This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.


Is the EPR Nuclear Reactor Fit for the Current Market?

Quentin Philippe's picture
Actis (Energy Fund)

Quentin works for Actis, a London-based Private Equity fund focusing on Emerging Markets. At Actis, Quentin is working for their Energy Fund, which invests in power generation and distribution...

  • Member since 2018
  • 2 items added with 4,507 views
  • May 23, 2016

Your access to Member Features is limited.


In this article, we would like to shed some light on some of the reasons why the EPR[1] reactors have suffered construction delays and are not encountering the commercial success that was expected. It should be noted that the EPR reactor is still new, and that like all new and large construction project, “infant diseases” are to be expected, and that the reactor’s long term commercial viability should not be written off. On top of this, we would also like to point in this article however that the reactor design’s philosophy was not conducive of short construction timings.

To develop these topics, we will remind the reader in a first section of the latest difficulties that the EPR is facing. We will also in a second section highlight the impact of the EPR’s design philosophy on its marketability, construction timing and costs. In the third paragraph, we highlight a few reasons why the issues faced are actually not relevant for the reactor’s long term commercial success. In the last paragraph, we conclude and offer a few suggestions.

1. Context

The construction of the four EPR reactors currently under construction has been plagued with delays and cost-overrun[2]. For instance, the Finnish Olkiluoto 3 reactor – to be commissioned in 2018 – is 10 years behind schedule and €5bn over budget[3]. Regarding Flamanville, as “Le Monde” puts it, there is ground for discontent with 6 years delay versus the original planning (implying a doubling of the construction time), a trebling of the budget[4], on top of the many “scandals” regarding the working conditions for instance[5],[6],[7] . These issues have not affected to the same extent the two reactors being built in Taishan[8], China, where Taishan 1 is likely to become the first EPR to be commissioned[9]. Nevertheless, these reactors have suffered delays as well, due to safety concerns, with the first reactor initially planned to be completed in 2013[10].

In addition to that, the economic performance of Areva – the EPR’s designer – has been dismal, which led to the sale of its reactor business to French State-owned utility EDF. This is further straining the finances of EDF. In turn, this raised concerns regarding financial viability of UK’s first EPR in Hinkley point, to the extent that Thomas Piquemal – EDF’s former CFO – resigned from the firm.

Let us also remind ourselves that EDF and Areva lost a major contract in 2009 for the EPR in Abu Dhabi. The tender was won by a South Korean JV led by Kepco thanks to a more competitive bid despite a less advanced technology[11]

2. EPR’s design philosophy, and consequences on its marketability

The EPR’s design – which started in 1992[12] – is underpinned by two objectives[13]: (1) improve safety via redundancy mechanisms, diversity and complementarity of safety systems[14], (2) reduce the cost of electricity production through economies of scale:

1. On safety: We would like to give an example of the increased safety through redundancy systems[15]: the EPR has four separate cooling systems, or trains. Each of these trains is located in a separate building and is capable of cooling the reactor on their own. Because of this redundancy approach to increase safety, the EPR reactor requires twice as much concrete and electric wires, and four times more steel, than previous nuclear plants built in France[16]. This differs from the strategy adopted by Areva’s competitors, such as Westinghouse, who opted for the use of more passive systems and for a simplification of the overall design[17]. Compared to the previous reactor – the N4 in France, the EPR’s footprint and size have not increased in line with the additional complexity. The increase in complexity (cfr. increase in electric wires or concrete) leads to an augmentation of the density of equipment in the buildings, making the EPR more difficult to build.

2. On costs: The EPR aims at reducing the cost of nuclear power per unit of production (by ~10%[18] compared to the previous French PWR[19] reactor N4) through economies of scale[20]. Therefore, a high power design (~1600 MWe) was selected. The EPR aims at reducing O&M costs (-20% vs other Gen3/3+ reactors) and fuel costs (up to -15% vs other Gen3/3+ reactors)[21]. Integrating such a high power unit of 1.6 GWe in the power system is more difficult than a smaller unit. The installation of an EPR necessitates a well advanced power transportation system (e.g. 450 kV power transportation network). In other words, installing an EPR in countries with weak Transportation and Distribution networks require additional investments to re-inforce them, thereby increasing the total cost of ownership of the EPR[22].

The complexity to build the reactor leads to (a) long construction times (and delays) and (b) high construction costs. On the construction time, the delays that are witnessed for the 4 reactors currently under construction speak for themselves.

Regarding the construction costs, Leveque mentions: “The cost escalation is mainly due to the scaling‐up strategy. The scaling‐up is associated with greater lead‐times and complexity which in turn meant an increase in costs per MW. The construction of Generation III reactors confirms that larger reactors are likely to be more expensive again”[23]. Some have argued that the high strike price for the Hinkley point reactor in the UK, i.e. £92.5/MWh, may indicate how costly the reactor has become, certainly when it is compared to the tariff for the regulated access to nuclear power in France (€42/MWh)[24]. We would like to highlight that the comparison between these two numbers is tricky for several reasons[25], and the reader must remain cautious. Nevertheless, it would be difficult to argue that the construction cost of the EPR is lower than that of the previous generation of reactors.

In other words, relatively long construction times and high construction costs do undermine the attractiveness of the EPR and certainly will not boost its commercial success.

3. Reasons not to be alarmed by the EPR’s performance so far

Not all is gloom in the EPR’s performance and many of its ills are not related to its specific design. We would like to highlight the following elements:

a. Overly generous contractual terms: The terms of contracts for the Finnish reactor were deliberately optimistic in order to win the bid in a competitive market: “When construction on Finland’s Olkiluoto 3 began in 2005, French nuclear company Areva had promised to be finished by summer 2009 – a record time for a prototype nuclear reactor. Rare is the nuclear commission that doesn’t fall at least somewhat behind schedule[26].

At the time when the contracts were signed, before Fukushima, the competition was fierce to win new markets. This is critical to understand why the terms were so generous in the civil nuclear industry which was forecast to grow considerably[27].

In other words, Areva had been overly generous for the first EPR for strategic reasons and different market conditions. This can certainly be avoided in the near future.

b. More stringent regulatory framework: Regulation has become much more stringent in the last 15 years, leading to miscalculations in the estimated construction time which was based on the construction of the reactors under the previous regulatory regime.

On top of this, each national regulator had different requirements for the EPR, meaning that the learnings that could be applied to one construction site to another are limited. For instance, given the specific requirements of the UK regulator, the learnings from the French or Finnish reactors are not to be applied directly to the UK. In addition, new requirements from the UK regulator – such as the existence of a non-computerized safety system[28] – imply that additional costs must be undertaken. The I&C (instrument and control system) – which was fully automatized and was the reason behind much delay due to its innovative functionality – has been specifically re-designed for the UK. The UK regulator accepted the new design in December 2012. The re-design required a substantial amount of time and delayed the accreditation of the EPR in the UK. Going forward, the construction of additional EPR reactors in the UK should not face these delays. One could argue that the construction of an EPR under a new regulatory framework is akin to the construction of a first-of-a-kind (FOAK) reactor.

c. Some lost contracts / opportunities were not due to EPR’s design: A lack of organisation of the French nuclear industry, and tensions between EDF and Areva, were key factors leading to the loss of the mega-contract in Abu Dhabi in 2009[29].

EDF’s unfortunate foray into the US market came to an end with its exit from the JV CENG (Constellation Energy Nuclear Group) with Exelon due to the poor economic climate in the US for nuclear energy[30]. Indeed, with an abundance of gas from shale to power CCGT and very low power prices, investors do not have the incentives to start nuclear projects in the US. In other words, one may argue that EDF’s withdrawal from the US is then rather linked to the US power sector attractiveness, not the EPR’s design[31].

d. Delays similar to those of the EPR are common for similarly large and successful projects: The delay that the EPR is facing is akin to those that the N4 (Gen 2 reactor) was facing[32]. Despite these delays, the nuclear power price in France has remained very low[33] and competitive[34], implying that delays at the construction of FOAK reactors do not have to lead to a demise of new nuclear EPR.

The Westinghouse AP1000 – the most direct competitor whose design is very different and much less complex[35] – also witnesses delays and cost overruns. At the time of writing these lines, there are four AP1000 being built in the USA. Two at VC Summer and two at Vogtle. For VC Summer, a delay of at least one year and extra costs of $1.2 billion were announced in October 2014, largely due to fabrication delays. Regarding Vogtle, the construction is also delayed (e.g. in June 2013, the construction schedule of unit 3 had slipped by 14 months). In addition, as Leveque mentions regarding the AP1000: “The overnight costs registered in the applications submitted to the Nuclear Regulatory Commission lie between USD2010 3,650/kW and USD2010 5,100/kW. This represents at least a 75% increase with respect to the USD2010 2,400/kW estimated in 2003. More recently, Rosner and Goldberg from the University of Chicago also updated their previous study and estimated an average cost for the AP1000 equal to USD2010 4,210/kW.

e. Purchase of the AP1000 technology – Westinghouse’s Gen 3 reactor – by China may indicate why fewer EPR are considered in China: The lack of recent commercial advances in China – the biggest market at the moment for nuclear new build – may simply be due to the fact that the Chinese companies have purchased the Gen 3 technology from Westinghouse. Given the upfront cost to Chinese companies for the purchase of this technology, the production cost of future AP1000 units in China is therefore lower than the cost that can be expected from EPRs[36]. In other words, should Areva and EDF decide to transfer (and be able to sell) more know-how to China, it could very well gain the upper hand in that market.

f. A supplying industry to the nuclear construction sector needs to be restarted: In France, the construction of the last reactor to be commissioned started in the 1991 (Civaux-2 reactor)[37]. The suppliers to the nuclear industry have therefore been without new orders for a long period, i.e. about two decades, and this industry needs to be restarted. In China, where the demand for reactors has been strong for the last decade, the capabilities of the supplying industry are sufficient to meet the demand of the constructors of nuclear reactors and operators. Construction delays are therefore also due to the lack of capability of the supplying industry, which is unrelated to the complexity of EPR’s design.

g. Abu Dhabi’s loss is not reflective of EPR’s competitiveness post-Fukushima: The loss of the contract in Abu Dhabi may be reflective of the fact that the EPR is over-designed for the then requirements of Abu Dhabi but is actually well-suited in a post-Fukushima world.

The Korean design was based on a Gen2+ reactor which is intrinsically less safe than the EPR. Following the Fukushima accident, Kepco has been asked to review its design to ensure increased safety[38]. In other words, if Abu Dhabi’s tender were to be re-open today, existing EPR’s design would stand a better chance of winning than in 2009. One could say that EPR’s design was “over-delivering” in terms of safety for 2009 Abu Dhabi.

 4. Conclusion

It is difficult not to acknowledge that the EPR is a complex and costly reactor. However, it is also clear that all its ills cannot be linked to its design. There is little doubt that in countries such as France or the UK, follow-on EPR reactors will be faster to build than the first-of-a-kind reactors currently being built or in the planning (Flamanville in France, and Hinkley Point in the UK).

Nevertheless, Areva should ensure that going forward, the reactor’s construction and costs are reduced and better managed. In order to do so, standardization could be increased by reducing the technological choices and aligning regulatory frameworks. Indeed, it is striking that three European regulators, i.e. the Finnish (STUK), French (ASN) and English ones, have different design requirements. The adaptation of the reactor to each country’s needs increased construction costs and reduced the learnings among the different countries. In a sense, three first-of-a-kind EPR will be built, one for each country. An alignment at European level of regulatory requirements would be a first step to reduce construction costs in Europe and improve credibility outside Europe.

Commercially, Areva, in order to secure a bigger slice of the Chinese market, may consider transferring more know-how to China. This choice is however not straightforward. By transferring its know-how to China, Areva would stand more chances to secure additional contracts at a low cost, given that the know-how is a sunk cost. However, it would help creating a new Chinese competitor, which in the long-run may turn out to be detrimental for Areva.

Photo Credit: Bjoern Schwarz via Flickr


[1] “European Pressurized Reactor” or “Evolutionary Pressurized Reactor”

[2] Reuters, ” Finnish nuclear plant delayed again as Areva, TVO bicker”, 28 February 2014,  ; Bloomberg, “France’s EPR Nuclear Reactor to Get Chinese Debut, Minister Says”,  12 November 2013,

[3] Financial Times, “Tale of woe in French nuclear sector”, 13 October 2015

[4] The FT mentions a cost overrun of €7bn (see ref. in footnote 7)

[5] Financial Times, “EDF in fresh delay for flagship nuclear plant”, 18 November 2014,

[6] Le Monde, “L’EPR est-il vraiment un fiasco industriel”, 26 November 2014. « According to the French trade union CGT, many of the 1000 foreign workers employed by Bouygues Construction at FA3 work between 10 and 15 hours per day, earning just €250 per month.” ; “. The deadly fall of a worker at the inner containment wall of the reactor building in January delayed work for nine weeks and is under investigation by authorities for possible manslaughter. Another worker plunged to his death in June, while a night-shift worker was killed recently in a car crash on his way home.” /

[7] Financial Times, “EDF’s real problem is Flamanville not Hinkley Point”, 14 May 2016,

[8] In October 2015, the FT highlights the greater expertise in large civil projects in China [than in Europe],

[9] World Nuclear News, “First Taishan EPR completes cold test”, 1 February 2016,

[10] Reuters, “China’s CGN says delayed Taishan EPR reactor still on track”, 23 February 2016,

[11] Nuclear Engineering, KEPCO wins UAE civil nuclear bid”, January 2010, ;  Michel Berthélemy

and François Lévêque , “Korea nuclear exports: Why did the Koreans win the UAE tender?

Will Korea achieve its goal of exporting 80 nuclear reactors by 2030? “, April 2011, ; Pierre Boucheny (analyst at Kepler Equities), “Il est possible que les Français aient à revoir le positionnement de l’EPR. En effet, il présente un prix d’entrée très élevé qui ne lui permet plus de capter l’ensemble du marché “,


[13], p.2


[15] Illustrations of implementation of other safety principles : (i) diversity : use of diversified emergency diesel generators ; (ii) use of complementary passive and active systems: core catcher and containment sprays

[16] Millicent Media,

[17] The AP1000 for instance has only two trains. Source:

[18] See

[19] Pressurized Water Reactor


[21], p18

[22] Expert interviews

[23] Source: Leveque, « Revisiting the Cost Escalation Curse of Nuclear Power. New Lessons from the French Experience”, Toulouse 2013,


[25] To name a few: the strike price in the UK includes a return on investment for the investors (the existing French reactors have been party amortised); the financing conditions are very much different between the Hinkley point nuclear reactors and the existing reactor fleet in France; the construction of the Hinkley point reactor includes also costs for the development of South West infrastructure (e.g. roads, jetties). Source: expert interviews


[27] Le Monde, “L’EPR est-il vraiment un fiasco industriel”, 26 November 2014

[28] Source: Expert interview


[30] ; FT, “EDF to exit US nuclear power over impact of shale gas”, 30 July 2013,

[31] Note that the reactor has not been licensed for the US yet

[32] Le Monde, “L’EPR est-il vraiment un fiasco industriel”, 26 November 2014.



[35] Safety systems rely more on natural phenomena (passive safety) instead of redundancy as in the EPR

[36] Source: expert interview

[37] Wikipedia

[38] Expert interview

Quentin Philippe's picture
Thank Quentin for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member
Spell checking: Press the CTRL or COMMAND key then click on the underlined misspelled word.
Nathan Wilson's picture
Nathan Wilson on May 23, 2016

It is hard to see how the EPR will every see much commercial success, beyond a couple more for the UK’s Hinkeley plant. A huge factor is that Areva doesn’t have the cash to help its customers finance construction. Given that France’s PWR fleet can likely continue to receive life extensions rather than near term replacements, the EPR will have trouble building the momentum that Chinese reactors have.

With Russia struggling with low oil prices and Japan cowering in nuclear fear, China could carry away the international market. The first units for China’s 1150 MW Hualong 1 design started construction in 2015, and the first CAP1400 (a 1400 MW unit derived from Westinghouse’s AP1000) should break ground this year.

Jarmo Mikkonen's picture
Jarmo Mikkonen on May 24, 2016

i.e. the Swedish (STUK),

STUK is the Finnish regulator.

I have no doubt EPRs will run well, once the construction is done. EPR might be suitable for markets after 2020 once the current low oil and gas prices are history and coal is phased out. 60 years worth of carbon-free electricity.

Engineer- Poet's picture
Engineer- Poet on May 24, 2016

I wonder if the double containment isn’t an implicit argument for allowing the unit to run until the RPV or some other major component fails?  The safety argument for retiring a unit on the basis of embrittlement becomes pretty weak, and IIUC the failure scenarios generally assume a thermal transient like a sudden cold-water dump like you’d have in a SCRAM scenario anyway (so the reactor would already be off with only afterheat being a factor).

It’s an awfully expensive way to get to that, though.  I’d rather design for the insertion of thermal blankets and heaters to allow annealing of neutron damage every couple of decades.

Quentin Philippe's picture
Quentin Philippe on May 24, 2016

Thanks for noting this.

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »