It is different to the system of going out to procure independent power through the series of bid windows the government has until now used for the Renewable Energy Independent Power Programme, as well as the recent emergency procurement programme.
In those programmes, the government chooses the producers (imposing conditions such as empowerment and localisation), the National Energy Regulator of South Africa (Nersa) licences them, Eskom agrees to buy the power from them, and the state guarantees Eskom will pay - putting more than R200bn of guarantees on the government balance sheet so far.
By contrast, now, companies that want to build their own power plants can just go ahead and do so, without having to go through the cumbersome and uncertain process of applying to Nersa for a licence and getting the minister's permission, as long as the project is below 100MW.
They will still have to get environmental impact approvals, as they would for any project. They will also, crucially, have to comply with the grid code and get permission from Eskom or a municipality to connect to the relevant grid. And once they have that, they will have to register with Nersa.
The government will not control the process - no doubt one reason Mantashe is so uncomfortable with the idea, along with the fact that ultimately the faster private producers bring new renewable energy onto the national grid, the easier it will be for Eskom to close its old coal-fired power stations, which may be unreliable and dirty, but support mines and whole communities.
But Mantashe still has to publish the relevant legislative amendments to put the new exemption in place. Nersa has to streamline its bureaucracy to expedite the permits. Eskom and municipalities have to streamline the grid connection process.
Embedded generation potentially opens the way to a very different energy market, with complex challenges SA will have to address
And Eskom has to set appropriate tariffs for companies wanting to "wheel" or transport their power across the national grid - as a mine wanting to use the solar resources of the Cape to generate power for its operations in North West might want to do, or, importantly, as a company generating power for more than one mine or manufacturer or shopping centre might want to do.
This is a key issue - that the new rules should let in not only the large power users that want to build their own power plants for their own use, but also power producers that set up plants to supply multiple end-users, as well as so-called aggregators who would buy and on-sell power to companies.
This is essential to "de-risk" new private power projects, as the Council for Scientific and Industrial Research's Jarrad Wright puts it, enabling producers to raise finance at cost-effective rates while taking advantage of economies of scale. That's key to enabling SA to realise the full potential of the private power. And though Mantashe has clearly been reluctant on the multiple customer idea - which was barred in the draft legislation he published when he agreed to a 10MW exemption - last week's media statement from the presidency makes it clear it is definitely in: "Generators will be allowed to sell electricity to one or more end-use customers, on the condition that they are registered and have secured grid connection approval."
South Africans have become so accustomed to central planning and state control that it's hard for many to imagine a system in which the government will not be calling all the shots. But embedded generation is potentially the fastest, cheapest and cleanest way to get large amounts of new capacity onto the grid in a reasonably short time. Once they have regulatory approval, many of these plants would take just 12-24 months to build.
If the government really does make embedded generation happen in a way that unleashes many thousands of new megawatts, the national grid will look very different, and there are large implications.
First, the task of balancing the power system will become much more complex, with multiple generators coming in and out, and users sometimes using their own power and sometimes needing their Eskom or municipal backup.
Too much supply is as much of a problem for a system operator as too much demand, and it will need skilful balancing - and facilities such as utility-scale batteries and Eskom's pumped storage stations will become ever more important.
Crucially, too, such a system will need a stronger national grid - which could require the investment of hundreds of billions of rands in coming years. All of this puts pressure on Eskom to finalise the unbundling of its transmission division into an independent company. A new independent transmission operator will be positioned to attract that investment, and will have the incentive to make the system work for all.
But the other huge issue is the impact it will have on Eskom and municipalities when more of their paying customers start producing their own electricity. Municipalities rely on revenue from those customers to cross-subsidise other services. And in the case of Eskom those most likely to go off-grid, at least in part, are its largest customers.
But Eskom could gain new sources of revenue from providing them with backup power and storage, as well as wheeling. And chances are it's losing more revenue by load-shedding than it will lose in revenue from those customers. If the new private power plants can cut load-shedding, and enable higher rates of economic growth, Eskom could win as much as it loses.
Embedded generation potentially opens the way to a very different energy market, with complex challenges SA will have to address. Now we wait to see the devil in the details. And then the ball will be in the private sector's court to deliver on its power promises.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.