Nuclear realignment

Global Energy Report 2023
19 min read

Is nuclear energy safe, or unsafe? Too expensive to deploy, or too cheap to meter? Putting populations at risk, or saving Planet A? There is space for even another elephant in the nuclear room, in the era of reverse globalisation: is nuclear energy a useful geo-political lever for hegemons, or a shield in times of energy warfare for the meek? In this article, project developer/nuclear adviser and ex-banker David Stearns sheds light on the uncanny staying power of nuclear energy. But staying power is not sustainability, so it is necessary to review the legacy business model to assess who actually benefits from nuclear energy. A realignment of costs, risks and benefits is needed to ensure the industry’s as-yet untested financial sustainability.

It’s only April, and the Year of Sustainability is poised already to be a banner one for the nuclear industry. After years of delays and derision from outside the industry, several large-scale new builds are finally coming online – the US’s Vogtle 3, seven years late, Finland’s Olkiluoto 3, 13, and Slovakia’s Mochovce 3, 10. In 2024, the once-tut-tutted French EPR reactor is likely finally to be in service at Flamanville, 12 years late, the third country in three years to operationalise this incredibly complex and expensive mega-project. By then, it is possible that there will even be a £30bn or so financial close on the 7th and 8th EPR units, at Sizewell C in the UK. Has the nuclear industry managed to snatch victory from the jaws of defeat?

A nuclear moratorium and extended winter in the OECD was all but declared with the Fukushima-Daichi triple meltdown in 2011. But it seems now a distant memory. Or perhaps memories are short. Westinghouse’s 2017 bankruptcy, after two poorly negotiated and executed EPC contracts in the US? Ask Brookfield and Cameco. They are purchasing all of Westinghouse on the basis of a US$3.4bn valuation. The seller, Brookfield Business Partners, announced an exit IRR of 60% to its LPs. Areva’s insolvency, restructuring and sale to EdF in 2017, after being unable to deliver a turnkey contract on mostly fixed-price terms, resulting in a series of multiple billion euro losses? Ancient history.

“We’re witnessing some of the best market fundamentals we’ve ever seen in the nuclear energy sector,” said Cameco’s CEO when the Westinghouse deal was announced in December 2022. Brookfield recently advised its unit-holders to expect significant “liquidity events” as their broader investment portfolio enters a new “monetisation cycle”.

Sure enough, Westinghouse has come roaring back, signing preliminary agreements in the past 12 months to deliver large-scale projects in Poland and Ukraine. Over that same timeframe, EdF was reported to have bid its EPR in Poland, India and Saudi Arabia. It was also just handed a mandate from the French government to deploy its simplified EPR2 across the country in an ambitious 14-unit newbuild programme – yes fourteen. Elsewhere, South Korea’s KHNP is expected this year to complete the 4th and final APR1400 unit at Barakah in the UAE, a crowning achievement for an inaugural export project to a newcomer country, delivering 5,600MW within 12 years. EdF may surpass that capacity if it can achieve financial close on Sizewell C, bringing UK EPR capacity to 6,400MW. For its part, KHNP’s next move is expected to be in Poland, where it has signed preliminary agreements to deliver up to 2 APR1400 units.

After Team Japan’s eye-wateringly costly efforts and honourable exit at Moorside, 2017 and Wylfa, 2019, in the UK, the competitive nuclear export market was meant to be dead for good. Well, slight disclaimer, I was in the room and that was my dire prediction. Is the Barakah project the new rule, or the exception to the old rule?

Estrangement from investibility

A tiny bit of industry history is in order. For decades, the nuclear industry suffered in the eyes of public opinion as being unsafe, undeliverable and increasingly uneconomical. Major safety incidents in 1979, Three Mile Island, 1986, Chernobyl and 2011, Fukushima-Daichi, brought newbuild projects and operating assets around the world to a screeching stop as national regulators re-assessed safety baselines and, helpfully or not, created new national standards.

“An accident somewhere is an accident everywhere” became both the industry’s cautionary tale and its battle cry. Predictably, continuously rising safety standards resulted in ever-greater system redundancies (in nuclear culture, this is meant to be a very good thing indeed) and “defence in depth”. But with increased system complexity came tighter programme couplings.

Not so long ago, a newbuild director stood up in front of a British parliamentary select committee and boasted that his employer’s product required the very skilful delivery of over 55m unique engineering-related tasks by thousands of specialists across multiple countries and languages over a period of about five years. Financiers instantly recoiled, realising investibility was just as much about the big tail risks – nuclear accidents and political U-turns – as the accumulation of small errors in first-of-a-kind (FOAK) and first-in-a-while programme trajectories. No amount of financial or schedule contingency could ever be sufficient to permit programme remediation and recovery for even “business as usual” risks. Delivery costs could only move one way, and no one could control them. Overruns were a fact, not a risk, and the unanswerable question became, “by how much and for how long?"

But the financial community did not know much about the internals of the nuclear delivery and business models until very recently. Projects were historically government guaranteed and export finance loans were ECA-covered. Diligence had been very limited. But thanks to the extensive market soundings of the 2000s, finance was able to generate feedback, albeit not positive. The project list reads like a history of the industry’s Great Battles with Public Opinion, Politics and Mammon: TVO’s Olkiluoto 3, 2003, NEK’s Belene 1/2, 2005, EdF’s Flamanville 3, 2007, Enel’s Mochocve 1/2 refurbishment and 3/4 newbuild, 2008, the slew of US projects – Vogtle 3/4, VC Summer, South Texas, Calvert Cliffs, Victoria County – and many more. ENEC’s Barakah 1-4 project was also launched in those heady times.

Most of these projects started to stumble, not for technical reasons. The technologies typically were acceptable and only lightly diligenced. It was the business model, or perhaps lack of one, that failed. Government guarantees and shareholder balance sheets were exceedingly thin. Most projects needed external, third-party finance to close the funding gap. But external financiers were not convinced that high capital costs could be recovered through the competitive power market. Wholesale prices were volatile and dropping rapidly, and large blocks of the contestable market were getting colonised by priority-dispatch renewable energies.

External financiers asked for extensive government guarantees. PPA providers asked for completion-cost and performance certainty. Owners asked for greater fixed-price and date certainty from the contractors/vendors. Every project party asked for performance commitments and security from other parties that could not, or in a few cases would not, provide them. Project size was also a problem – the risk-underwriting and financing gaps were in the multiple dollar billions. A badly structured project could kill you. For Westinghouse, VC Summer and Vogtle, and Areva, Olkiluoto 3, they actually did.

It was in this financial environment that virtually every horizontally-integrated programme failed or was abandoned, while vertically-integrated programmes moved forward. The UAE’s original third party-led finance plan for Barakah was ultimately chucked after five years of iterations with the market, replaced with cheap and almost bottomless bilateral government-backed funding. Similarly, multiple projects led by Rosatom with integrated construction finance went ahead – in Belarus, Hungary, Egypt, Bangladesh and Turkey. Tellingly, Rosatom’s only projects that specifically required third-party finance for completion – Bulgaria’s Belene and Finland’s Fennovoima – were never successful in attracting outside financiers. Neither, for that matter, has Rosatom’s innovative PPA-backed BOO approach in Turkey, Akuyyu. Designed specifically to offer investment opportunities to investors and lenders, it has not succeeded in delivering external finance. With Unit 1 now approaching completion this year, that financial design, intended to attract third-party construction finance in order to limit the owner/contractor’s funding commitments, is on the verge of failing completely.

2023 – A fork in the road

Between 1999 and 2020, there were almost as many plant closures, 103, as grid connections, 104, around the world. Nuclear market share has been stagnating at some 10%. The OECD fleet has an average age of almost 35 years. The vertically-integrated programmes in China, India and Russia accounted for the lion’s share of starts. In this context, the growing climate, energy affordability and energy security poly-crisis has been an opportunity for OECD governments to reassess the un-monetised benefits of nuclear energy. In essence, with the support of public opinion, dozens of governments have decided they want to increase nuclear’s share in the energy mix, and have challenged the industry to come up with a proposal it can deliver. Nice work if you can get it.

In the UK, EdF asked for, and got, a new financing framework called regulated asset base (RAB) for Sizewell C. RAB features significant risk- and cost-share with consumers. In France, the government fully nationalised EdF, removed the nuclear market-share cap, and instructed EdF to define a programme it can actually deliver. EdF’s initial six-reactor price tag: €50bn. In countries without delivery capabilities or fiscal flexibility, such as Belgium, the Netherlands, Romania, and Slovenia, existing reactors are planned to undergo more economical life-extension works. The Czech government, having recognised that OECD vendors cannot mobilise large amounts of funding or balance sheet to absorb completion risks, is seeking to run a competitive tender featuring full host government-backed construction financing. The large-reactor industry has been invited to the government’s table, and dinner is being served.

It is not entirely clear yet if, or how, these governments are developing their requirements. On the one hand, these are bold top-down measures, though some may fall afoul of traditional EU state aid rules. On the other, at the bottom-up project level, it is not clear if the industry is being incentivised to improve its delivery model.

Of all the countries seeking to expand nuclear capacity, Poland seems to have the highest hurdles to clear. Or rather, the case for the economic delivery of large-scale nuclear will face its highest hurdles: a newcomer client, transitioning urgently from domestically sourced coal-fired power stations, with limited fiscal flexibility. At first glance, a perfect opportunity for the sole-sourced vendors to use massive information asymmetry to their advantage and replicate the somewhat predatory and short-sighted contracting practices of the past. If they are not required to do so, most vendors will not address owner’s gaps (scope, capabilities, funding)until the gaps have crystallised, generating change orders and plenty of revenue. The industry will not have advanced or innovated if Poland needs to lose (financially) for the vendors and contractors to win. Poland’s PEJ and PGE-led projects are a twin test case and likely leading indicator on if/how the industry can self-correct.


The SMR business model relies on economies of multiples, rather than the elusive economies of scale of the large-reactor mega-projects. Manufacturing smaller, standardised units in a factory environment can vastly improve “production runs”. Reducing interface risk and variances in quality reduces rework and compresses costs and schedules. The US DoE recently estimated that the industry must sustain a 12+% learning rate over some 20 units in order to reduce installation costs from the current US$9,000/kW to the commercial lift-off target of US$3,600/kW. With vastly greater potential client and site numbers than for the large reactors, SMRs can potentially achieve that scale of delivery better and faster than their large-reactor brethren.

2023 has been a banner year for SMRs as well. The US licensed its first SMR design, NuScale’s Gen 3 PWR, and the government committed critical development cost-share funding to projects led by both NuScale, Gen 3 PWR, and x-Energy, Gen 4 HTGR. In Canada, GE-H, Gen 3, and ARC, Gen 4, first-of-a-kind designs were competitively selected by customers backed by Canadian government funding. More recently, Terrestrial achieved Canada’s first level 2 design approval for its Gen 4 integral molten-salt reactor, paving the way for a design freeze for the purpose of commercial contracting and delivery.

In the EU and UK, independent SMR vendors are using their initial VC funding to progress technical definition and lobby for modest dollops of government grant funding. Their priorities are to take their designs through generic licensing approvals, and to implore host governments to facilitate the common infrastructure workstreams for site selection and approvals, supply chain mobilisation, operating procedures and decommissioning requirements.

These vendors are seeking more welcoming infrastructure-hosting regimes than in the the US or Canada, but may be surprised when they realise regulators and governments are sovereign, and, when it comes to nuclear, just as finicky. Furthermore, the European market is more fragmented from a licensing, supply chain and government support perspective. As a result, the first SMR generic design licence seems a few more years away in the UK/EU. On the other hand, latent market demand continues to strengthen as the UK/EU's pivot from fossil fuels becomes more pronounced, making the potential payoff for these vendors much greater.


So the nuclear industry continues to lumber ahead. If properly structured, the large-scale reactor programmes will deliver high learning rates for vendors and owners, albeit starting from a low baseline. And the SMR side is spending enough money to maintain the credibility and promise of their radical new business model. Despite the many challenges still ahead, the promises of commercialisation between 2030 and 2035 from about a dozen large-scale and SMR vendors seem credible enough at this stage.

The SMR business model will not be validated until those first units are fully designed and licensed, financed, procured, manufactured, installed, commissioned and operated. Does the sector have enough sweat equity (and OK, government grants) to eke it out until the 20th unit is delivered and perhaps investment-grade ratings are achieved for the asset class on the whole? VC funding becomes incrementally harder, as investors facing dilution become more demanding, competing designs vie for the spotlight and cash burn rates ramp up with the work multiplier effect. Not a problem, apparently. Bloomberg recently estimated that there is no trouble on the horizon, with over US$5bn of VC funding mobilised just last year, suggesting the momentum on innovation will keep going for at least several more years.

The SMR market seems to be saying that they do not need to speak much with Ministries of Finance, or even external finance, since their funding model is essentially fully equity-underwritten until their product is proven and operationally good enough to attract regular investors and lenders. Where that refuelling point lies on the timeline depends on each developer. Some are already engaging with future external financiers, taking on board their interest in things that governments can do, or cannot do, around setting rules for the sale or purchase of low-carbon power and heat, taxation and subsidies of various energy forms, procurement of nuclear technology and the transport of fissile materials, waste treatment and disposal, nuclear insurance and liability, and decommissioning. But for now not many VCs have been that exacting. The main difference with the previous nuclear renaissance is that this time, contrarian financial market feedback is not welcome.

On the helpful assumption that the VC tap will never go dry, project developers and financiers have little more to do than encourage the SMR vendors not to deposit any obscene cash balances with the next Silicon Valley Bank, and bid them a good day.

If the large-reactor vendors and delivery partners are properly engaged, their promises – construction schedules below five years, operating lives of up to 60 years, and load factors consistently above 90% – would be entirely credible. Across multiple units and fleets, as safety standards continue to ratchet upward, costs can be expected to plummet below US$4m/MW, delivering LCoEs in the US$40–60/MWh range.

Who can get in on the action? The large-reactor sector seems to be at capacity despite no new orders. With lead times up to 10 years, a foreign host who has not meaningfully engaged by now – through siting, grid and demand studies, licensing framework, technical pre-qualifications and nuclear infrastructure building – would be well-advised to wait and see how the current announced/development projects progress (UK, Poland, Czechia, Slovakia, Slovenia, Romania, France). Issuing an EPC/Investor RFP in 2030 is probably preferable to issuing one in 2023.

For hosts and offtakers seeking to engage with the SMR market, there are new learnings and much more to play for. There is a natural symmetrical relationship between hosts seeking the benefits of NOAK SMR units, and SMR developers raising cash on the strength of a robust NOAK pipeline. Nevertheless, just more funding or orders cannot, and will not, compress the SMR development timeline. NOAK units cannot credibly be ready at scale before 2030 at the earliest, or more likely 2035. More impatient or ambitious hosts, with demand potentially in the 20–100 unit range, can engage immediately, since SMR developers are keen to share risks – and future rewards – with their early clients.

The first-mover developers with the first committed orders will likely be the future nuclear industry leaders. The client side of the nuclear industry has never had a stronger hand to play, whether in large-scale or SMR nuclear projects.

To see the digital version of this report, please click here.

To purchase printed copies or a PDF of this report, please email