‍

After decades of flirting with the idea of nuclear energy, thwarted by a combination of government incompetence and lawfare from foreign-funded environmentalist NGOs, South Africa has entered into a nuclear cooperation agreement with China.

The deal focuses on small modular nuclear reactors (SMRs), which promise a scalable and incremental development schedule. Aimed at replacing the energy deficit while avoiding carbon regulations, they stand a chance of seriously impacting the problem of loadshedding.

The deal builds on a letter of intent signed in November 2024 with the China National Nuclear Corporation (CNNC). Electricity and Energy Minister Kgosientsho Ramokgopa has pushed nuclear power as vital to the nation’s “just energy transition,” a vision based on Western financing, and articulated in the Integrated Resources Plan (IRP) from 2019. The IRP targets 2500 megawatts (MW) of new nuclear capacity, with plans for two SMRs by 2030.

The exact designs for the SMRs are still not public, but there are two likely models: China’s ACP100 and South Africa’s HTMR-100. The ACP100 is a 125 MW pressurized water reactor (PWR) that can be modified for adding desalination capacity. A demonstration unit is underway in Changjiang, Hainan, with completion slated for 2026. Each unit can produce 1 billion KWh a year, enough for about 526,000 households. The HTMR-100 is derived from the Pebble Bed Modular Reactor (PBMR) tech tree, and is a 35 MW helium-cooled reactor designed for remote deployment. It is still in development, but Stratek Global, its developer, is seeking R10 billion to push it forward.

To hit the 2500 MWe goal, the Department would have to commission 20 ACP100 units, or 70 HTMR-100s. The choice will hinge on cost, scalability, and whether South Africa prefers to empower the Chinese or local producers, though a combination is certainly possible, opening the window to bilateral technology transfer.

Cost estimates are difficult in South Africa, due to corruption and incompetence in policy implementation, but just at sticker price, this could get expensive.

For the ACP100, using the R75,000 ($4000) per kilowatt benchmark of China’s Hualong One, each 125 MWe unit might cost R9.4 billion, totaling R188 billion for 20 units. The HTMR-100’s costs are tougher to estimate. The R10 billion development goal hints at ambition, but per-unit estimates are unavailable. Broader SMR cost ranges, from R940 million for microreactors to R56.4 billion for bigger units, suggest the ACP100 might land between R9.4 billion and R18.8 billion, while the HTMR-100 could be cheaper due to local production.

SMRs often promise lower costs via modularity, but in reality, scaled implementation tends to accrue administrative costs and waste due to government inefficiency and patronage. Even in the United States, taking NuScale’s project in Utah: R174.84 billion for 462 MWe, or R377.8 million per MW - a 54% cost above initial estimate. Ultimately, it was canceled in November 2023 due to insufficient subscriber commitments and escalating costs, despite $4.2 billion in federal subsidies, including $1.4 billion from the U.S. Department of Energy and additional support from the Inflation Reduction Act.

For South Africa, this could get much more expensive. The Kusile Power Station is often cited as the most expensive coal-fired power plant in the world - its initial budgeted cost estimate in 2007 was just R81 billion, and was expected to take six years to build. Instead, it ended up costing R240 billion, and took 18 years.

For 2500 MWe, South Africa could be looking at R190-380 billion, depending on the reactor and scale, but since South Africa can run overbudget by a factor of three, that range should really be R570 billion - R1.1 trillion. Funding could lean on international loans, private investment, or Chinese support, but South Africa is already under significant fiscal strain, with estimates from economist Dawie Roodt showing that our debt-GDP ratio is ow at 95%, taking SOEs and municipalities into account.

China’s ACP100 demo promises a 58-month build, while local Stratek Global targets a pilot HTMR-100 by 2028. Multiply the 3-5 year estimates by three, and we’re looking at a 9-15 year waiting time for the very first module to come online. Scaling to 71 units, though, would take far longer. The Chinese partnership could speed things up, but this offers no guarantees, as China still has to deal with the same red tape and organised labour issues as other South African project managers.

This is not yet taking into account delays and raodblocks presented by the potential reaction from environmentalists.

‍