17 August 2026

Who Really Pays When Eskom Goes Green?

Inside South Africa's first-loss infrastructure guarantees — and who actually holds the risk.

South Africa has a R13 trillion infrastructure problem and no clean way to pay for it. Government doesn't have the money. Eskom, under the terms of its own debt-relief deal with Treasury, isn't even allowed to borrow more to build new generation capacity. So the country has turned to a financing trick borrowed from global development finance: get private investors to put up the money, and have government quietly absorb the risk of it going wrong. The mechanism is called a first-loss guarantee, and understanding how it works is the key to understanding who actually stands to gain — and who's left holding the risk — as South Africa tries to build its way out of an energy crisis.

The gap that started it all

The scale of the problem is worth sitting with. A joint study by the Development Bank of Southern Africa and the World Bank put South Africa's total infrastructure financing shortfall at roughly R13 trillion. Energy transmission alone accounts for a huge slice of that: the country needs about 14,000 kilometres of new transmission line, at an estimated cost of R450 billion, just to move renewable power from where it's generated to where it's needed. Government officials have been blunt that public money can't cover this. Private capital has to fill most of the gap — the question has always been how to make that capital willing to show up.

What "first-loss" actually means

Here's the plain-language version. Imagine an infrastructure project needs R1 billion in financing. If Treasury commits to a 20% first-loss position, it means: if the project underperforms, defaults, or fails outright, the first R200 million of losses come out of Treasury's pocket — before a single rand of private investor money is touched. Only once losses exceed that 20% cushion does private capital start losing anything at all.

This is exactly the structure South Africa has built for transmission investment. Treasury has committed an initial US$100 million to a Credit Guarantee Vehicle designed to take that first-loss position, with the goal of unlocking R10 billion in financing from development finance partners. In practice, this means private and development-finance lenders sit in a protected, senior position — they get paid first if the project succeeds, and they're shielded from losing money unless the project fails badly enough to wipe out Treasury's entire cushion.

It is not a risk-free deal for private capital — losses beyond that 20% still land on them — but it is a substantially de-risked one. That's the entire point: government is deliberately absorbing the part of the risk that private investors find most unappealing, so they'll accept financing terms government can actually afford.

Eskom Green: the same logic, applied to a utility

Eskom Green, a wholly owned Eskom subsidiary approved for its own project financing in mid-2026, runs on a related version of the same idea. Because Eskom itself is barred from taking on new generation debt under its debt-relief conditions, Eskom Green was built as a ring-fenced, project-finance structure: each renewable project sits inside its own special purpose vehicle, with debt and risk contained to that vehicle. Eskom's exposure is limited to the equity it puts in — if a project fails, the losses don't flow back to Eskom's main balance sheet or, by extension, to the national fiscus in the same direct way a guarantee would.

The purpose is explicit: Eskom's own leadership has said the structure exists to "crowd in" external capital and expertise, letting private investors partner on utility-scale renewable projects that Eskom currently can't finance on its own books. It's a different mechanism from the Credit Guarantee Vehicle — ring-fencing rather than a direct loss guarantee — but the underlying goal is the same: make it structurally safer for private capital to say yes.

Who's protected, and who isn't

Laid out simply, the risk stack looks like this:

• Treasury / the fiscus: absorbs the first losses, up to the guaranteed percentage. This exposure is real but contingent — it only costs money if a project actually fails.
• Development finance institutions and private lenders: sit above that first-loss layer, protected from moderate losses but still exposed if problems are severe enough to exceed the guarantee.
• Eskom's core balance sheet: shielded from renewable project failures by the ring-fenced SPV structure, so a failed Eskom Green project shouldn't drag down the utility's broader finances or force further bailouts through that channel.
• Ordinary South Africans: exposed indirectly, in two ways — as taxpayers, if guarantees are triggered, and increasingly as pension fund members, since Regulation 28 of the Pension Funds Act was amended specifically to allow retirement funds to allocate up to 45% of assets to infrastructure, meaning some of the "private capital" in these deals is, in practice, retirement savings.

The cost that doesn't show up on a budget line

This is where the structure gets genuinely tricky to evaluate, and it's the central reason these deals deserve more public scrutiny than they typically get. A first-loss guarantee doesn't cost government anything up front — it's a contingent liability, not a cash outlay. It only becomes a real cost if a project actually fails and the guarantee gets called. That means the fiscal exposure is real, but invisible, right up until the moment something goes wrong. There's no line item in a budget that says "cost of subsidizing private infrastructure risk this year" — because most years, that cost is zero. It only turns into a number when a solar farm underperforms, a transmission project stalls, or a borrower defaults.

That invisibility is precisely what makes the arrangement attractive to government in the short term, and precisely what makes it hard to hold anyone accountable for in the long term. There's no annual reckoning, no obvious moment to ask whether the terms of the guarantee were generous or measured, fair to taxpayers or overly generous to private capital — until a failure forces the question.

The open question

So is it worth it? That depends on a comparison nobody has published a clear answer to: does the value of getting transmission lines and renewable capacity built now — sooner than government could otherwise afford — outweigh the implicit subsidy handed to private capital in the form of reduced risk for a return that doesn't fully reflect that reduction? If the projects succeed, the guarantees cost nothing and South Africa gets infrastructure it badly needs. If they don't, taxpayers and — depending on how deeply pension funds are involved — ordinary savers absorb losses that were largely invisible until the moment they weren't.

Nobody has run that number publicly yet. Given how much of South Africa's energy future now depends on this financing model working as intended, that's the number worth asking for.