Your vaccination jab against ignorance.
20 March 2018 08:08 (South Africa)
South Africa

South Africa’s Electricity Choice (Part 1): Why Ramaphosa must kill the nuclear dream

  • Anton Eberhard & Amory Lovins
    Anton Eberhard & Amory Lovins
  • South Africa
Photo: A file picture dated 11 September 2006 shows the Units 5 and 6 of the Bulgarian nuclear power plant during a sunrise in the town of Kozloduy. EPA/VASSIL DONEV

Shifts ever more adverse to nuclear new-build have accelerated on every front. The global nuclear enterprise is slowly dying of an incurable attack of market forces. By ANTON EBERHARD and AMORY LOVINS.

Deputy President Cyril Ramaphosa has indicated his determination to improve the economy and attract more investment to South African shores in every one of his speeches since his election as ANC president in December 2017. But to succeed in his mission to put the country back on the economic track once he becomes president, he will need to act firmly and decisively to close the chapter on the Jacob Zuma administration’s obsession with nuclear power.


The South African government’s obsession over the past decade with nuclear power may finally come to an end with a new president and administration committed to restoring good governance and securing reliable electricity supply at least cost to support economic growth and developmentii,iii. The inherent complexity of nuclear procurement, financing, and especially construction, means no new nuclear generated electricity could flow for at least a decade even with immediate procurement, and then only at prices well above other electricity optionsiv. The risk that the project might collapse as unfinanceablev,vi, or through corruption, creates further uncertainty around South Africa’s electricity choices.

Some South African advocatesvii,viii,ix assert that nuclear costs will prove affordable for five reasons:

vendors are eager to sell;

vendor financing might come at low cost and long debt tenors;

higher prices or subsidies are worth paying for such a “strategic input” as nuclear electricity;

South Africa is “ideally positioned” to help satisfy the African continent’s “huge appetite for nuclear energy”; and

Western countries’ “severe over-regulation and a culture of being overly risk averse”, over-reacting to the ostensibly innocuous events at Chernobyl and Fukushima, caused “severe cost and schedule overruns” that South Africa, presumably with a more sensible risk perception, could avoid.

Whether capital markets would validate these assertions or view them as signs of disconnection from technical and market realities is a testable empirical question.

The country’s latest two flirtations with modern nuclear power have not gone well. First, plans for a home grown “pebble bed” modular reactor (PBMR)x were abandoned in 2010 after 12 years and $1.8-billionxi had been wasted – much of it after the project was acknowledged to be a non-starter (as Eskom presumably knew when its US utility partner pulled out in 2002). Now rarely mentioned, the PBMR found no customers or investors, but its advocates’ ambition and influence live on.xii

In 2008, perhaps slightly chastened by the unfolding PBMR episode, the government rejected as “uncompetitive and unaffordable” bids to build two light-water reactors totalling 3,2 gigawatts (GW or billion watts). The two undisclosed bids were reportedly more than twice government’s expectationxiii. Despite such prices, the rejected nuclear vendors, which didn’t include Russiaxiv, failed to thrive: Areva became insolvent and is now being bailed out for €5-billion by French taxpayersxv, while Westinghouse sold its nuclear business to the now junk-rated Toshiba, and is currently in bankruptcy proceedingsxvi.

Despite these setbacks, the South African government’s ambition mushroomed into a plan three times bigger than in 2008 – 9.6 GW of nuclear power – even though each GW will now cost more because real costs have risen by more than one-fourthxvii while the rand has lost half its value. With the national debt downgraded to junk-bond statusxviii, “affordability” can hardly have improved.

Electricity planning

Government’s nuclear policy consistently cites as its analytic basis the Integrated Resource Plan (IRP) for Electricity 2010 to 2030, which was updated twice in 2010 after public consultation, promulgated 25 March 2011, and formally published 6 May 2011xix. It called for 13% nuclear capacity and 21% solar and wind capacity in 2030. The IRP’s analysis found new nuclear power not economically optimal, even at real dollar costs far below today’s (though 40% above initial assumptions) and assumed to fall by a further 3%/y (they instead rose by roughly 5%/y).

The IRP nonetheless committed to 9.6 GW of nuclear power – as a matter of “policy” despite its inconsistency with the National Development Planxx. This “non-negotiable” nuclear policy has been enunciated consistently by President Zuma and his energy ministers from 2010xxi, even as its stated rationale has steadily evaporated. That rationale was chiefly “to account for the uncertainties associated with the costs of renewables and fuels”. As we’ll see, South African renewable energy costs were then discovered in the marketplace and proved to be far lower than empirical nuclear costs, with the spread between them continuing to widen; renewables, just like nuclear power, displace the same fuels whose uncertain prices were of concern; and renewables meet all the other stated policy objectives as well as or better than nuclear power.

A 2013 draft IRP updatexxii, though somewhat flawedxxiii, was posted on the Department of Energy (DoE) website in November 2013xxiv, soliciting public comment by February 2014 for consideration “in preparing a final draft IRP 2010 Update which (was to) be submitted to Cabinet for final approval by March 2014”. That updated draft was required, said the DoE’s website, because “there have been a number of developments in the energy sector” since 2010 and “the electricity demand outlook has changed markedly”.

The 2013 draft update strengthened the previous finding that nuclear new-build lacked economic rationale. It concluded that “in 2015, even if demand is high and there is no prospect of shale-gas power plant,…if nuclear costs exceed $6,500/kW, then the (nuclear) procurement programme should be abandoned.” Specifically, the draft found new nuclear plants clearly uneconomic if they got costlier, or renewables got cheaper, or demand growth slackened. In fact, all three strongly occurred. The plan indicated new nuclear to be unneeded for 15 to 25+ years – or ever if its capital costs rose another 12%. They have since risen much higher than this, with costs in some countries exceeding $6 500/kWxxv,xxvi while renewables cost less and (using the best U.S. data) performed one-third betterxxvii,xxviii than the 2013 draft update had assumed. Two years on, the gap between nuclear and renewable costs had widened furtherxxix,xxx.

However, the IRP 2013 was never finally approved and the parliamentary energy portfolio committee chair was reported at the time saying it “will never see the light of day”; indeed, it was never gazettedxxxi and can thus be officially ignored.

By 2016, many of the 2013 draft update’s data had become even more antiquated and a decision was made by the DoE to produce a further IRP update, which was gazetted on 25 November 2016 for public consultationxxxii. In the base (or reference) case, nuclear energy appears only from 2037, but after questions posed formally to the department by the Ministerial Advisory Council on Energy (MACE), it became clear that the modellers had placed artificial constraints on the amount of solar or wind energy that could be built in any one year. Once these constraints were removed, nuclear energy did not appear in any of the optimal least-cost scenariosxxxiii.

The government ignored rational warnings and requested Eskom to expedite the procurement of nuclear power. However, it soon started butting up against increasingly vocal challenges from various quarters. In 2016, civil society organisations launched a court challenge to government’s continued preference for nuclear power. On 26 April 2017, the Western Cape High Court not only set aside a number of international nuclear co-operation agreements, including that with Russia, but also invalidated the ministerial determination (made in terms of section 34 of the Electricity Regulation Act) to procure nuclear electricity. The court found that the National Energy Regulator of South Africa (NERSA) had concurred with this ministerial determination (as required in law) but had done so without adequate public consultation. In effect, the ruling forced government to restart the procurement process from scratch, including producing a new update of the IRP and a new official ministerial procurement determination, but only after the concurrence of NERSA is obtained, and this only after adequate public consultationxxxiv.

During the last quarter of 2017, the DoE asked the IRP modellers (based in Eskom) to produce another updated IRP to reflect lower actual electricity demand and the latest comparative energy prices. The DoE may have noticed the global publicity about plummeting renewable energy prices, now so low in many good sites, without subsidies, that they rival or undercut just the operating costs of many existing nuclear or fossil-fuelled power stations, let alone the total cost of power from new onesxxxv,xxxvi.

The resulting new electricity demand forecasts of the modellers are almost certainly still too high, as are their assumptions of renewable energy costs. In a number of emerging economies in 2017, we have seen solar and wind auctions deliver prices that are half those registered in the last renewable energy auction in South Africa held in November 2015, after which Eskom had stalled further bids by refusing to sign PPAs xxxvii,xxxviii,xxxix. And the DoE’s nuclear cost assumptions seem much lower than many recent contracts, such as the Rosatom deal for up to $21bn in Egyptxl.

But even with these conservative assumptions, nuclear energy is not picked in any of these new South African modelling scenarios, other than one where artificial constraints are placed on how much solar and wind energy can be built and where additional carbon budget limits are imposed. Even in this extreme scenario, nuclear energy would only be required after 2039.

These outcomes are consistent with the IRP update undertaken by the DoE in 2016 under the direction of the Ministerial Advisory Council on Energy and with the Council for Scientific and Industrial Research’s (CSIRs) model which used the same PLEXOS software and assumptions as the DoE in 2016.xli A few years back, the National Planning Commission asked South Africa’s premier university-based energy modelling group – the University of Cape Town’s Energy Research Centre – to run an independent model. They too came to the same conclusionxlii.

These outcomes were not welcomed by the nuclear section in the DoE and the modellers were asked to run a scenario where 9,6 GW of nuclear power stations are “hard-wired” or forced into the model. Fortunately, the modellers also calculated the additional wholesale cost of electricity in this scenario – R800-billion over 20 years at net present valuexliii.

The optimal, least-cost mix recommended by all these models includes solar and wind energy complemented by pumped storage, hydro-electricity and gas turbines or engines, on top of legacy coal fired powered stations.

South Africa’s Energy Minister decided to ignore the conclusions of his modelling team and presented a revised IRP2017 to Cabinet on 6 December 2017, which still included the procurement of 9,6 GW of nuclear power but over a longer period to reflect lower electricity demand.

South Africa’s Constitution requires administrative action to be lawful, reasonable, and procedurally fair. Almost certainly, civil society organisations will go back to court to challenge the rationality of the latest IRP and any actual procurements that might follow.

The IRP should be a living planning document that is regularly revised and updated as necessitated by changing circumstances. At the very least, it should be revised by the DoE every two years and should provide a rational and informed base for power procurement and investment decisions. The way in which DOE has continued to manipulate and adjust least-cost, optimal plans to incorporate nuclear power reduces stakeholder confidence that Cabinet is fully informed about the very fast-moving competitive landscape in electricityxliv. Just within one year (2016), levelised world costs for new onshore wind power fell by 18%, for offshore wind by 28%, and for utility-scale photovoltaics (PV) by 17%xlv, while low bids fell 37% for Mexican PV and 43% for EU offshore windxlvi. Yet South Africa’s Cabinet is still apparently relying on data using 2010 cost data if not earlier – myopia that would not pass muster in any private-sector boardroom.

Shifts ever more adverse to nuclear new-build have accelerated on every front. The global nuclear enterprise is slowly dying of an incurable attack of market forcesxlvii. Financial distress stalks vendors, with cascading insolvencies spreading in the past two years. Construction cost and delays keep rising worldwidexlviii. As a Frost & Sullivan analysis found in 2014xlix, whatever is driving South Africa’s nuclear push, it’s not economics. And as financial consultant Dirk de Vos concludedl, “The nuclear industry can only exist when the state underwrites its cost. Countries that procure their power through democratic, transparent and market-based methods are not building new reactors.”

All this raises several questions: what could emerge from the “full, transparent and thorough”li cost-benefit analysis previously requested by the ANC national general council in 2015lii,liii, and from the Constitutionally required transparent competition demanded by the National Union of Mineworkersliv? What salient evidence was absent when Cabinet formally launched the nuclear procurement programme based on the gravely outdated 2010–11 analysis? What major decision factors have changed since the failed 2008 nuclear plan? We explore these questions in the context of the two most dynamic competitors – renewable generation and efficient use – then discuss policy risks. DM

In Part 2, we explore renewable energy.

Physicist Amory Lovins is a renowned, leading world energy expert. A former Oxford don, honorary US architect, Swedish engineering academician, adviser to business and government leaders for 44 years in over 65 countries including South Africa, he has won many of the world’s top energy and environment awards, received 12 honorary doctorates, taught at 10 universities, and written 31 books and over 625 papers. Time named him one of the world’s 100 most influential people; Foreign Policy, one of the 100 top global thinkers

Anton Eberhard was inspired to undertake his PhD in the field of energy and development in 1979 after reading Amory Lovins’ seminal publications on renewable energy and energy efficiency. He has recently been elected to the rank of Emeritus Professor and Senior Scholar at the University of Cape Town Graduate School of Business after 35 years of research, teaching and policy advocacy in energy and sustainable development in Africa.

Photo: A file picture dated 11 September 2006 shows the Units 5 and 6 of the Bulgarian nuclear power plant during a sunrise in the town of Kozloduy. EPA/VASSIL DONEV


i Ref. 25.
ii Phakathi, B. Zuma vows nuclear option is a given for South Africa. Business Day LIVE. 3 November 2017.
iii SA to proceed with nuclear power expansion: Mahlobo. ENCA. 7 December 2017.
iv Eberhard, A. SA is switched on to Mahlobo’s wheeling and dealing in nuclear bid. Business Day LIVE. 30 November 2017.
v De Vos, D. Can any South African nuclear energy procurement ever succeed? Daily Maverick. 5 May 2017.
vi Cohen, M. Opposition Grows to South Africa’s $100 Billion Nuclear Plan. Bloomberg Business. 27 August 2015.
vii Cilliers, A. Cost of nuclear versus other sources. 21 September 2017.
viii Slabbert, A. Nuclear: ‘It’s a buyer’s market’. The Citizen. 25 January 2016.
ix Slabbert, A. SA nuclear affordable if IPP model employed – South Korea. Moneyweb. 17 March 2016.
x Thomas, S. The Pebble Bed Modular Reactor: An obituary. Energy Policy. 26 January 2011.
xi Thomas, S. The demise of the pebble-bed nuclear reactor. Bulletin of Atomic Scientists. 22 June 2009.
xii Yelland, C & Potgieter, E. Eskom renewed interest in the PBMR. 29 March 2017.
xiii Eberhard, A. Towards a secure, competitively priced and environmentally sustainable electricity future. Keynote address: South Africa’s Electricity Supply Conference. 15 August 2013.
xiv South Africa regards Rosatom as most likely contractor to build nuclear power plant. TASS. 11 June 2014.
xv EDF approves Areva reactor purchase, clearing path to bailout – sources. Reuters. 15 November 2016.
xvi Hals. T. Westinghouse reaches deal for $800 million U.S. bankruptcy loan. Reuters. 23 May 2017.
xvii Thomas, S. Nuclear Technology Options for South Africa. University of Greenwich. July 2014.
xviii Donnely, L. Global credit ratings agency has downgraded South Africa to junk status. Mail & Guardian. 25 November 2017.
xix Electricity Regulation Act No. 4 of 2006 and Electricity Regulations on the Integrated Resource Plan 2010 – 2030. Government Gazette. 6 May 2011.
xx Marrian, N. ANC allies alarmed by ‘big rush’ nuclear deal. Business Day LIVE. 25 September 2014.
xxi Ref. 1,2.
xxii Posted, with additional EPRI input data used, at (accessed 14 February 2016).
xxiii Comments on the draft Integrated Energy Plan issued by the Department of Energy in 2013. Energy Research Centre, University of Cape Town. 13 December 2013.
xxiv Department of Energy. The draft update is linked at
xxv Ref. 16.
xxvi Schneider, M. et al. World Nuclear Industry Status Report 2017. 12 September 2017.
xxvii Bolinger, M. & Seel, J. Utility-Scale Solar 2014: An empirical analysis of project cost, performance, and pricing trends in the United States. Lawrence Berkeley National Laboratory. September 2015.
xxviii Wiser, R. & Bolinger, M. 2014 Wind Technologies Market Report. LBNL-188167. Lawrence Berkeley National Laboratory. August 2015.
xxix Bolinger, M., Seel, J., Hamachi LaCommara, K. Utility-Scale Solar 2016, LBNL-2001055. Lawrence Berkeley National Laboratory. September 2017.
xxx Wiser, R.H., & Bolinger, M. 2016 Wind Technologies Market Report. LBNL-2001042. Lawrence Berkeley National Laboratory, 2017.
xxxi Martin, B. & Fig. D. Final Report—Findings of the African Nuclear Study, p 17. Heinrich Böll Stiftung Southern Africa. Updated 27 July 2015.
xxxii Government Gazette, Republic of South Africa. Vol 617. 25 November 2016. No. 40445
xxxiii Eberhard. A. Expect legal challenge if Eskom sticks with its irrational procurement course. Business Day LIVE. 25 October 2016.
xxxiv Staff reporter. Read the full judgment: High court finds nuclear deal unconstitutional. Mail & Guardian. 26 April 2017.
xxxvi IRENA. Renewable energy auctions: analyzing 2016. IRENA. 2017.
xxxvii Dipaola, A. Saudi Arabia gets cheapest bids for solar power in auction. Renewable Energy World. 3 October 2017.
xxxviii Deign, J. The world’s cheapest solar bid is actually for wind. Greentech Media. 26 November 2017.
xxxix Safi, M. Indian solar power prices hit record low, undercutting fossil fuels. The Guardian. 10 May 2017.
xl Reuters. Rosatom says planned Egypt nuclear plant to cost $21 bn. Engineering News. 11 December 2017.
xli Wright, J.G., Bischof-Niemz, T., Calitz, J., Mushwana, C., Van Heerden, R. & Senatla, M. Formal comments on the Integrated Resource Plan (IRP) update assumptions, base case and observations 2016. CSIR. 3 April 2017.
xlii Energy Research Centre. Towards a New Power Plan. University of Cape Town. April 2013.
xliii Ref 34
xliv Ref. 3.
xlv McCrone, A., Moslener, U., d’Estais, F., & Gruning, C. Global Trends in Renewable Energy Investment 2017. Frankfurt School-UNEP Centre/BNEF. 6 April 2017.
xlvi T. Kåberger, Opening address, REvision 2017 symposium, Renewable Energy Institute (Tokyo), 8 March 2017.
xlvii Ref. 25.
xlviii Ref. 25.
xlix Harris, T. Clarity required on drivers of South Africa’s nuclear policy. Engineering News. 6 August 2014.
l De Vos, D. Should SA put its nuclear dreams aside for now? Mail & Guardian. 12 April 2013.
li Wild, F. South Africa’s ANC Voices Caution on Nuclear Power Plan. BloombergBusiness. 18 August 2015.
lii Ref. 5.
liii Le Cordeur, M. ANC warned: Abandon nuclear plans, or else. Mail & Guardian. 6 October 2015.
livRef. 19.

  • Anton Eberhard & Amory Lovins
    Anton Eberhard & Amory Lovins
  • South Africa

Get overnight news and latest Daily Maverick articles

Reader notice: Our comments service provider, Civil Comments, has stopped operating and will terminate services on 20th Dec 2017. As a result, we will be searching for another platform for our readers. We aim to have this done with the launch of our new site in early 2018 and apologise for the inconvenience.