(1)

By letter dated 21 June 2024, Belgium notified the Commission of the measure to support the lifetime extension of two nuclear reactors in Belgium (Doel 4 (‘D4’) and Tihange 3 (‘T3’), together the long-term operation (‘LTO’) Units. It provided the Commission with further information by letters dated 9 and 10 July 2024, following the Commission’s request for information of 4 July 2024.

(2)

By letter dated 22 July 2024, the Commission informed Belgium that it had decided to initiate the procedure laid down in Article 108(2) of the Treaty on the Functioning of the European Union (‘TFEU’) in respect of the measure (the ‘Opening Decision’).

(3)

Belgium sent its comments on the Opening Decision on 22 August 2024.

(4)

The Commission decision to initiate the procedure was published in the Official Journal of the European Union (2). The Commission called on interested parties to submit their comments.

(5)

The Commission received comments from interested parties. It forwarded them to Belgium, which was given the opportunity to react; its comments were received by letter dated 30 October 2024.

(6)

Further information by Belgium was provided on 2 September 2024, 23, 27 and 30 October 2024, 8, 14, 18, 22, 25, 27 and 30 November 2024, 2, 4, 5, 12, 13, 17 and 20 December 2024, 7, 9, 17, 18, 20, 27, 29, 30 and 31 January 2025, 1 February 2025 (3).

(7)

By letter dated 24 October 2024, Belgium agreed to exceptionally waive its rights deriving from Article 342 TFEU in conjunction with Article 3 of Regulation 1/1958 (4) and to have this Decision adopted and notified in English.

(8)

Until 2022, Belgium’s nuclear fleet consisted of seven nuclear reactors, four located in Flanders (Doel) and three located in Wallonia (Tihange). All reactors came into operation between 1975 and 1985 (5) and were built by public utilities (Ebes, Intercom and Unerg) that were eventually merged to become Electrabel SA (majority owned by Tractebel) in 1990 (hereafter referred to as Electrabel). In 1996, the Société Générale de Belgique (‘SGB’) became the majority shareholder of Tractebel and in 1999, Suez acquired nearly 100 % of SGB. Following the merger between Suez and Gaz de France (‘GDF’) in 2008, the ultimate owner of Electrabel has been Engie S.A. (hereafter referred to as Engie).

(9)

Electrabel, a fully-owned subsidiary of Engie, has been the nuclear operator and majority owner of the seven Belgian nuclear reactors since the start of their operations. Today, the ownership of the Belgian nuclear reactors is as follows:

(10)

As for the LTO Units capacities (see Table 2),

(11)

Belgium explains that all Belgian nuclear reactors, including the LTO Units, have been designed and operated as baseload units based on the technology of pressurised water reactors (‘PWR’). The design dates from the 1970s and is typical for baseload units, like older nuclear reactors around the world. These units are commonly operated at maximum rated capacity whenever online, contrary to newer technologies (Generation III) which are capable of flexible operation, including changing power output over time (‘ramping’ or ‘load following’) and providing frequency regulation and operating reserves (11). In baseload operation, power is usually only reduced or shut down when required for planned refuelling and/or periodic maintenance, unplanned urgent maintenance to correct plant equipment issues, or unexpected design and/or safety constraints. Hence, reduction in power output is rather driven by the needs of the nuclear operator or of the transmission system operator (‘TSO’), than by market signals.

(12)

Belgium clarifies that the Belgian nuclear reactors are technically constrained by their design, which strongly restricts the possibilities of output modulation, which was for a long time limited to grid regulation. This is due to the fact that Belgian nuclear reactors are exclusively equipped with black reactor control rods, which strongly restrict the ability to modulate, contrary to grey ones (12). By contrast, for example, some of the latest French nuclear reactors are equipped with both black and grey control rods, so that they are designed and built to perform modulations and load following. As a result of the design based on black control rods in the Belgian nuclear reactors, until recently, modulation for economic reasons was forbidden by the Belgian Nuclear Safety Agency (‘AFCN/FANC’) (13). Modulation was only acceptable for technical reasons or upon request from the Belgian TSO (Elia), to avoid a blackout (14).

(13)

However, Belgium submits that in response to the general evolution of electricity markets and the need to introduce flexibility into its operation, studies were performed concerning allowing some instances of modulation for economic reasons (‘economic modulation’). In 2015, Engie, Electrabel and the Belgian regulators launched safety studies (15) to analyse which conditions would have to be met to allow some modulations for economic reasons, by looking among other things at the impact of such modulation on the nuclear fuel. The study concluded that a maximum of 30 economic modulations per fuel cycle could be acceptable for Tihange 3 and Doel 4, provided that a number of technical conditions - mainly linked to the heating up and cooling down of the fuel element - are respected:

(14)

Furthermore, modulations would not be possible during certain periods (17), such as:

(15)

Following these studies, the AFNC/FANC authorised the operator to perform up to 30 modulations per cycle (18), which corresponds to the maximum number of modulations that can be performed within the boundaries of the current operating licence. Given these constraints, Engie pointed out that each decision to modulate is cautiously assessed by the operator as it bears multiple risks (e.g., production of liquid effluents because of modulation, faster ageing of the installations, risk of triggering an automatic shutdown, potential damage that would render it impossible to restart the reactor, etc.) which are widely documented by academic literature (19).

(16)

According to Belgium (based on input by Electrabel), at the current state of knowledge, there are no technological solutions available to increase flexibility at a reasonable cost and within a reasonable timeframe. The technical upgrade to turn a baseload reactor into a load following plant and to allow more flexibility would require an entire change in design, and particularly the replacement of the reactor pressure vessel head, which is a very lengthy and complex operation. Besides, it would require new Safety Analysis Report (SAR) studies and a significant modification of their operating licence which would have to be issued by the Belgian nuclear safety authorities.

(17)

Furthermore, Belgium (based on input by Electrabel) explains that the reactors’ flexibility is also constrained by the specificity of their fuel and the outages management:

(18)

As described in section 2.1 of the Opening Decision, in 2003, the Belgian federal Parliament adopted a law which prohibited the construction of new nuclear units aimed at the industrial production of electricity by nuclear fission in Belgium and limited the operation of the already existing reactors to 40 years, hereby establishing a nuclear phase-out between 2015 and 2025 (‘2003 Law’ or ‘Nuclear Phase-Out law’) (23). As initially foreseen by the Nuclear Phase-Out law, Doel 3 and Tihange 2 were permanently disconnected from the grid on 23 September 2022 and 31 January 2023 respectively. By the laws of 18 December 2013 and 28 June 2015 (24), the Nuclear Phase-Out law was amended and the lifetime of the three oldest reactors, Tihange 1, Doel 1 and Doel 2, was extended by 10 years, until 30 September 2025, 14 February 2025 and 30 November 2025 respectively (10-year lifetime extension) (25).

(19)

According to the Nuclear Phase-Out law, Doel 4 and Tihange 3 were to close by 2025. As a consequence, since 2020, Engie’s strategic objectives concerning nuclear activities were to (i) withdraw from nuclear power generation activities in Belgium to de-risk its exposure as a nuclear operator to market price volatility, and (ii) no longer position nuclear power generation as part of Engie’s core business. Since 2020, this withdrawal has resulted in a halt to all studies relating to the extension of its nuclear power plants (all located in Belgium). Engie’s financial communication since 2020 is in line with this strategic objective of withdrawal and has been taken into account in the accounting assumptions used to prepare the consolidated financial statements, in particular in impairment tests (26).

(20)

Table 1 provides an overview of the seven Belgian nuclear reactors, including their ownership status, net capacity, and initial deactivation dates according to the Nuclear Phase-Out law and the subsequent revisions.

Table 1

Overview nuclear power plants in Belgium

(21)

On 18 March 2022, the Belgian Federal Government (hereafter also referred to as ‘Belgian Government’) decided to reassess the nuclear phase-out, by allowing the extension of the operating lifetime (Long-Term Operation (‘LTO’)) of the two youngest nuclear reactors, Doel 4 and Tihange 3, for a period of 10 years (‘LTO Project’). The decision by Belgium was made in the context of the European response to the Russian war against Ukraine (including the need for EU Member States to reduce their gas consumption and gas dependency) and the resulting gas crisis, as well as the already existing concerns about security of supply in Belgium (see also section 2.4 below) given the increased electrification needs (to enable the energy transition) and the low availability of the French nuclear fleet in 2021-2022 (due to unforeseen corrosion issues and extensive maintenance to prolong its operation lifetime) (27).

(22)

Subsequently, in 2022, the Belgian Government started negotiations with the Belgian nuclear operator, Electrabel, regarding the implementation of the lifetime extension. The reasons for not having organised a tender, which were provided by Belgium, and on which the Commission did not raise doubts in the Opening Decision, are mentioned in recital 38 of the Opening Decision, and include:

(23)

Engie, the parent company of Electrabel, was initially hesitant to enter into negotiations with Belgium regarding the LTO Project, claiming that nuclear technology had become too expensive and too risky, and referring to Engie’s intention to stop nuclear operations in Belgium after 2025 (see recital 19). Therefore, the Belgian State agreed with Engie to set up a mechanism to share, in a balanced and transparent manner, the risks and rewards in relation to the lifetime extension of the two reactors. According to Belgium, Engie made it clear from the start that without a risk sharing mechanism and an agreement on nuclear waste stemming from the operation of the seven nuclear power plants in Belgium, it would not consider the lifetime extension of the two nuclear reactors, which forces Engie to substantially modify its company strategy and risk exposure (28).

(24)

The negotiation process has been described in detail in section 3.1 of the Opening Decision and led to the conclusion of an Implementation Agreement on 13 December 2023, with a view to restarting operating the LTO Units before the Winter of 2025-2026. As mentioned in recital 36 of the Opening Decision, the Implementation Agreement consists of three main components, which all have the same purpose of supporting the execution of the LTO Project:

(25)

The Implementation Agreement has been amended twice in the course of 2024 (29). Its content is described in more depth in section 3.

(26)

Some important milestones of the LTO Project are the following:

(27)

As a result, the electricity production during the period 2026-2028 will be relatively low due to more than usual scheduled outages of the two reactors during the Restart Phase (‘scheduled LTO outages (31)’). The unavailability of the LTO Units during the scheduled LTO outages is expected to be 24 weeks per year and this during the first 3 years after the LTO restart date.

(28)

On top of the scheduled LTO outages, a yearly normal outage for nuclear power plants (‘scheduled non-LTO outages (32)’) is expected for the whole period of the lifetime extension (i.e. during the Restart Phase and the Running Phase), up to 1 year before the end of operations for Doel 4, and until the last year of operations for Tihange 3. Each scheduled non-LTO outage is expected to last 6 weeks. As a result, during the first 3 years after LTO restart, the two nuclear reactors are expected to be shut down for 30 weeks per year.

(29)

On top of the scheduled (LTO and non-LTO) outages, there can be unplanned and unforeseen problems which require additional shutdown of the LTO Units. A forced outage rate (‘FOR’) of 10 % has been assumed in the Signing Financial Model underlying the Remuneration Agreement (‘RA’) (33). This implies that both Doel 4 and Tihange 3 have a target availability rate of 90 % over 10 years, when not considering the scheduled LTO and non-LTO outages.

(30)

When including all scheduled and non-scheduled outages, Doel 4 and Tihange 3 have a target availability rate of approximately 68,4 % and 67,4 % respectively.

(31)

The nominal electricity production capacity, annual electricity production and share in national electricity demand in Belgium for Doel 4 and Tihange 3, before and after the lifetime extension, have been summarised in Table 2. As is clear from Table 2, the estimated annual electricity production of the LTO Units is expected to double after the Restart Phase.

Table 2

Key characteristics of Doel 4 and Tihange 3 (pre- and post-LTO)

(32)

As mentioned in recital 17 of the Opening Decision, since 2019, the Belgian TSO, Elia, conducted three national resource adequacy studies (‘2019 NRAA’, ‘2021 NRAA’ and ‘2023 NRAA’), which all identified a need for new capacity by the Winter of 2025-2026, as a consequence of the (partial) nuclear phase-out in Belgium, which started with the decommissioning of Doel 3 and Tihange 2 in 2022 and 2023 (see Table 1), reinforced by the decommissioning of thermal generation capacities in neighbouring countries and problems with the French nuclear assets.

(33)

In order to address these resource adequacy concerns, Belgium set up a capacity mechanism (‘CM’), approved by the Commission in 2021, and amended twice since then (34). The CM is a market-wide measure compensating the readiness of plants to supply electricity in pre-defined periods, regardless of whether they produce or not. The CM aims at addressing resource adequacy concerns in electricity, while supporting the energy transition, and will kick in as of Winter 2025, at the time the two LTO Units are also expected to restart (see recital 26).

(34)

Belgium’s energy mix is currently dominated by gas and nuclear-based electricity generation, although the share of renewables has been steadily increasing over the past years. In 2023, the share of nuclear, gas and renewables in the generation mix was 39,9 %, 21,4 % and 32,7 % respectively (35).

(35)

According to Belgium, the main players in the electricity generation market, in terms of installed capacity connected at the transmission grid level, are Electrabel (9,3 GW in 2023 – 65 %), Luminus (2,2 GW in 2023 – 15 %), RWE (0,7 GW in 2023 – 5 %), Eneco (0,7 GW in 2023 – 5 %) and TotalEnergies (0,6 GW in 2023 – 4 %). Belgium submits that the HHI (36) indicator for market concentration decreased from 5 510 in 2016 to 4 431 in 2023, partially explained by the increased development of renewable energy sources (solar and wind) by non-incumbent market players (37).

(36)

According to Belgium, the main players in terms of electricity generation from installations connected to the transmission grid in Belgium are Electrabel (39,1 TWh in 2023 – 70 %), Luminus (6,8 TWh in 2023 – 12 %), Eneco (2,5 TWh in 2023 – 4 %), TotalEnergies (2,4 TWh in 2023 – 4 %) and RWE (2 TWh in 2023 – 4 %). Belgium submits that the HHI indicator for market concentration decreased from 6 372 in 2016 to 5 143 in 2023, thanks to the increased electricity production from renewable energy sources (38).

(37)

According to Belgium, at the retail level, there were in total 16 electricity suppliers present in Belgium in 2023. The main suppliers in terms of supplied electricity are Electrabel (47 % in 2023), Luminus (18,2 % in 2023), TotalEnergies (5,5 % in 2023) and Eneco (5,3 % in 2023), while many players are very small (39).

(38)

The lifetime extension of the two LTO Units will facilitate the development of an economic activity and contribute to addressing the resource adequacy concerns in Belgium (see section 2.4). In addition, Belgium submits that the lifetime extension of the nuclear reactors also aims at reducing the dependency on imports in general and imported fossil fuels (in line with the REPowerEU objectives) in particular, thereby also contributing to decarbonisation of the Belgian electricity system, as well as supplying baseload capacity in the context of the increased electrification needs in the near future in Belgium.

(39)

Belgium submits that the objective of the aid measure is to overcome a number of market failures which prevent Electrabel from continuing operating the nuclear reactors in Belgium without additional input by the Belgian Government (40).

(40)

Belgium argues that, in the light of these general market failures related to low carbon investments in electricity markets, as well as the additional risks to which a nuclear operator is exposed, a commitment to support the lifetime extension of the LTO Units is needed by the Belgian Government. Therefore, Belgium submits that the objective of the notified measure is to overcome these market failures.

(41)

In response to the doubts the Commission raised in its Opening Decision (see section 3.8), Belgium modified certain parts of the measure.

(42)

In particular, those modifications concern:

(43)

The above-mentioned modifications are set out in detail in section 3.3 of this Decision.

(44)

This subsection describes the different components of the support package for the LTO Project, including the amendments made by Belgium following the Commission’s doubts raised in the Opening Decision (see sections 3.3.1, 3.3.2 and 3.3.3), as well as the alternative financing options considered by Belgium (see section 3.3.4).

(45)

A set of financial support mechanisms has been foreseen to allow for the financing of a timely and safe lifetime extension of the two nuclear reactors. All sub-measures of Component 1 are described in detail in the rest of this section.

(46)

During the formal investigation, Belgium presented the rationale behind all parts of the financial support:

(47)

As mentioned in recital 22, due to the strict timing for the LTO restart, the nuclear operator, Electrabel, identified and agreed with Belgium to undertake certain development activities, which are necessary to enable the LTO restart in time and necessary to meet the Safety Authority’s requirements and expectations, prior to entering into the final transaction. These development activities have been laid down in the JDA, which was last amended on 18 July 2024 (JDA++) (44).

(48)

The JDA++ lays down the conditions under which Belgium pre-funds Electrabel’s costs and expenses for the development activities (45) until all required legislative changes have been adopted and have entered into force (the ‘Legislative Condition’). Shortly after fulfilling the Legislative Condition (on 15 July 2024), Electrabel started to fund its own costs and expenses for the development activities and will continue to do so until the amount of Electrabel’s funding equals the amount pre-funded by the Belgian State (expected by early 2025), after which Electrabel and the Belgian State will fund the costs and expenses for the development activities on a 50/50 basis.

(49)

Belgium submits that the pre-funding by the Belgian State of the costs and expenses for the development activities is limited to any costs and expenses actually (to be) borne by Electrabel, corresponding to 89,807 % of the total costs for the development activities. A control mechanism is set up, as well as a ‘True-Up’ at the end of the contract period. Belgium also submits that the funding arrangements under the JDA++ are on arm’s length and value for money basis.

(50)

The Belgian State will invest, together with the nuclear operator Electrabel, in a joint venture (‘JV’), named BE-NUC, which will own 89,807 % of the LTO Units (as Electrabel currently does). The remaining 10,193 % of the LTO Units will stay in the hands of Luminus. Electrabel and the Belgian State will each own 50 % of BE-NUC and act as equal shareholders in terms of financial participation (provision of equity and Shareholder Loans) and share of power sales earnings. BE-NUC, as co-owner, will bear 89,807 % of the investments needed to extend the operation and the Belgian State will therefore indirectly bear 44,9035 % of the investment costs related to the LTO Project. The remaining 10,193 % will be borne by Luminus.

(51)

Electrabel is and will remain the sole operator of the two nuclear reactors through an Operations and Maintenance (‘O&M’) Agreement (see section 3.3.1.4). BE-NUC will not become a nuclear operator but will have control rights over the operating costs through the O&M Agreement.

(52)

Belgium submitted that it expects no possibility of raising external debt to finance the LTO Project. The LTO Project will be fully financed by its shareholders through equity and Shareholder Loans (see section 3.3.1.2.2 below).

(53)

There will be no purchase by the Belgian State of co-ownership stakes in the LTO Units, but rather a transfer (partial demerger of the relevant assets) from Electrabel to the JV, according to the steps described in recital 49 of the Opening Decision.

(54)

Electrabel will transfer its 89,807 % ownership rights regarding the LTO Units (as well as the related permits and any other assets required) to BE-NUC, in return for the distribution of BE-NUC shares to Engie (being at that time Electrabel’s sole shareholder). The contribution of Electrabel to BE-NUC will be valued in consideration of the scrap value of the building, the value of the land and the value of immovable installations, as described in recital 51 of the Opening Decision.

(55)

This valuation of the contribution of Electrabel is reflected in the purchase price of EUR 24,7 million (subject to adjustments) to be paid by the Belgian State to acquire new shares and retain a 50 % stake in BE-NUC. The Board of Directors of BE-NUC will request a (statutory) auditor or a certified accountant to prepare a report regarding the contribution in kind, assessing notably the applied valuation and the valuation methods used for that purpose.

(56)

Belgium submits that the valuation of assets has no impact on the counterfactual scenario analysis and it provided evidence that the transfer of assets (buildings and land) under the LTO Project is neutral. According to Belgium, the use of conservative methods of asset valuation (based on scrap or fair value) ensures that Engie does not derive any advantage from the transfer of assets.

(57)

In addition, the immovable installations correspond to LTO equipment, which are purchased in the context of the LTO Services, the O&M Services and the JDA Services, and which were installed and incorporated in the LTO Units prior to the partial demerger. For reasons linked to Belgian property law and to avoid double counting, these specific pieces of LTO equipment are not paid directly but are considered in the partial demerger, together with the land and buildings. Furthermore, Belgium confirms that the related cash flows for these pieces of LTO equipment are already reflected in the Signing Financial Model.

(58)

A Shareholders’ agreement between Electrabel, the Belgian Government and BE-NUC was concluded to set up the corporate governance of BE-NUC and each of its shareholders’ rights. According to this agreement, the board of directors of BE-NUC is composed of four directors, two appointed upon nomination of the Belgian Government and two appointed upon nomination of Electrabel. BE-NUC’s chairperson and chief financial officer will always be directors appointed by the Belgian Government. The quorum at the board of directors is a simple majority, and its resolutions are voted by simple majority. Conflict of interest provisions have been put in place.

(59)

As mentioned in recital 55 of the Opening Decision, non-European assets currently held by Electrabel will be transferred to Engie. Engie, as mother company of Electrabel, guarantees that assets worth at least EUR 4 billion remain in Electrabel at the time of the closing of the transaction (based on an equity value calculated as of 30 June 2023). In addition, after closing, other safeguards apply such as the continued and enhanced monitoring of the financial position of the nuclear operator by the CPN/CNV, and the uncapped and uncancellable parent company guarantee granted by Engie for certain obligations of the nuclear operator.

(60)

Belgium argues that the JV constitutes a pari passu investment, as the two shareholders enter into the JV under equal terms and conditions and, as shareholders, with the same level of risk and rewards.

(61)

The role of the Belgian State as shareholder presupposes, inter alia, the financing of BE-NUC’s capital costs (CAPEX) and operating costs (OPEX), the management of shares and the exercise of shareholder rights (e.g., voting rights), and the support of the two directors of BE-NUC, appointed upon nomination of the Belgian State (see recital 58).

(62)

The share capital contribution consists of:

(63)

Belgium explains that the terms and conditions of the Electrabel Shareholder Loan and the Belgian Government Shareholder Loan are identical. Both Shareholder Loans will be granted on market terms, at interest rates that have not yet been set, but would be, according to the Shareholder Loan Agreements, set by the board of BE-NUC in accordance with the Shareholders’ Agreement by reference to prevailing market rates and any comparable third-party debt financing which may be available at the relevant time. Under the procedure agreed between the Belgian Government and Engie, Engie has prepared a term sheet setting the terms and conditions of both Shareholder Loans. The term sheet describes the methodology to set the interest rate. Belgium submits that this methodology is consistent with Engie’s transfer pricing loan policies and is in line with the OECD BEPS principle (46) ensuring that the interest rate it set at an arm’s length level. Pursuant to the term sheet, the interest rate is expected to be a floating interest rate, set at the EURIBOR 6-month rate (floored at 0 %) plus an expected margin of approximately [0-3] %.

(64)

Belgium submits that the introduction of a Shareholder Loan in addition to the equity injection follows from financial and transactional considerations. On the one hand, the provision of the Shareholder Loan grants more flexibility in the design of drawdown and repayment schedules. In particular, loan repayment provisions may be agreed upon with less regulatory constraints than dividend payments or equity repayments. On the other hand, the loan might optimise the financial structure with respect to taxable income (up to 30 % of EBITDA is redeemable to deduct interest).

(65)

Based on preliminary computations in the Signing Financial Model, the total share capital contribution (equity injection and Shareholder Loans) amounts to EUR [2 000-2 500] million (47), to be provided at 89,807 % by both the Belgian Government and Electrabel on pari passu terms in […] instalments from […] to […] to finance, amongst others, the CAPEX of the LTO Project (the remaining 10,193 % being funded by Luminus). The share capital contribution is to be paid back to BE-NUC’s shareholders through a series of share capital reductions and Shareholders loan repayment and is to be remunerated through the distribution of dividends and Shareholder Loan interest. The split of the share capital contribution (EUR [2 000-2 500] million) into equity injection and shareholder loans has not yet been decided.

(66)

Belgium submits that the shareholder funding obligations and the Shareholder Loan can be considered as pari passu financing, and that the Shareholder Loan will be granted on market terms, so that the shareholder funding does not grant a selective advantage to the beneficiaries.

(67)

A Remuneration Agreement (RA) has been concluded between BE-NUC, Luminus and the RA Counterparty, which will be an autonomous service with accounting independence within the Belgian State (‘BE-WATT’). The RA aims at addressing market price uncertainty and de-risking the LTO Project revenues for the owners of the LTO Units, BE-NUC and Luminus. Thereby, BE-NUC and Luminus should receive sufficient revenues from the operation of the LTO Units to ensure their safe and reliable operation and economic viability, while allowing shareholders to contemplate the required market-conform financial return.

(68)

In particular, Belgium clarifies that the financing principles of the transaction are as follows:

(69)

The financial support mechanisms (CfD, MOCP and SDC Loans) are explained in more detail in sections 3.3.1.3.2, 3.3.1.3.3 and 3.3.1.3.4 below.

(70)

In general terms, the (target) IRR is the minimum level of return that investors accept to be compensated for a certain level of risk in an investment project. The target IRR is often referred to as the ‘hurdle rate’, which is the threshold that a project’s IRR needs to equal or exceed before the project will be undertaken. In theory, the hurdle rate equals the sum of the cost of capital (e.g., weighted average cost of capital or ‘WACC’) and a hurdle premium. In other words, it is a combination of the rates of return from comparable projects/assets and adjustments to align with a particular risk profile including project-specific premia to cover additional non-diversifiable risks.

(71)

In the present case, the Levelised Cost of Electricity (‘LCOE’), corresponds to the minimum level of the average electricity price (or the strike price in the CfD design) to be obtained for the LTO Project to achieve its target IRR. As mentioned in recital 72 of the Opening Decision, the preliminary strike price in the Signing Financial Model, based on the assumption of a cost to modernise the LTO Units of approximately EUR [2-2,5] billion, was set at EUR [80-90] per MWh (in 2022 values). Belgium submits that, when considering the application of the (intensified) pain/gain sharing mechanism (MPRA), the strike price, based on the Signing Financial Model, can vary from EUR [80-90] per MWh to EUR [80-90] per MWh (in 2022 values).

(72)

Belgium argues that the target IRR is a conservative rate of return given the risk exposure of the LTO Project, based on: (i) a benchmarking exercise of the rates of return of other nuclear investments globally and other power sector assets in Belgium (see section 3.3.1.3.1.1), (ii) an assessment of BE-NUC’s cost of capital (both WACC and cost of equity (‘CoE’)) based on the CAPM pricing model (see section 3.3.1.3.1.2), and (iii) a simulation of the expected MPRA-adjusted IRR in view of updated market circumstances that shows that the expected IRR of the LTO Project has decreased compared to the time of the negotiation of the Implementation Agreement (see section 3.3.1.3.1.3).

(73)

Belgium provided an international benchmark of the target rates of return and cost of capital for companies operating nuclear plants across various geographies, as well as gas-fired power plants remunerated under the Belgian CM and companies operating other types of energy transmission and storage infrastructure in Belgium. The benchmarking exercise was fine-tuned after the adoption of the Opening Decision, notably regarding a further description of the nature of the regulatory frameworks and possible additional schemes affecting the risk allocation for the companies and projects considered. Table 3 provides an overview of the public target rate of return or WACC estimates of the benchmarked projects and companies.

Table 3

Public WACC estimates for benchmarked projects/companies

(74)

Table 4 provides an overview of the elements addressing risks in the benchmarked projects or companies.

Table 4

Elements addressing risks for benchmarked projects/companies

(75)

According to Belgium, this benchmarking exercise highlights three elements:

(76)

Therefore, Belgium considers the 6 %-8 % target IRR of the Project to be within the likely range of market-based returns that will not result in overcompensation.

(77)

In addition to the benchmarking exercise, in order to confirm whether the target IRR of 7 % lies within the likely range of market-conform returns, Belgium provided an assessment of the cost of capital using two metrics, the WACC and the Unlevered CoE, given that BE-NUC is not debt-financed.

(78)

Belgium obtained estimates for the WACC and unlevered CoE lying in the range of 6,2 % to 7,4 % (which amount to a risk premium for the project of 3,1 % to 4,3 % above the risk-free rate). According to Belgium this demonstrates and further confirms that the target IRR of 7 % is proportionate and does not result in overcompensation. The sections below explain in more detail how the WACC and CoE range were estimated.

(79)

The first metric to estimating the cost of capital is the WACC, which represents the minimum rate of return that the project must offer to attract capital. This minimum return represents the return that potential investors could earn if they decided to invest in another project with equivalent characteristics in terms of cash flow, timing and risk. The post-tax WACC is computed using the below formula:

(80)

The WACC is thus the sum of (i) the CoE weighted in proportion to the market value of equity within the total capital (wE), and (ii) the cost of debt (post-tax) weighted by the share of the market value of debt in the total capital (1-wE). The CoE is calculated on the basis of the Capital Asset Pricing Model (‘CAPM’), a standard model for measuring the return required by an investor to cover the opportunity cost of the value of money over a set investment horizon (the risk-free rate) and for the exposure to market risk (51). The exposure to market risk is captured by the levered equity beta value (ß E ). Due to the volatility of parameters, the inputs to CAPM were computed in ranges.

(81)

Regarding the computation of the levered equity beta, Belgium argues that, as the JV is a new and privately held company holding two existing nuclear LTO Units with expected generation of 10 years, its beta cannot be computed from market or historical data and therefore has to be estimated through a group of comparable listed companies. However, given that there are no comparable utilities that have 100 % exposure to nuclear generation, Belgium considered a sample of five listed European utility companies with the highest exposure to nuclear activities, i.e. comparable projects with 15 %-65 % generation using nuclear power plants, within a diversified portfolio (see Table 5).

Table 5

Asset beta values for comparable utility companies with various nuclear exposure

(82)

Belgium explained that each of the utilities is a diversified publicly traded utility company with some notable exposure to RAB-regulated network assets and/or supported renewable energy assets, unlike the JV which is a private entity with a 100 % exposure to nuclear generation subject to the RA. Therefore, the systemic risks estimated based on these comparable companies would likely underestimate the systemic risk of BE-NUC.

(83)

Belgium used the parameters of Table 6 to put into the WACC formula and, based on these and additional inputs, observed throughout 2023, obtained a theoretical WACC which lies in the range of 6,2 to 7,4 %, including a premium of 3,1 %- 4,3 % over the risk-free rate.

Table 6

BE-NUC’s WACC estimate and key underlying parameters

(84)

The second methodology is based on the CoE, which measures the required shareholder project return, assuming that there is no debt financing, as it is the case for the LTO Project. Although the CoE approach also uses CAPM, it relies on the unlevered (i.e. with no gearing) CoE (52). The unlevered CoE is computed as follow:

(85)

The following approach was used to estimate each parameter of the CoE equation (see Table 7):

(86)

Putting the estimates above in the CoE equation provides an estimate for the unlevered CoE in the range of 6,2% to 7,5% (see Table 7), including a premium of 3,1% to 4,3 % above the risk-free rate, which is in line with the results obtained for the WACC.

Table 7

BE-NUC’s unlevered cost of equity estimates and underlying parameters (full range)

(87)

Taking into account the parameters estimated above (for possible ranges throughout 2023), Belgium estimates three plausible scenarios for the unlevered CoE at the time of the agreement between the parties in 2023, corresponding to a range of 6,8 % to 7,3 % (see Table 8).

Table 8

Three scenarios for BE-NUC’s unlevered cost of equity (reduced range)

(88)

Belgium acknowledges the limitations of the CAPM approach to estimate BE-NUC’s cost of capital given the lack of directly comparable publicly listed companies and the specificities of the risk profile and financing approach of the project (54). In particular, Belgium puts forward the specificity of the nuclear sector and the limited comparability to other groups of assets.

(89)

In addition, Belgium discusses different premia typically required by investors to invest in practice in comparable long-term investment projects over the WACC estimate derived by applying the CAPM approach. Belgium puts forward one possible justification for the need to account for this premium over the WACC estimate in relation to the risk of capital lock-in for long-term investment projects that cannot be diversified away and is not rewarded under CAPM (55). Moreover, another argument put forward to justify the need to account for a premium relates to the shortage of capital to achieve energy policy goals in Europe, as there is substantial competition for investments among available projects which increases the required rate of return to attract capital. Belgium argues that it is common practice among utility companies to add a premium in the range of 1,5 % to 4 % and provides some examples to prove this (56). Finally, Belgium also submits that another justification for accounting for a financing premium over the WACC estimate is the illiquidity premium (57) given the nature of this investment, which is backed by both academic studies (which point to a relevant range of 0,7 % to 7,3 %) (58) and industry practitioners (59). These premia could be added to the estimates of WACC or CoE derived through the CAPM approach, as explained in previous sections, and justify Belgium’s argument that the target IRR is a conservative rate of return given the risk exposure of the project.

(90)

Belgium explains that the MPRA has the effect that the target IRR after MPRA adjustment varies dynamically between a range of 6 % to 8 %. Belgium argues that even if the target IRR is set at 7 % before MPRA-adjustment, under current market circumstances (with lower expected market prices than at the time of the Signing Financial Model) a lower IRR can be expected than what was estimated at the time of the negotiation of the RA.

(91)

To demonstrate this, Belgium provided an analysis of the evolution of market expectations over the period 2025-2035 since the negotiation of the RA based on various sources, and the impact of these revised projections on the expected IRR of the LTO Project (through the application of the MPRA).

(92)

Figure 1 shows the initial price curves at the end of 2023 (i.e. at the time the agreements were signed). Across the various price projections, market prices start - in the central scenario shown in full lines - at a value of EUR 120 per MWh in 2025, declining towards an average value of EUR 87 per MWh in the period until 2033.

Figure 1

Initial baseload price curves in Signing Financial Model (EUR/MWh) (2022 values)

(93)

Figure 2 shows that the current central scenario power price projections (60), displayed with full lines, have decreased compared to the ones used at the time of the Signing Financial Model. […] price projection in the central scenario in Q4 2022, used amongst others as a reference curve in the Signing Financial Model, is for instance presented with a red dotted line in this graph. While, at the end of 2022, prices were projected at EUR 120 per MWh in 2025, they are currently projected at EUR 85 per MWh or even lower in 2025. Market developments since 2022 have reduced projected wholesale electricity prices in the next decade, notably due to decreasing commodity price projections (gas and CO2), among others. Figure 2 also shows that current central scenario projections forecast relatively flat prices between 2025 and 2035, ranging from EUR 69 per MWh to EUR 86 per MWh.

Figure 2

Baseload price curves updated in fall 2024 (EUR/MWh) (2022 values)

(94)

Belgium concludes on the basis of the above figures that market price expectations have decreased.

(95)

Belgium submits that the grey area shown in the figures above represents the price corridor (as initially foreseen in the RA) for MPRA adjustments of the target IRR, i.e. the market price range within which the target IRR is adjusted (61). The lower market price expectations then suggest a decline in the expected IRR adjusted by the MPRA. Under the MPRA parameters initially foreseen in the agreement, the expected IRR of the LTO Project would in Q3 of 2024 be anticipated at 6,7 % (if all the assumptions from the Signing Financial Model are met). With the updated MPRA parameters (see section 3.3.1.3.2), which makes BE-NUC’s revenues and IRR more sensitive to market price conditions, the expected IRR would be reduced to 6,5 % (if all the assumptions from the Signing Financial Model are met) considering the same price projections (see Table 9).

Table 9

Expected IRR (with original and updated MPRA) versus target IRR (Q3 2024)

(96)

The CfD will apply between the Belgian State and BE-NUC, as well as between the Belgian State and Luminus, the other co-owner of the two nuclear reactors. This means that a predefined target price (the ‘strike price’), that will be indexed throughout the RA period, will be guaranteed by the Belgian State for the aggregate metered electricity output of the LTO Units. If the market reference price (‘MRP’) is higher than the strike price, the positive difference multiplied by the actual metered output will be paid by BE-NUC and Luminus to the Belgian State. If the MRP is lower, the negative difference multiplied by the actual metered output will be paid by the Belgian State to BE-NUC and Luminus. The difference payments become payable on the first power date (the date on which the relevant LTO Unit injects electricity into the high-voltage grid for the first time after its initial legal end date) and will be made in proportion to BE-NUC’s and Luminus’ share of the power generated by the LTO Units.

(97)

The main parameters of the CfD are:

(98)

According to Belgium, based on an independent analysis by Compass Lexecon (62), the initial choice of the DAM as MRP in the CfD is suitable in the context of the LTO Project, particularly in the period of LTO works (the Restart Phase until 31 December 2028). Moreover, Belgium argues that flexibility built into in the RA opens the door for alternative MRP choices during the contract duration, depending on the evolution of the Belgian market context and needs, whilst maintaining the risk-return balance of the initial CfD design. Belgium further justifies the use of the DAM based on the following elements:

(99)

A study on the market risk management strategy of BE-WATT (the CfD counterparty) has been procured by the Federal Public Service (‘FPS’) Economy. The ongoing study aims at taking preparatory steps in the operationalisation of BE-WATT, specifically regarding this entity’s task on risk management. In particular:

(100)

The initial choice of the MRP may be revisited by the Belgian Government, as RA Counterparty, up to three times over the contract duration, subject to BE-NUC’s and Luminus’ agreement, as from the end of the initial period of the LTO works, e.g., depending on the evolution of the Belgian market context and needs, whilst maintaining the risk-return balance of the initial CfD design.

(101)

The strike price will be indexed annually by reference to a weighted indexation calculation and can be revised at particular times:

(102)

The strike price will be calculated using information from the detailed financial model which will be produced and updated by BE-NUC, within the parameters notified by Belgium. The financial model (and any updates thereto) is subject to the approval of the RA Counterparty. Where such approval is withheld, BE-NUC and the RA Counterparty may refer the determination of such a financial model issue to an independent expert in accordance with a specified expert determination procedure.

(103)

Belgium submits that the CfD reduces BE-NUC’s exposure to market risk and market price variations.

(104)

Belgium submits that, while the CfD reduces BE-NUC’s exposure to market risk and market price variations, it includes risk-sharing mechanisms which should ensure that BE-NUC is still exposed to some market risk and incentives. In particular, the RA includes a pain/gain sharing mechanism (the MPRA) when market prices turn out to be lower or higher than the strike price.

(105)

The objective of the MPRA is to incentivise BE-NUC to optimise its cost structure prior to setting and revising the strike price (2025/2029), maximise the output of the reactors when high prices are expected and the electricity system nears scarcity, and mitigate windfall profits.

(106)

The MPRA is derived and calculated as follows:

(107)

In the initial design of the MPRA (‘initial MPRA’), the range between the strike price and upper (respectively lower) threshold was divided into 30 positive (respectively negative) ‘MPRA values’. Compared to the ‘initial MPRA’, the ‘updated MPRA’ is triggered faster, i.e., the pain/gain sharing mechanism corresponds to a market price corridor of intervals of +/-20 % instead of +/-30 %.

(108)

As a result, Belgium submits that the (updated) MPRA provides BE-NUC with strong(er) incentives to:

(109)

Moreover, Belgium submits that, in the modified CfD design, the decision-making authority on when an economic modulation decision should be taken is transferred to the EMSA partner (see section 3.3.1.5). The EMSA remuneration will not only be fixed as initially planned, but will in addition to a fixed portion be made variable, and thus incentives will be provided for making best use of the limited stock of modulations of 30 per fuel cycle (see recital 13). Compared to the initial CfD design, by which modulation decisions were taken based on fixed modulation threshold, the modified CfD design does not rely upon a fixed modulation threshold anymore but incentivises that modulations are executed at times when they pay off most (i.e., when prices are expected to be deeply negative for a sufficiently long period of time). The EMSA remuneration is detailed in section 3.3.1.5.2.

(110)

Finally, Belgium submits that, for the reasons provided in recital 80 of the Opening Decision, the CfD is an appropriate and proportionate instrument to tackle the identified market failures and specific nuclear risks, hereby achieving the objectives of the measure, while preserving efficient market signals. Belgium clarifies that any proceeds from the CfD will flow into the general State budget but will be subject to separate accounting. They will be used primarily to fund the payments of the RA Counterparty under the CfD. Where the CfD proceeds would exceed the amounts necessary to finance the costs of the CfD, they could then be used to finance the costs of another CfD. Belgium commits that any remaining CfD proceeds would be used for purposes of distributing them to undertakings. Belgium further commits that it will inform the Commission in case CfD proceeds would be distributed to undertakings, and, if need be, notify such a measure. Belgium also confirms to have included penalty clauses in case of undue unilateral early termination of the CfD contract.

(111)

If BE-NUC’s revenues are not sufficient to cover the costs payable in any month under the O&M Agreement, as well as any other operating, fuel and maintenance CAPEX costs required for the operation of the LTO Units, the RA Counterparty has to make a shortfall payment to BE-NUC to ensure sufficient cash flow to meet these costs, in order to ensure the long-term economic viability of the JV (63).

(112)

The MOCP consists of two components: Minimum OPEX: (i) a revenue shortfall payment to ensure sufficient cash flows to meet costs required for the operation of the LTO Units and safeguard the long-term economic viability of the JV and (ii) a 50 % cost protection on the invested capital in relation to the amortised capital costs of the LTO extension (every three years).

(113)

Belgium argues that any single event that would reduce the availability of the LTO Units for a significant period during the year and/or repeated outages over several years may result in significant losses for BE-NUC.

(114)

Belgium asked Electrabel to evaluate the risk of major unexpected events throughout the lifetime of the LTO Units based on historical data. Electrabel argues that there is a high probability of a significant unplanned event (64) for the 10-year period of the lifetime extension of the LTO Units. In order to evidence this, Electrabel identified historical instances of unavailability for each of the seven nuclear reactors over the 11-year period between 2012 and 2022 (see Table 10). Electrabel claims that this historical approach provides useful insights for the following reasons:

(115)

The data set out in Table 10 show that there were events of significant unplanned unavailability outside of the nuclear operator’s control in all seven nuclear units over the period 2012-2022, which resulted in an average unavailability of 59 % per unit (in additional to planned outages), corresponding to an average unavailability of 7 months/per event. For the seven Belgian nuclear units together, there were 11 events of significant unplanned unavailability over the period 2012-2022. The reasons for unexpected unavailability events included hydrogen flakes in reactor pressure vessel, degradation of concrete (requiring important repairs), equipment failures, leaks and a sabotage. Therefore, Belgium submits that there was an average of around 1,6 (11/7) significant unplanned unavailability events per reactor per decade.

(116)

Belgium submits that, when there are multiple nuclear units (or a nuclear fleet) in operation, the revenue loss during a breakdown period of one or more units can be partially compensated by the revenue generated by the other units. In the case of the LTO Project, however, these systemic risks (of unavailability events affecting most units at the same time) are expected to intensify as Belgium's nuclear fleet shrinks from 2025 onwards and becomes less diversified, even more so as the two LTO units are of the same technology. A major event affecting both LTO Units for a substantial part of the year is therefore not deemed unlikely.

(117)

The 59 % probability of unit unavailability, which is equivalent to approximately 7 months, corresponds to the average duration of one significant unplanned unavailability event. Electrabel considers this duration to be a realistic estimate based on its own experience with the operation of the LTO Units in Belgium, as significant events require important engineering works, the procurement of spare parts, a safety impact analysis, and extended discussions with the nuclear safety authority before the nuclear reactors can start up again.

(118)

Belgium further explains that the regular PSRs, based on IAEA Guideline SSG-25 (66), which have been conducted every 10 years from the start of operation, have resulted in significant design upgrades and investments to cope with the ageing and obsolescence of the nuclear units and help to maintain their safety and reliability. Nevertheless, despite these periodic reviews, the nuclear operator cannot exclude the possibility that new problems will occur which trigger significant unavailability periods for the nuclear units.

(119)

In order to estimate the impact on BE-NUC of the occurrence of a significant unplanned unavailability event during the period of the lifetime extension of the LTO Units, and the effect of the MOCP, Belgium considered plausible scenarios of major unavailability events (see Table 11) based on data from Table 10. As with other assessments, for the illustrative computations, Belgium relied on the Signing Financial Model dated 13 December 2023.

(120)

Belgium submits the analysis above shows that the MOCP is designed to cover the effects of potential major events threatening the economic viability of the JV, and leaves the JV to bear many risks of unavailability: in Scenario 1, the MOCP is not triggered if an average major event were to take place, and does not guarantee the returns to the shareholders who continue to bear the risk of not recovering their investments.

(121)

The MOCP might in some scenarios potentially lead to a relatively significant cost for the Belgian State, as it will get triggered to address the high-impact unavailability events, as illustrated above. In order to address this concern, Belgium introduced a cap on the MOCP payments. Accordingly, the RA Counterparty (the Belgian State) will exercise its termination right under the RA (67) in case the MOCP reaches a paid-out amount of EUR 2 billion (‘MOCP Trigger’), provided this appears to be appropriate in light of, among others,

(122)

Belgium submits that there are circumstances in which a termination would not be appropriate:

(123)

In case the RA Counterparty decides not to exercise its termination right(s), Belgium will submit its reasoning to the Commission for evaluation purposes.

(124)

The amount of the MOCP Trigger corresponds to the upper bound of MOCP payments in a scenario when both LTO Units are simultaneously unavailable for two consecutive years. The expected MOCP payments from two-year closures will indeed depend on the year when the said closures take place. Belgium provided yearly computations for annual MOCP payments and totals over two consecutive years (see Table 12) based on preliminary forecasts from the Signing Financial Model dated 13 December 2023.

Table 12

Simulation MOCP payments

(125)

To estimate the two-year reference period used to calculate the proposed EUR 2 billion cap on the MOCP, Belgium provides some examples of nuclear reactors that have been permanently shut down for technical reasons in the United States (since this has never occurred in Europe so far) (68). According to Belgium, these examples show that the two-year reference period is in line with industrial reality, as a period of two years after the detection of a problem requiring the shutdown of both LTO Units can usually be considered as a reasonable cut-off point for questioning the continuation of repairs or for considering the final shutdown.

(126)

In order to make the MOCP operational, BE-NUC will procure, either from its shareholders or an external party, a working capital facility (WCF) at the latest on the first LTO restart date to occur. The WCF serves to fund the need in working capital stemming from the operation of the LTO Units. BE-NUC will be allowed to draw down the WCF if the difference between its cash inflows and cash outflows is smaller than the estimated operational expenditures of the upcoming spending period set out in the RA. The amount of the WCF shall be at least the average aggregate estimated operational expenditure for a period of three months. In effect, the WCF serves as an intra-year bridge to the annual MOCP, acting as a revolving credit facility that would be repaid yearly, if drawn down, by the MOCP provided by Belgium.

(127)

The terms of the WCF, which shall be procured on market terms at the latest on the first LTO restart date to occur, are not yet known. However, Belgium further clarified that the interest rate of the WCF will be determined using the same methodology as the one agreed upon for the Shareholder Loans (see recital 63).

(128)

In addition to the MOCP which provides for financial stability throughout the entire 10-year period of the lifetime extension, during the first three years of the LTO Project (Restart Phase), the Belgian Government will grant a loan to BE-NUC and Luminus on identical terms. The loans will be sized by reference to their proportionate share in the LTO Units, and consequently their respective share in the shutdown and operating costs as of 1 July 2025. As mentioned by Belgium, the SDC Loans are provided to ensure liquidity through 2028, as operating costs cannot be financed through cash flows due to LTO works.

(129)

The SDC Loans, which are provided at a capped interest rate and repayable according to a specified repayment schedule, are composed of two different facilities (i.e., one per LTO Unit), each formed of two tranches:

(130)

The terms of the SDC Loans are provided in recital 96 of the Opening Decision and have not been amended. The SDC Loans are repayable at an interest of up to 6 % and pari passu with shareholders’ returns. Regarding the methodology used to determine the interest rate, Belgium explains that this was the result of a negotiation between the parties to the RA, and that the maximum rate of 6 % corresponds to the lower bound of the target IRR range of 6 % to 8 % (considering the MPRA).

(131)

As stated in recital 97 of the Opening Decision, the SDC Loans are expected to be drawn down for an aggregated amount of EUR [500-700] million in […] instalments from […] until […], repaid in three instalments from […] until […] including capitalised interests. These computations and amounts will be updated in the financial model approved by the RA Counterparty in the course of 2025, prior to the LTO restart date, based on the cost of extending operation under nuclear safety requirements set out by the Belgian nuclear safety agency, estimated on the basis of submitted quotes by contractors.

(132)

Belgium argues that the SDC Loans do not disproportionally benefit the shareholders: the risks of not realising the expected returns are shared proportionally by the JV shareholders and the RA Counterparty as with lower profits, both the IRR of the LTO Project and the SDC Loans will decrease. In fact, the SDC Loans will not be repaid (at all) only if the LTO Project does not generate sufficient profits to pay any returns on the shareholders’ investment (69). Consequently, Belgium argues that the SDC Loans bring no additional risks to the RA Provider beyond the risks of the LTO Project’s performance and profitability, as the facility is not used to finance the capital expenses (only the amortised capital costs of the LTO extension during the Restart Phase).

(133)

Under the O&M Agreement, Electrabel is to perform for BE-NUC:

(134)

Certain services are explicitly excluded from the O&M Agreement, including services, works or activities in respect of decommissioning and dismantling of the LTO Units, which remain under the responsibility of Electrabel (see section 3.3.2.4).

(135)

Pursuant to Article 12.1 of the O&M Agreement and subject to certain adjustments and exceptions, BE-NUC will pay Electrabel 89,807 % (reflecting Luminus’ holding of 10,193 % of the LTO Units) of all costs incurred in the provision of the LTO services and O&M services plus the relevant margin, being:

(136)

Belgium submits that the levels of margins are aligned with those applied under the LTO Partnership Agreement with Luminus (which itself covers a wide range of services including but not limited to O&M). The original agreement covering similar services to a third party, Luminus, concluded on 26 June 2003 and re-confirmed on 13 December 2023, is a relevant reference to support the position that the O&M Agreement reflects arm’s length costs for nuclear operations. In addition, Belgium argues that the financial risks borne by Electrabel are greater than under the Partnership Agreement with Luminus, since, under the O&M Agreement, the margin of Electrabel will be reduced in case of (non-excusable) cost overruns (i.e., costs not included in the budget as proposed by Electrabel and validated by the parties) and in case of unavailability of the plant beyond a target.

(137)

In addition, Belgium submits that the O&M Agreement includes certain cost controls, including rights for BE-NUC to audit Electrabel’s calculation of the fees and performance of the services and to request a benchmark review of the prices charged by Electrabel for technical affiliate services.

(138)

Finally, as the (sole) operator of the LTO Units and a service provider to BE-NUC under the O&M Agreement, Electrabel will be incentivised to achieve technical and economic performance of the LTO Units. In particular, under the O&M Agreement:

(139)

As a consequence, Belgium concludes that the O&M Agreement is limited to covering of costs incurred and that the financial conditions of the O&M Agreement are set to reflect market terms.

(140)

Although BE-NUC will be the technical owner of 89,807 % of the electricity produced by the LTO Units (the remaining 10,193 % is owned by Luminus), it transfers the ownership of that electricity to the energy manager, who will sell it (70). To this purpose, BE-NUC will enter into an EMSA with a partner purchasing the electricity from BE-NUC and selling it.

(141)

The EMSA will set out key terms, conditions and risk allocations and will, consequently, set out in a detailed manner a BIS and the role of the energy sales manager (‘EMSA partner’) in relation thereto, within the set parameters. The EMSA partner appointed under the EMSA will be the purchaser and the owner of BE-NUC’s share of the electricity produced by the LTO Units and is subject to a pre-defined BIS. The BIS can be reviewed and amended from time to time following a procedure detailed in the RA. The Belgian Government (BE-WATT), in its capacity as RA Counterparty, can impose the BIS to the extent it respects the BIS conditions as laid down in the RA. The EMSA partner has an advisory role in the matter.

(142)

In order to bring the electricity produced to the wholesale electricity market in a competitive and transparent manner, Electrabel and the Belgian State have agreed to run a public tender for these services, ensuring transparency and competition (see section 3.3.1.5.1). Public procurement law and all related necessary safeguards apply to the tender procedure.

(143)

Belgium submits that the tender procedure is being voluntarily conducted in accordance with the Belgian law on public procurement of 17 June 2016 and the Royal Decree of 18 June 2017 on public procurement for the utilities sectors (71).

(144)

The EMSA contract will be awarded through a negotiated procedure with a prior call for competition (72), which is a standard procedure in the utilities sector. The call for tenders has been published on the relevant Belgian (73) and EU platforms (74), allowing all potential candidates to take part in the tender procedure.

(145)

Given the importance of the service tendered and the sensitivity from a competition point of view, Belgium implemented additional provisions and safeguards to ensure that the objectives of the EMSA will be attained and that the expected services will be delivered adequately:

(146)

In addition, Belgium submits that the tender includes a set of selection criteria to ensure that the conditions for professionalism and required infrastructure for an orderly execution of the EMSA are met by any tender candidate. These qualification criteria are however limited to the minimum extent possible, referring each to a necessary requirement for the services intended, to ensure maximum competition during the tender.

(147)

In comparison to the information at the Commission’s disposal at the time of the Opening Decision, Belgium provided more details on the qualification criteria. The tenderer (or, for the first four criteria described below, through reliance on the capacities of another entity) must:

(148)

During the term of the EMSA contract, the following conditions apply among others to the EMSA partner:

(149)

With respect to the timing of the tender, Belgium submits that it aims to conclude the tender in due time, at the latest by 15 May 2025, so that the EMSA partner would have enough time to do the necessary preparations before the restart of the LTO Units.

(150)

The tendered services comprise, in relation to the power assets, among others (i) ‘Forward and realised Power Assets representation services’, (ii) ‘Day-ahead (‘DA’) and intraday (‘ID’) Optimisation and Operations services’, (iii) services related to market regulatory obligations, (iv) Belgian TSO contracting services, and (v) trading, market access and market analysis. Because of the interdependency between the various tasks listed, and the regular market practice to combine these services into a single commercial contract to ensure operational efficiency, these services will be integrated into a single EMSA service definition.

(151)

Belgium has clarified that the EMSA partner is not trading as an agent or on behalf of BE-NUC but is selling the purchased electricity since it is the owner thereof. In addition, in the modified CfD design, the EMSA partner obtains the decision-making authority regarding economic modulations in order to incentivise an efficient use of the limited stock of modulations, and to better respond to market signals (see recital 153 and following).

(152)

Belgium submits that the settlement between the EMSA partner and BE-NUC is based on the DAM, in order to limit the basis risk for BE-NUC, as this implies that, in combination with BE-NUC’s difference payment through a CfD where the MRP is the DAM price, BE-NUC receives the strike price (see section 3.3.1.3.2) (82). Belgium confirms that the EMSA partner will be free to trade in the market of its choice and is not obliged to bid on the DAM.

(153)

As explained in section 2.1 and recital 109, since the LTO Units are limited to 30 modulations per cycle, and hence do not have the technical and regulatory possibility to modify their output at any point in time, the remuneration of the EMSA partner has been modified in order to provide incentives to make the best use of the limited amount of authorised modulations. Instead of having merely a fixed fee (determined through the tender), the modified remuneration formula also includes a variable fee (based on a predetermined formula and depending on the modulation and correction settlement payment). Such remuneration exposes the EMSA partner to part of the potential revenue upsides from modulation at times of negative prices (as well as part of the potential revenue downsides from modulations at times of positive prices, if any), while exposing the EMSA partner directly to imbalance and intraday costs as set out below. Therefore, the variable fee will consist of a fixed percentage share (‘alpha’) of the revenue upsides triggering a modulation in DA and a differentiated percentage share (‘beta’) of intraday and imbalance costs and revenues, i.e., the correction settlement payments (which may comprise revenue upsides from triggering a modulation in ID), whereby:

(154)

The formula set out in Figure 3 applies to the entire volume during modulations, including if modulation occurs when prices are positive (for instance during the ramping up/down periods) and not only during negative price periods (84). The variable fee is paid by BE-NUC under the EMSA, but these costs and/or revenues will be passed-on to BE-WATT under the RA.

(155)

According to Belgium, through the modified remuneration formula, the EMSA partner is incentivised, among others, to use economic modulations in an optimal way and throughout the day-ahead and intraday timeframe and to reduce intraday and imbalance costs. Belgium submits that the alpha and beta parameters are chosen to preserve efficient market functioning, while avoiding deterring risk-averse tender participants.

(156)

The modified remuneration formula will affect the bidding behaviour of the tender participants for the fixed fee. The bidders will factor in the expected upsides from modulation (option value) in their bids, as well as the costs and potential revenues related to intraday and imbalance payments. Belgium submits that the tender documents will explain the remuneration (incentive) structure so that the upside/downside potential of the modified remuneration formula is transparent for the bidders. Belgium will evaluate the appropriateness of the EMSA partner’s (modified) remuneration formula and the market-conformity and efficiency of the incentives it provides on the different markets. The evaluation will be done at the latest two years after the start of operations, so as to allow a timely revision of the contract or the organisation of a new tender, which is possible after 3,5 years (85). However, in order to keep the appropriate incentives for the EMSA partner, only an upward revision of the parameters alpha and beta is possible (86).

(157)

Belgium argues that the adaptation of the EMSA remuneration aligns the incentives of the EMSA partner, who has the economic modulation decision-making authority (subject to the operator’s final decision, e.g., in case of safety or regulatory concerns), to market signals, hereby addressing the Commission’s concerns in this regard. At the same time, the EMSA partner has the appropriate incentives to fulfil its task as BRP without being fully exposed to imbalances that are outside its own control. In addition, Belgium submits that there will be continuous accurate information sharing (including regular updates on the availability, technical capabilities, etc. of the LTO Units) between BE-NUC, the nuclear operator and the EMSA partner to allow for informed economic modulation decisions by the EMSA partner, which will be equipped to make efficient decisions.

(158)

Belgium also confirms that the CfD counterparty (BE-WATT) will develop a risk management strategy for its open position, as is legally foreseen, and that its implementation will contribute to liquidity to the forward electricity markets (see recital 99). The adoption of the strategy is subject to an advice of the regulator, which will include an assessment of the impact of the strategy on the relevant electricity markets.

(159)

The EMSA partner will be in principle selected through a transparent, open and competitive tender (see section 3.3.1.5.1). Engie’s trading entity GEMS (Global Energy Management & Sales), a business unit of the Engie group managerially independent from the business unit Nuclear, can also participate in the tender. Specific provisions and measures have been foreseen in case GEMS participates in the tender procedure in order to ensure a fair process.

(160)

Belgium submits that sufficient measures are and will be taken and implemented to effectively identify and prevent any potential conflicts of interest. These measures can be summarised as follows:

(161)

If the tender procedure fails to appoint an appropriate candidate, by way of ultimate fallback, if no EMSA has been concluded in due time, at the latest by 15 May 2025, GEMS will temporarily perform the EMSA services. These services will be rendered in accordance with terms to be agreed between the parties during a limited period, in order for BE-NUC to award the EMSA via another tender procedure. In the case of a successful tender procedure, the successful candidate will take over the performance of the EMSA services from GEMS once the relevant period has lapsed during which GEMS has performed these services. There is no limit to the number of re-tenders.

(162)

Belgium argues that this arrangement is necessary and adequate to ensure the continuity of the public service (i.e., the sale of the electricity generated by the nuclear units), but this solution is strictly limited in time and to the minimum necessary.

(163)

In conclusion, Belgium submits that it will ensure that BE-NUC will follow public procurement legislation and principles rigorously. The extensive consultation process provides additional safeguards ensuring that the purchase of services is carried out through a competitive, transparent, non-discriminatory, and unconditional tender procedure. Belgium therefore considers that any risk of market foreclosure and other potential anticompetitive practices by Engie is avoided and that all measures and safeguards have been adopted to ensure that an independent EMSA partner will be appointed. Only as a fallback, GEMS would temporarily provide these services, which is necessary and adequate to ensure the continuity of the public service.

(164)

Two other sub-measures that are part of the transaction constitute the Administration Services Agreement (‘ASA’) between Electrabel and BE-NUC and an agreement stipulating indemnification arrangements for cost coverage losses in the event of no closing of the agreement. These sub-measures have been explained in sections 3.3.10 and 3.3.11 of the Opening Decision.

(165)

As explained in recital 121 of the Opening Decision, the legislative framework applicable to radioactive waste and spent fuel in the EU is grounded on the following two fundamental principles: (i) the ‘polluter pays’ principle (87), and (ii) Member States’ ultimate responsibility (including financial responsibility) for the responsible and safe management (including disposal) of spent fuel and radioactive waste (88).

(166)

As explained in section 3.4.2 of the Opening Decision, under the current regulations in Belgium, the nuclear operator is financially (through the nuclear provision company, Synatom, and together with EDF Belgium and Luminus (the ‘Contributing Companies’)) and operationally responsible for the decommissioning of the seven nuclear power plants, as well as for the conditioning and management of the radioactive waste and spent fuel, and its long-term storage after its acceptance by ONDRAF/NIRAS until its final disposal. The nuclear provisions for spent fuel and decommissioning waste are funded by Electrabel and the Contributing Companies, managed by Synatom, and are subject to the prudential control of an independent public authority, the Nuclear Provision Commission (‘CPN/CNV’). As noted in recitals 123 to 125 of the Opening Decision, Belgium submits that Electrabel will keep certain responsibilities as sole operator of the LTO Units, resulting from (i) European and Belgian legislation, and (ii) contractual obligations of the Implementation Agreement, hereby respecting the ‘polluter pays’ principle.

(167)

CPN/CNV and ONDRAF/NIRAS are the supervising authorities in Belgium:

(168)

As mentioned in recital 129 and Table 4 of the Opening Decision, based on the last triennial revision of the CPN/CNV of July 2023, the current total amount of provisions related to nuclear liabilities amounts to EUR 18 225 million, consisting of provisions for dismantling activities (EUR 8 122 million), spent fuel management (EUR 9 070 million) and operational waste (EUR 1 033 million) (89).

(169)

As part of the negotiations on the LTO Project, Engie and the Belgian State agreed on a cap on the long-term liability of producers of radioactive waste resulting from the production of electricity through nuclear energy (the ‘Waste Cap’ agreement), to reduce uncertainty in this respect (see section 3.4.3 of the Opening Decision). The ‘Waste Cap’ agreement provides for the transfer of financial liabilities in relation to the production, detention or ownership of conditioned radioactive waste and spent fuel of all seven Belgian nuclear units, subject to, and after compliance with, the relevant contractual transfer criteria, from the nuclear operator (Electrabel) to the Belgian State against the payment of a lumpsum amount (90).

(170)

This lumpsum amount has been established per category of nuclear waste. As explained in recital 131 of the Opening Decision, radioactive nuclear waste can be divided into three categories, in line with the historical radiological classification by ONDRAF/NIRAS and consistent with the International Atomic Energy Agency (‘IAEA’) classification: category A-waste (short-lived waste with a low or intermediate level of radio activity); category B-waste (long-lived waste with a low or intermediate level of radio activity); and category C-waste (short- and long-lived waste with a high level of radio activity) and spent fuel.

(171)

Table 13 below shows the allocation of the existing provisions of EUR 18 225 million between Electrabel (EUR 8 410 million) and the Belgian Government (EUR 9 815 million), according to the Waste Cap agreement, and includes the allocation per category of waste.

Table 13

Breakdown of the Waste Cap agreement

(172)

The principles of the Waste Cap agreement have not been amended and are described in recital 133 of the Opening Decision. They can be summarised as follows:

(173)

As mentioned in recital 107(a) and explained in detail in sections 3.4.5.1 and 3.4.5.2 of the Opening Decision, the establishment of the EUR 15 billion of Capped Amounts is based on the current amount of nuclear provisions of the nuclear operator (the base amount) and a risk premium. Belgium has clarified that the base amount includes already margins for contingencies, uncertainties and other risks that may arise in relation to dismantling, radioactive waste management, and spent fuel management. The provisions regarding the payment of the Capped Amounts are explained in section 3.4.5.3 of the Opening Decision and have not been amended.

(174)

Belgium has clarified that the Capped Amounts to be transferred to the Belgian State already include all the historical and estimated future nuclear waste liabilities (until the original legal end date of all the Belgian nuclear power plants in 2025). Therefore, all nuclear waste and spent fuel produced, or to be produced, by the Belgian power plants during their legal operating life are already taken into account in the Capped Amounts.

(175)

Belgium also clarifies that the actual volume of waste will only be known once the waste is conditioned and physically transferred to the Belgian State. The lumpsum of EUR 15 billion therefore only covers the predetermined volumes of the concerned waste (the volume credits) agreed in the Implementation Agreement. If the actual volume of the concerned waste produced is larger than the volume credits, Engie will pay a volume adjustment fee for each additional volume credit (see recital 133 of the Opening Decision).

(176)

Similarly, all operational waste (‘LTO Waste’) and spent fuel (‘LTO Spent Fuel’) resulting from the lifetime extension of the LTO Units, will be financed by BE-NUC and Luminus for each incremental waste amount. For each incremental amount of LTO waste and LTO spent fuel, two sets of payments will be required:

(177)

Belgium further clarifies that the Signing Financial Model considers the LTO Waste and the LTO Spent Fuel management or back-end costs, estimated at approximately EUR 0,9 million per assembly (in 2022 values), of which approximately EUR 0,3 million per assembly for the on-site storage and approximately EUR 0,6 million per assemble for the LTO Waste Volume Adjustment Fee for Spent Fuel.

(178)

As explained in section 3.4.4 of the Opening Decision, Belgium created a new sui generis public institution, Hedera, whose role is twofold:

(179)

The amounts of the Waste Cap received by Hedera must be secured and invested, to generate the necessary return to pay the costs for the management of the transferred waste when they are due. Hedera is a segregated fund under the supervisory control of CPN/CNV, the dedicated national body playing a key role in ensuring and securing the proper management, control and use of the funds (93). This is reflected in the double role of Hedera, by ensuring that: (i) the amounts received are secured and generate the necessary returns (94), and (ii) the costs of the management of the waste and spent fuel and of the transferred liabilities are sufficiently controlled (95). In addition, the financial resources will have to be sufficiently ringfenced from the general budget of the Belgian State (as provided for in the Hedera Law), so that they are only used to pay for the costs for the long-term storage and final storage and cannot be used for other purposes or to absorb any future budget deficits.

(180)

Belgium submits that the Waste Cap agreement includes and is accompanied by other risk mitigating measures, including:

(181)

As noted in recital 172 of the Opening Decision, Belgium acknowledges that the agreement regarding the nuclear waste has a positive impact on the risk profile of the nuclear operator, since a significant part of its nuclear liabilities will be covered by the Capped Amounts, and that this different risk profile justifies and requires a revision of the package of measures to ensure the safety and supervision of the nuclear operator’s financial situation: the lumpsum payment of EUR 15 billion to Hedera justifies the release of Electrabel’s non-European assets from Electrabel’s perimeter (and the accompanying monitoring of the CPN/CNV). Engie, as the French parent company of Electrabel, will ensure that, at the time of closing the agreement between the Belgian State and Electrabel, at least EUR 4 billion of assets (equity value as of 30 June 2023) will remain in Electrabel. In addition, Engie grants an unlimited and non-cancellable parent company guarantee at first request for (i) Electrabel’s decommissioning obligations (which also includes the risk that the value of the provisions is insufficient), (ii) volume risk under the cap and (iii) the repayment of (current or future) loans with Synatom.

(182)

Belgium considers that the transferred liabilities regarding nuclear waste and the additional decommissioning liabilities resulting from the LTO Project do not provide an economic advantage to Electrabel and the Contributing Companies, considering that they adequately reflect the risk of subsequent cost fluctuations taken over by the State, and that the set-up of the risk transfer is such that a private investor would have accepted to bear it.

(183)

Belgium submits that the nuclear provisions that have been collected by the nuclear operator/Synatom before the Waste Cap agreement, were calculated by using a discount factor of 2,5 % for dismantling activities (including dismantling waste) and a discount factor of 3 % for all activities concerning spent fuel (as decided by the CPN/CNV during the triennial review of end 2022).

(184)

Belgium clarifies that the use of two different discount rates to compute the provisions (dismantling versus spent fuel liabilities) is justified by the difference in duration and reflects the time value of money (97).

(185)

As noted in recital 198, the Waste Cap agreement only entails a transfer to the Belgian State of transferred nuclear waste and spent fuel liabilities, while decommissioning and dismantling liabilities remain with the nuclear operator (except for the dyssynergies caused by the LTO Project, see recital 199). Therefore, Belgium argues that for the computation of the present value of the transferred nuclear waste and spent fuel liabilities a unique discount factor of 3 % (1 % real rate + 2 % inflation) was retained to account for the fact that the nuclear waste and spent fuel liabilities have a longer duration than the decommissioning liabilities (see recital 148 of the Opening Decision).

(186)

Belgium asked the CPN/CNV’s advice regarding the most appropriate discount factor to be used for the calculation of the Waste Cap (98). As mentioned in recital 149 of the Opening Decision, the CPN/CNV proposed to adjust the discount factor for nuclear waste/decommissioning and spent fuel liabilities, applying a two-step approach, consisting in applying a discount factor based on the actual 30-year OLO rate of 3,17 % for the first 30 years, and applying a discount factor of 2,17 % based on an estimation based on OLO rate of the risk-free rate for the 30 years thereafter.

(187)

Belgium recalls that the negotiation period was long (1,5 years) and that the choice for a single discount factor of 3 % was motivated by the need to have a certain degree of predictability with respect to the discount factor used in the negotiations with Engie. This is because the discount factor has a direct impact on the calculation of the lumpsum to be transferred. Belgium explains that the methodology used by the CPN/CNV in its advice of 2023 was later used, together with the Ultimate Forward Rate (‘UFR’) from the European Insurance and Occupational Pensions Authority (‘EIOPA’) (99) as an ex-post confirmation that the discount factor of 3 % was a prudent long-term rate.

(188)

First, Belgium requested an independent consultant to calculate the ‘single’ discount factor equivalent to the two-stage approach of the CPN/CNV, using the risk-free rates suggested by the CPN/CNV, as well as using different other OLO-rate data (100). The analysis shows that the approach suggested by the CPN/CNV is particularly sensitive to the period retained for the calculation. For instance, the two-stage approach results in an equivalent discount factor of 2,8 % using the risk-free rates suggested by the CPN/CNV, an equivalent rate of 3,1 % when applying the CPN/CNV methodology with the OLO-rate data available at the time of the Opening Decision, and an equivalent rate of 3,2 % at the time of the signature of the Implementation agreement. As shown in Table 14, the two-stage approach suggested by the CPN/CNV yields two equivalent discount rates with a 40 base points difference when considering two different time periods that are only 9 months apart.

(189)

Second, Table 14 also shows the comparison with the UFR from the EIOPA, since the EIOPA reference was also used in the German precedent case (SA.45296 (101)), where a discount factor of 4,58 % was found proportionate.

Table 14

CPN equivalent rate and EIOPA rate evolution since the CPN Advice of March 2023

(190)

As is clear from Table 14, the retained discount factor of 3 % falls between the

(191)

In addition, the 3 % rate appears more conservative than the CPN/CNV suggestion and the UFR at the time of signature of the agreement and thereafter.

(192)

Belgium further submits that a discount factor of 1 % in real terms appears conservative in comparison with the rate of return that can be expected by Hedera using an appropriate prudent asset-liability management approach. As a comparative example, Belgium refers to the dedicated asset portfolio of EDF to ensure secure financing of its long-term nuclear liabilities, described in the Universal registration document 2023 of EDF (103). This portfolio is managed based on an asset-liability approach and consists of a portfolio of diversified assets comparable to the description of the hypothetical portfolio managed by Hedera, for which a comparable return can therefore reasonably be assumed.

(193)

Finally, Belgium assumes in the calculation of the Waste Cap a constant inflation rate of 2 %, which is the explicit long-term inflation target of the European Central bank (‘ECB’). The ECB considers that the Consumer Price Index (‘CPI’) is the appropriate measure of inflation. As mentioned in recital 150(b) of the Opening Decision, the CPN/CNV expressed concerns that the actual inflation of nuclear construction costs (based on the ABEX index) might be higher than the CPI inflation. Belgium argues that the 2 % ECB inflation target is the correct proxy of inflation for the Waste Cap calculation, for the following reasons:

(194)

For the reasons noted above, Belgium considers that the 3 % discount factor is an appropriate and conservative rate for the discounting of long-term liabilities.

(195)

As noted in recital 153 of the Opening Decision, Belgium submits that to the base amount of EUR 9 815 million a significant additional risk premium of EUR 5 185 million has been added to cover remaining uncertainties. As stated in recital 154 of the Opening Decision, the risk premium is based on a technical note by ONDRAF/NIRAS (104), analysing the uncertainties and risks associated with the transfer of financial responsibility for the management of radioactive waste and spent fuel from the seven Belgian nuclear power plants to the Belgian State (see footnote 80 of the Opening Decision). As stated in recital 156 of the Opening Decision, Belgium argues that the risk premium of 52,83 % is an adequate one and exceeds the risk premium of 35,47 % applied in the German case (SA.45296 (see footnote 101)).

(196)

In addition to the clarifications regarding the risk premium set out in recital 154 of the Opening Decision, Belgium has clarified in more detail during the formal investigation how the risk premium was computed based on the technical note by ONDRAF/NIRAS. Table 15 provides the detailed clarifications and composition of the risk premium, which led to amounts of EUR 5 033 million and EUR 5 133 million as lower and upper bound respectively. As part of the negotiations on the Waste Cap agreement, the risk premium has been set at a slightly higher amount of EUR 5 185 million, so as to arrive at a lumpsum payment of EUR 15 billion (see Table 13).

Table 15

Detailed risk premium composition and computation

(197)

Belgium submits that the ONDRAF/NIRAS opinion in its technical note provides a catalogue of risks associated with the transfer of nuclear liabilities to the Belgian State, which were added, with the assistance of ONDRAF/NIRAS experts, to the 2022 provision file to estimate what is the minimum risk premium required to cover all known risks. Therefore, Belgium argues that the risk premium in the Waste Cap agreement is meant to cover all the ‘less-likely-than-not’ risks identified in the ONDRAF/NIRAS note and additional risks identified during the negotiation that were derived from interactions with Electrabel and experts’ opinions. Finally, Belgium recalls that that experts from ONDRAF/NIRAS were involved in the negotiations of the Waste Cap agreement to assist the Belgian Government with regard to how to take into consideration their recommendations in their technical note of March 2023.

(198)

As mentioned in recital 163 of the Opening Decision, in contrast to the transferred nuclear waste and spent nuclear fuel liabilities, the nuclear operator remains solely responsible for the decommissioning and dismantling of the whole nuclear park (whereby the Contributing Companies contribute financially to the decommissioning and dismantling costs). The decommissioning and dismantling liabilities are thus ‘uncapped’.

(199)

However, as part of the transaction, the Belgian State took financial responsibility for the ‘additional decommissioning liabilities resulting from the LTO Project’ or ‘decommissioning dyssynergies’. The two main categories considered are the increases in post operational phase costs (operational and project (‘POP’) costs) and in decommissioning and dismantling (‘D&D’) costs caused by the LTO Project (operational, project and investments costs). Examples of what these decommissioning dyssynergies concretely consist of have been provided in footnotes 85 and 86 of the Opening Decision. The final amount of the ‘decommissioning dyssynergies’ is the result of the evaluation of the impact of the LTO Project and does not cover any operating costs relating to day-to-day management or usual activities. The final amount of the ‘decommissioning dyssynergies’ is the result of the evaluation of the impact of the LTO Project on different cost subcategories, since the shift of the decommissioning of two reactors impacts the planning, organisation, timing and need for certain investments at a given time (105). Neither BE-NUC nor the Belgian State bear any other liability for decommissioning liabilities.

(200)

As explained in recital 168 of the Opening Decision, since there was a disagreement between Electrabel and the Belgian State regarding the extent of the amount of decommissioning dyssynergies, the CPN/CNV was tasked to take the decision, and the final amount was not yet known at the time of the Opening Decision. On 24 June 2024, the CPN/CNV approved the amount of EUR [100-500] million (in nominal terms), and, on 16 October 2024, the CPN/CNV notified Belgium and Engie that the final amount to be paid amounts to EUR [100-500] million at the end of 2024 (106). The amount of decommissioning dyssynergies to be paid by the Belgian Government is a fixed lump sum that cannot be increased in case the additional costs are higher than foreseen; however, if the LTO Project does not proceed as expected, the Belgian State will be reimbursed.

(201)

Belgium submits that the waste deal also foresees an extensive security package to guarantee that ‘uncapped nuclear liabilities’ are borne by an economically viable nuclear operator:

(202)

Belgium submits that the waste deal can be assessed separately from the other parts of the measure.

(203)

First, Belgium submits that the waste deal concerns not only the LTO Units, but all seven nuclear reactors in Belgium. The nuclear provisions on which the Capped Amounts are based take into account all nuclear provisions to be made until the original legal end dates of the seven nuclear reactors (see Table 1).

(204)

Second, as mentioned in recital 170 of the Opening Decision, costs related to the operational waste and spent fuel produced by the LTO Units during the LTO Period will be paid by the co-owners of the LTO Units (BE-NUC and Luminus), which includes the costs for making the operational waste compliant with the CTC and a volume adjustment fee for these additional volumes. Electrabel and the Contributing Companies will establish provisions for radioactive operational waste in accordance with the applicable accounting rules. These additional costs are reflected in the Signing Financial Model on which the financial support measures of Component 1 of the LTO Project are based and are therefore taken into consideration regarding the calibration of the sub-measures in Component 1.

(205)

Third, the nuclear operator remains responsible for the decommissioning and dismantling of all the seven nuclear reactors in Belgium. These costs are neither reflected in the waste deal, nor in the Signing Financial Model. There is one exception related to the increase in decommissioning liabilities (for the LTO Units as well as the non-LTO Units) resulting from the LTO Project, for which a fixed lumpsum amount was established and approved by the CPN/CNV and that will be paid as a one-off by the Belgian authorities. The amount of these dyssynergies has been taken into account in the Signing Financial Model.

(206)

As noted in section 3.5 of the Opening Decision, the agreement between the Belgian State and Engie also includes provisions on legal protections, defining the risk-sharing in the event of certain future legislative changes. Chapter 4 of the Phoenix Law provides the legal basis to protect Engie against certain changes in the law. However, the relationship between the parties regarding indemnifications in case of the concerned changes in law will solely be governed by the provisions of the Implementation Agreement.

(207)

These provisions concluded with Engie and Electrabel provide that if the Belgian Federal Government or the Belgian Federal Parliament adopts new regulations specifically concerning nuclear operators in Belgium or Electrabel’s nuclear activities, which have a negative impact on the material terms of the transaction, the Belgian State will indemnify Engie (or one of the affected Engie group companies) for the direct losses it actually incurs as a consequence thereof. This also includes the payments Engie has to make to Luminus in the context of this indemnification. In accordance with Belgian law, the claimant must prove its claim and the amount of indemnification will be determined by a court or an arbitration procedure. Belgian courts are competent, but there is a reciprocal arbitration option for UNCITRAL arbitration.

(208)

This provision does not apply if the legal amendment results from the transposition of European or international law, unless the Belgian Federal Government or the Belgian Federal Parliament has induced or actively promoted the existence and the content of such legislation at another level (international, supranational, European regional, municipal, etc.) or has induced or actively promoted a judicial decision.

(209)

As stated in recital 180 of the Opening Decision, Belgium submits that the Commission’s decisional practice suggests that protections against change in law can constitute State aid. Belgium refers, in this respect, to the Commission decision regarding the lifetime extension of three other nuclear reactors in Belgium, in which the Commission examined the indemnification clauses contained in the agreements concluded between the Belgian State and the owners of nuclear power plants. As a consequence, Belgium concludes that the legal protections agreement between the Belgian State and Engie could imply the granting of a selective economic advantage to Engie.

(210)

Belgium submits that the lifetime extension of the two nuclear reactors requires a specific support package, because of the specific economic situation and the specific risk profile of nuclear energy.

(211)

As noted in recital 22 of the Opening Decision, no provisions for the LTO were made by Electrabel since, until March 2022, all the Belgian nuclear assets were planned to shut down by - at the latest - 2025, following the Nuclear Phase-Out law. Following the decision on the lifetime extension of Doel 4 and Tihange 3, there was an urgent need to refurbish the LTO Units. Given the tight time schedule and the increased costs of fuel and other necessary parts in recent years, the uncertainties regarding the investment costs of the LTO Project are considerable.

(212)

The Belgian authorities examined alternative support mechanisms, such as participation of nuclear technology in the Belgian capacity mechanism, a fixed feed-in premium, a one-way CfD, or a RAB model.

(213)

According to the Belgian authorities, given the market failures mentioned in section 3.1, those alternatives were considered as less appropriate to support the LTO Project, for the following reasons:

(214)

Belgium submits that the two-way CfD, in combination with the other sub-measures of the support package, provides the required support at a lower cost to consumers compared to alternative remuneration support mechanisms.

(215)

As noted in recital 181 of the Opening Decision, the ultimate beneficiaries of the notified measure are (i) Engie, as parent company of Electrabel, which is the sole operator and co-owner of the LTO Units (89,807 %) and as direct party to the Implementation Agreement concluded with the Belgian Government, and (ii) EDF S.A. (‘EDF’), as ultimate parent company of Luminus, which is co-owner of the LTO Units (10,193 %) and part of the Contributing Companies and, as the parent company of EDF Belgium, as part of the Contributing Companies. The Commission has not identified any reasons to change its assessment of the main beneficiaries during the formal investigation.

(216)

However, the Commission agrees with Belgium’s point of view regarding the identification of the beneficiaries of Component 1 of the measure. In particular, Belgium argued in its comments to the Opening Decision that BE-NUC (the JV between the Belgian State and Electrabel) and Luminus, rather than Electrabel and Luminus, are the direct beneficiaries of the CfD, SDC Loans, MOCP and WCF. Electrabel, as shareholder of BE-NUC and current co-owner of the LTO Units, can nevertheless be considered as an important beneficiary of these sub-measures of Component 1 as well. Regarding the other sub-measures under Component 1, which provide support for the operation and maintenance of the LTO Units, Electrabel, is a direct beneficiary as nuclear operator and shareholder in BE-NUC with the Belgian State, while Luminus is an indirect beneficiary as co-owner of the LTO Units. Under points 115 and 116 of the Commission’s Notice on the Notion of Aid, an advantage can also be conferred on undertakings other than those to which State resources are directly transferred (indirect advantage). Luminus benefits from the combination of sub-measures under Component 1 through the reduction of operational and insolvency risks they provide to the LTO Project, without itself being directly involved in these sub-measures. Luminus can therefore be considered to obtain an indirect advantage from these sub-measures. This indirect advantage does not apply to the EMSA, since Luminus is the owner of its share of the electricity produced by the LTO Units and will manage the sale of the electricity independently, nor to the ASA to be concluded between Electrabel and BE-NUC.

(217)

Regarding Component 2 of the notified measure, the transfer of nuclear waste and spent fuel liabilities and the agreement on decommissioning liabilities benefit the nuclear operator, Electrabel, as well as Luminus and EDF Belgium in their role as Contributing Companies, which are, together with the nuclear operator, financially responsible for the nuclear waste and decommissioning liabilities.

(218)

Regarding Component 3 of the notified measure, the legal protections provide that unilateral measures adopted by the Belgian State specifically affecting or applying to the operators of the nuclear units in Belgium and adversely modifying the material terms of the transaction would give rise to a right to indemnification. Therefore, the direct beneficiaries are Engie as mother company of Electrabel, the operator (and co-owner) of the nuclear reactors in Belgium, and BE-NUC. Similar as for some of the sub-measures under Component 1, although Luminus is not directly involved in the legal protections measures of Component 3, an indirect advantage is conferred on Luminus, who will also be indemnified in case of legal changes affecting the LTO Project (see section 3.3.3). Therefore, Luminus is an indirect beneficiary of Component 3 of the measure.

(219)

As explained in section 3.7 of the Opening Decision, the LTO Project requires a number of legislative changes, summarised in the present section.

(220)

The Nuclear Phase-Out law organises, since 2003, the gradual phase-out of electricity production through nuclear power in Belgium and has been amended already three times to allow for the lifetime extension of Tihange 1, Doel 1 and Doel 2.

(221)

The 10-year lifetime extension of Doel 4 and Tihange 3 requires another amendment of the Nuclear Phase-Out law. The modifications have been implemented by the ‘Law amending the Nuclear Phase-Out Law’, approved by the Parliament in plenary session on 18 April 2024, signed by the King on 26 April 2024, and published in the Belgian Official Gazette on 5 June 2024.

(222)

In accordance with Directive 2011/92/EU of the European Parliament and of the Council (107), Council Directive 92/43/EEC (108) and Directive 2009/147/EC of the European Parliament and of the Council (109), Belgium submits that an environmental impact assessment with cross-border consultations (110) is required and has been made.

(223)

The different elements of the LTO Project are implemented through a separate ‘Law to guarantee security of supply in the energy sector and reforming the nuclear energy sector’, also referred to as the ‘Phoenix Law’. The detailed description of the different chapters of the Phoenix Law has been provided in recital 191 of the Opening Decision. The Phoenix Law has been approved by the Parliament in plenary session on 18 April 2024 and signed by the King on 26 April 2024. It has been published in the Belgian Official Gazette on 5 June 2024.

(224)

As noted in recital 193 of the Opening Decision, the Belgian State established two new public entities that will take up certain responsibilities in relation to the LTO Project:

(225)

The BE-WATT Law and the Hedera Law have been approved by the Parliament in plenary session on 18 April 2024 and signed by the King on 26 April 2024. They have been published in the Belgian Official Gazette on 5 June 2024.

(226)

First, the Royal Decree Authorisations regulates the authorisation of installations that condition nuclear waste. This Royal Decree has been modified to allow the authorisation of conditioning installations that condition nuclear waste in accordance with the contractual transfer criteria. A separate category will be introduced for authorisations for conditioning installations that condition in accordance with contractual transfer criteria. It has been adopted on 11 July 2024 and published in the Belgian Official Gazette on 15 July 2024.

(227)

Second, the Royal Decree on Contractual Transfer Criteria determines the contractual transfer criteria and the categorisation of the waste package, as well as the way they consume the volume credit. It has been adopted on 11 July 2024 and published in the Belgian Official Gazette on 15 July 2024.

(228)

Belgium will finance the total funding requirement of the LTO Project via the general State budget, including the potential paybacks from the two-sided CfD.

(229)

As noted in recital 200 of the Opening Decision, Belgium estimates that the CAPEX costs of the LTO Project amount to EUR [2-2,5] billion, and the total operating costs over the lifetime to be EUR [7 000-8 000] million.

(230)

The net impact on the Belgian State budget is twofold: first, through the capital contribution of EUR 24,7 million to the JV, and second, through the net cost of measures payable by the RA Counterparty. Belgium submits that the expected budget spending depends not only on costs projections, but also on energy market price/revenues projections, because the LTO Project will be funded by a combination of market revenues, CfD difference payments, minimum OPEX and capital payments, and SDC Loans.

(231)

Under the base case projection of the Signing Financial Model, over the lifetime of the project, the RA Counterparty will receive a total net nominal amount of EUR [0-500] million.

(232)

However, if electricity prices were to evolve according to a lower projection, the cost of the measure to the Belgian Government would increase by EUR [4 000-4 500] million due to higher CfD payments. In case of an unexpected event whereby both nuclear plants would be unavailable, the RA Counterparty would also be exposed to providing additional support of EUR [500-1 000] million per year of unavailability. In a negative scenario of low electricity prices and a 12-month unavailability event, the RA Counterparty would have to provide a total net (nominal) amount of EUR [4 000-4 500] million over the project lifetime, i.e., the budget of the measure would be close to [40-50] % of total capital and operating costs.

(233)

Belgium confirms that the measure cannot be cumulated with other aid received to cover the same costs to be incurred under the LTO Project. Belgium submits that no other funds than the State budget are foreseen to be involved in the financing of the LTO Project and confirms that no Union funds will be involved.

(234)

The Commission adopted the Opening Decision on 22 July 2024. The Commission came to the preliminary view that the measure constitutes State aid and raised doubts regarding its compatibility with the internal market under Article 107(3), point (c), TFEU. In the Opening Decision, the Commission raised doubts regarding the necessity, appropriateness and proportionality of the aid, regarding the non-violation of Union law and regarding potential undue distortions of competition and trade.

(235)

First, regarding the necessity of the aid, the Commission raised doubts as to the financial support measures of Component 1. Although the Commission recognises that there might be a need for the nuclear operator and the owners of the nuclear reactors to have a stable source of revenues given the uncertainties related to the future electricity market price, the Commission questioned whether the additional financial support measures on top of the two-way CfD are all necessary. In particular, the Commission had doubts as to the creation of a JV to which the Belgian Government will be a shareholder, the minimum OPEX and capital payment and the SDC Loans.

(236)

Second, regarding the appropriateness of the aid, the Commission raised again doubts as to the financial support measures of Component 1. The Commission wanted to further inquire the design of the two-way CfD proposed by Belgium, since it seemed to lack appropriate incentives to react to market circumstances and to schedule maintenance in the most efficient way (including the use of the DAM price used as reference market price in the CfD design). In addition, the Commission questioned whether the combination of remuneration measures does not relieve the beneficiaries from too big a share of the market and operational risks.

(237)

Third, regarding the proportionality of the aid, the Commission raised doubts as to the proportionality of several of the financial remuneration measures (including the SDC Loans and the MOCP). The Commission considered that those measures are combined with a CfD and they are set to reach, by design, the target rate of return of 7 %, which can only be assessed with respect to the measures themselves and the risk-reduction they provide. Therefore, the Commission considered that it could not assess - at the time of the Opening Decision - the proportionality of this target rate of return in abstracto and concluded that the proportionality assessment can only follow the assessment of the appropriateness of the measures. In addition, the Commission had doubts as to the establishment of the amount of the lumpsum payment of EUR 15 billion for the transfer of the nuclear waste and spent fuel liabilities (in particular on the used discount rate and the height of the risk premium), as well as on the amount of the additional decommissioning liabilities resulting from the LTO Project (which was not yet known at the time of the Opening Decision).

(238)

Finally, the Commission questioned whether an inappropriate design of the CfD and the cumulation of all financial support measures could not lead to undue market distortions. In addition, the Commission needed further reassurance regarding the identity and independence of the entity selling the LTO Units’ output in the market under the EMSA.

(239)

Belgium sent its response to the Opening Decision on 22 August 2024. Belgium’s submission provided additional independent analyses prepared by Compass Lexecon on the CfD design, the SDC Loans and the MOCP (provided on 22 August), as well as on the Waste Cap (provided on 30 August 2024) in response to the Commission’s doubts raised in the Opening Decision regarding the necessity, appropriateness and proportionality of the package of financial support measures (Component 1 of the aid measure), regarding the proportionality of the waste deal (Component 2 of the aid measure), regarding the compliance of the CfD design with Union law and regarding potential undue distortions of competition and trade.

(240)

Belgium’s response to the Opening Decision discusses the initial measure (LTO Project) as described in the Opening Decision. The current measure includes changes (introduced to address the doubts raised by the Commission in its Opening Decision) compared to the initial measure and is set out in detail in section 3 of this Decision.

(241)

The arguments put forward by Belgium are outlined in more detail below.

(242)

Belgium agreed with the Commission that fostering market-conform incentives and minimising potential undue distortions are fundamental to the design of support mechanisms.

(243)

According to Belgium, supported by an independent additional economic analysis prepared by Compass Lexecon (111), the LTO Units’ technical, regulatory, and economic constraints (see also section 2.1) shape the feasibility and appropriateness of the CfD design. Belgium notably argued that providing economic incentives for the nuclear operator to react to market circumstances in situations where the ability to respond is restricted may allocate risks in a suboptimal way and increase the uncertainty of future revenues without any mitigation possibility. Hence, higher risks for the beneficiary that cannot be acted upon through operational levies may translate into a degraded risk-return profile for the investment and economic imbalance of the scheme (e.g. requiring a higher target IRR). The resulting expected impact on support costs needs therefore to be balanced with the costs for the RA counterparty of directly bearing that risk.

(244)

Belgium further argues that the merchant operation of the LTO Units and the corresponding bidding decisions are the economic benchmark for a CfD design that upholds market-conform behaviour.

(245)

In the case at hand, Belgium recalls that various technical and operating constraints strongly limit the leeway to modify the maintenance schedule and to modulate, and thereby the potential for value creation associated with increased economic incentives:

(246)

Belgium argued that, given the LTO Units’ operating and safety constraints, some alternative CfD designs features detailed below would not allocate risks and incentivise market-conform behaviour more efficiently than the proposed design. In particular, Belgium submitted that, in some cases, using long-term reference prices and reference quantity may foster efficient dispatch/operational decisions, as it decorrelates the CfD payments from the captured market price. However, this may not be appropriate in the case at hand given the LTO Units’ operational constraints and the unique availability pattern in the period of the LTO works, as explained below:

(247)

According to Belgium, the chosen CfD design in combination with its specific arrangements (the MPRA and the modulation arrangement), aimed to balance the market-risk exposure with operating incentives to support both investment and short-term market efficiency. It encourages timely responses to market signals without causing unnecessary complexity and is consistent with the technical and regulatory constraints of the LTO Units. Belgium referred in this respect to:

(248)

For the reasons outlined above, Belgium concluded that the Commission’s concerns regarding the appropriateness of the CfD design in the Opening Decision were unfounded.

(249)

However, during the formal investigation, Belgium decided to amend the CfD design and the MPRA mechanism with a view to reinforcing market-conform incentives and account for market evolutions.

(250)

Electrabel and the Belgian State envisage the creation of a 50/50 joint venture, which will hold the relevant assets enabling the continued operation of the LTO Units.

(251)

Belgium submitted that this approach enables the Belgian State to co-control the project company, as the two shareholders enter into the JV under equal terms and conditions and, as shareholders, with the same level of risk (corresponding to regular shareholder risk and funding) and rewards (in particular, dividends). It also enables the Belgian State to retain some degree of ownership over critical infrastructure.

(252)

Belgium explained that, if BE-NUC’s revenues are not sufficient to cover the costs required for the operation of the LTO Units (the O&M Agreement costs and other operating, fuel and maintenance CAPEX costs), then the Belgian State, as RA Counterparty, is required to make a shortfall payment to BE-NUC (and Luminus). BE-NUC has to, in that regard, submit an annual reconciliation report. This Minimum Opex payment is to ensure sufficient cash to meet these costs, in order to safeguard the long-term viability and solvency of BE-NUC. In addition, the RA Counterparty will provide a Capital Payment, which consists of a 50 % cost protection on the invested capital in relation to the amortised capital costs of the lifetime extension of the LTO Units. Both Minimum Opex and Capital Payment are referred to as MOCP and apply from the LTO Restart Date onwards.

(253)

Belgium explained that it will also make drawdown facilities available to both BE-NUC and Luminus with effect from 1 July 2025 in order to fund the operation and maintenance costs incurred during certain shutdown periods under clause 3.2(A) (‘Procurement of SDC Loans’) of the RA. Specifically, the Belgian State will provide SDC Loans at a capped interest rate and repayable according to a specified repayment schedule, rather than additional capital to BE-NUC and Luminus, to fund in two tranches:

(254)

Belgium explained that the MOCP addresses high impact risks resulting in losses during the operating phase. Along with the SDC Loans during the pre-Restart Phase, these measures are essential to ensure that the project generates sufficient cash flows to pay the nuclear operator and to maintain the long-term operational viability. Belgium argues that, without these measures, the nuclear power plant could face bankruptcy in the event of prolonged outages or regulatory (policy) changes that cause unavailability.

(255)

In order to further substantiate these claims, Belgium submitted the following observations, supported by an independent analysis prepared by Compass Lexecon (112):

(256)

Belgium concluded therefore that the MOCP and SDC Loans are appropriate and proportionate:

(257)

Following the formal investigation procedure, Belgium decided to introduce a cap on the potential MOCP payments equal to a maximum of EUR 2 billion by exercising its termination right(s) under the RA, except if an analysis by the Belgian State demonstrates that such termination could adversely affect Belgium’s security of supply and/or that a termination would not be appropriate from a financial perspective. See more details in section 3.3.1.3.3.

(258)

According to Belgium, the Belgian State and Electrabel, as shareholders, agree to fund their pro rata share of BE-NUC’s funding requirements or unexpected funding shortfalls (to the extent these concern costs and expenses that are not funded by the other mechanisms such as BE-NUC revenues, difference payments to BE-NUC under the RA, the SDC Loan, etc. ensuring that costs are not covered twice).

(259)

Belgium explained that Electrabel and the Belgian State entered into individual and identical agreements regarding the Shareholder Loans with BE-NUC. The interest rates have not been fixed but the Shareholder Loan Agreements provides that the interest rate of each Shareholder Loan will be an arm’s length rate determined by the board of BE-NUC in accordance with the Shareholders’ Agreement, by reference to prevailing market rates and any comparable third-party debt financing which may be available at the relevant time. The repayment terms of the amounts and interest rates are identical.

(260)

Belgium explained that the WCF serves to fund the need in working capital stemming from the operation of the LTO Units. BE-NUC will be allowed to draw down the WCF if the difference between its cash inflows and cash outflows is smaller than the estimated operational expenditures of the upcoming spending period set out in the RA. The amount of the WCF is to be at least the average aggregate estimated operational expenditure for a period of three months. Belgium clarified that, in effect, the WCF serves as an intra-year bridge to the annual MOCP, acting as a revolving credit facility that would be repaid yearly, if drawn down, by the MOCP provided by the Belgian State.

(261)

Belgium clarified that Electrabel will propose a methodology for the Electrabel Shareholder Loan, described in a term sheet, to be communicated to the board of BE-NUC and then replicated for the Belgian State loan, as well as for the WCF. The repayment terms of the amounts and interest rates are identical. Belgium submitted that the methodology was still under discussion and would be communicated at a later stage.

(262)

Regarding the purchase of energy manager services under the EMSA, Belgium submitted that the purchase of these services is carried out through a competitive, transparent, non-discriminatory, and unconditional tender procedure. In particular, the EMSA contract will be awarded through a negotiated procedure with a prior call for competition (Article 120 of the Belgian law on public procurement of 17 June 2016), which is a standard procedure in the utilities sector. Belgium therefore concludes that the EMSA does not cause an advantage for the energy manager (EMSA partner) and that the EMSA partner does not benefit from State aid.

(263)

Belgium submitted that, while the application of the public procurement directives already provides a number of guarantees, given the importance of the services tendered and the sensitivity from a competition perspective, additional provisions and safeguards have been implemented, to ensure that there will be sufficient competition in the tender.

(264)

More specifically, Belgium has explained that, prior to the drafting of the tender documents, an RFI on the EMSA was launched to gather views from the market, including a large number of subjects related to the content of the EMSA contract. In addition, the tender imposes specific selection criteria ensuring that the qualified/selected participants are fit for purpose in view of the importance and sensitiveness of the service tendered. More information on the tender procedure and the selection criteria is provided in section 3.3.1.5.1.

(265)

As noted in section 3.3.1.5.2, following the formal investigation procedure, Belgium decided to grant the decision-making authority regarding economic modulations to the EMSA partner and by replacing the EMSA partner’s fixed remuneration by a combination of a fixed and a variable fee, thereby providing stronger incentives to modulate according to market signals, while taking into account the technical constraints (and potential automatic shutdown risks) of the LTO Units.

(266)

Belgium submitted that specific measures are foreseen concerning the potential participation of GEMS, Engie’s trading entity (a business unit of the Engie group managerially independent from the business unit Nuclear), in the tender procedure.

(267)

Belgium further clarified that sufficient measures are, and will be taken, and implemented to effectively identify and prevent any potential conflicts of interest. These measures can be summarised as follows:

(268)

Belgium also clarified that, by way of ultimate fallback, if no EMSA has been concluded in due time, GEMS will temporarily perform the EMSA services in order to ensure continuity. Belgium argued that this arrangement is necessary and adequate to ensure the continuity of the public service (the sale of the electricity generated by the LTO Units), but this solution is strictly limited in time and to the minimum necessary.

(269)

Belgium concluded that the EMSA tender procedure excludes any potential State aid, since: (i) the Belgian State will ensure that BE-NUC will follow the public procurement legislation and principles rigorously, and (ii) the extensive consultation process before the actual launch of the tender provides additional safeguards ensuring that the purchase of services is carried out through a competitive, transparent, non-discriminatory, and unconditional tender procedure.

(270)

In addition, Belgium concluded that any risk for market foreclosure and other potential anticompetitive practices by Engie is avoided and that all measures and safeguards have been adopted to ensure that an independent EMSA partner will be appointed. Only as a fallback, GEMS would temporarily provide these services, which is necessary and adequate to ensure the continuity of the public service.

(271)

Belgium clarified that the LTO decommissioning and dismantling liabilities resulting from the LTO Project (‘dyssynergies’), if proven by Electrabel, will be borne by the Belgian State by the means of a one-shot (full and final) lumpsum payment on the closing date of the transaction. Belgium explained that the Implementation Agreement provides that the Belgian State and Electrabel should agree on the amount within a set timeframe, and if they fail to do so, the matter will be submitted to the CPN/CNV for decision.

(272)

Belgium recalled that, since the issue was pending at the level of the CPN/CNV at the time of the Opening Decision, its proportionality was not assessed in the Opening Decision.

(273)

Belgium explained that the Belgian State and Electrabel failed to agree on the amount within the set timeframe, so that the matter was submitted to the CPN/CNV for decision. More precisely, Belgium clarified that:

(274)

Belgium therefore concluded that the amount for the transfer of additional decommissioning liabilities resulting from the LTO Project as set out in the CPN/CNV in its Advice of 24 June 2024 is proportionate.

(275)

In addition, Belgium submitted, on 30 August 2024, an additional independent analysis by Compass Lexecon, presenting an economic analysis of the Waste Cap design, including an analysis of the 3 % discount rate and including further explanations on how the additional risks identified by the CPN/CNV are either already covered by the risk premium calculated by the Belgian State or of limited likelihood/impact.

(276)

Belgium submitted that the Commission’s concerns related to the discount rate originate from an analysis of the CPN/CNV in its advice of 7 March 2023 to the Minister of Energy (113), in which the CPN/CNV considers that the discount rate that should be used for spent fuel expenses has to gradually approach the long-term risk-free rate of return (calculated as 2,17 %), and suggests the use of a two-stage discount rate (3,17 % for the first 30 years, and 2,17 % after the first period of 30 years).

(277)

Belgium referred to an independent analysis prepared by Compass Lexecon (114), which demonstrated that the 3 % discount rate retained by the Belgian State is proportionate, based on: (i) an economic analysis of the CPN/CNV recommended approach, (ii) other relevant approaches, and (iii) the German precedent in case SA.45296.

(278)

Belgium submitted that the Commission’s concerns related to the risks considered in the Waste Cap agreement also originate from the CPN/CNV’s Advice to the Minister of Energy (see footnote 113), in which three risks should be considered:

(279)

The independent analysis by Compass Lexecon showed that these risks have been considered, and demonstrated that the approach taken by the Belgian State is prudent and includes sufficient contingencies to cover these risks.

(280)

This analysis has been completed by the response to question 5.4 of the RFI of 1 October 2024. Belgium submitted that, in practice, the portfolio of Hedera should be designed in order to face all the economic scenarios plausible for the period, notably in terms of the inflation level. Hedera’s target is to achieve a 1 % return in real term (i.e., 1 % above the inflation level). To this end, Hedera will develop a strategy that takes into account the risk that such return may not be achieved under certain economic regimes (for example, in case of an extended period of high inflation) with a portfolio composed only of bonds. Hedera will therefore focus on a dynamic asset-liabilities management, based on a diversified portfolio which includes assets that allow the fund to hedge against inflation. For example, Hedera could design a portfolio based on (but not limited to) bonds and debt instruments, equity, real estate, and derivatives, with the long-term goal of achieving a 1 % return in real term, while minimising the volatility risk. Finally, it should be noted that Hedera will limit its exposition to Belgian equity, bonds, and real estate. This is to ensure diversification and that the fund is sufficiently decoupled from the Belgium economy.

(281)

Belgium submitted that the different components and sub-measures of the LTO Project were part of the requests by Engie to agree on the lifetime extension and were all included in the global transaction documents signed on 13 December 2023. However, Belgium also observed that the components of the LTO Project differ in a number of elements and are complementary to each other, which limits any potential cumulative effects to arise.

(282)

Belgium referred to the case-law, which considers the following topics for the assessment of consecutive measures, including:

(283)

Belgium submitted that the subject matter and nature of the three components of the LTO Project are inherently different and noted that the Commission, in its Opening Decision, did not specifically consider the differences or similarities of the subject matter and nature of Components 1, 2 and 3. In this regard, Belgium submitted the following:

(284)

Belgium noted that the overall purpose of the three components is related to the lifetime extension of the LTO Units. However, Belgium submitted that the specific purpose of each component individually is a relevant element in the assessment:

(285)

Belgium also submitted that Components 1, 2 and 3 have different direct beneficiaries. Belgium acknowledged that the identification of beneficiaries of each measure is an important point in the assessment of the existence of cumulative effects. In this respect, the Belgian State submitted the following observations:

(286)

Components 1, 2 and 3 differ significantly in terms of subject matter and nature, purpose and beneficiaries, which have been entered into at the same time and serving the overall purpose of the lifetime extension of the LTO Units. As demonstrated above, each Component corresponds to a specific purpose rendering them complementary to one another, which limits any potential cumulative effects arising.

(287)

With respect to a potentially changed risk profile of Electrabel thanks to Components 2 and 3, these measures may indeed benefit Electrabel as a nuclear operator and consequently improve Electrabel’s risk profile overall. Components 2 and 3 do not alter the risk profile of BE-NUC, the main beneficiary of Component 1. As assessed in an independent analysis conducted by Compass Lexecon, the risk profile of BE-NUC given the specific risks affecting the project, and the risk allocation between BEGOV and BE-NUC, constitute the relevant factor to assess the proportionality of Component 1.

(288)

Engie submitted its comments to the Opening Decision on 9 September 2024. In its submission, Engie provides evidence and analysis in support of their argument that the doubts raised by the Commission in its Opening Decision would not be founded. Engie’s response to the Opening Decision discusses the initial measure as described in the Opening Decision. The current measure includes changes (introduced to address the doubts raised by the Commission in its Opening Decision) compared to the initial measure, and is set out in detail in section 3 of this Decision. The position of Engie is aligned with the position of the Belgian authorities.

(289)

Engie argued that the LTO Project is an overall agreement and the result of extensive and sustained negotiations with the Belgian Government, with as a guiding principle to remain committed to Belgium and to contributing to the country’s security of supply by accepting Belgium’s request to extend the lifetime of two nuclear reactors, thus overturning its public strategy to stop nuclear operations in Belgium after 2025 (see recital 23). In addition, Engie submitted that the LTO Project has positive effects on the internal market, for the following reasons:

(290)

Engie also argued that there are several differences between the LTO Project (lifetime extension of Doel 4 and Tihange 3) and the lifetime extension of Doel 1, Doel 2 and Tihange 1 in 2015 (‘existing LTO’):

(291)

Engie submitted that all these factors led Engie and the Belgian State to conclude a comprehensive agreement providing both for the requested lifetime extension (subject to a mitigation of Engie’s risk exposure) and the management of transferred spent fuel and nuclear waste and considered that the existing LTO is not a relevant reference to assess the necessity and appropriateness of the various measures of the present LTO Project.

(292)

The main arguments of Engie regarding the Commission’s doubts raised in the Opening Decision, are outlined in more detail below.

(293)

The Commission considers that the three components of the measure can be assessed as part of one single intervention from the Belgian State, insofar as they are all related to the same event, i.e., the lifetime extension of the LTO Units.

(294)

Engie submitted that it has no objection to the qualification as single intervention, insofar as it corresponds to the criteria laid down by the Court of Justice of the European Union’s case law (119). In particular, Engie agreed with the Commission’s view, in recital 208 of the Opening Decision, that the three components of the measure are interdependent and contribute altogether to the performance of the agreement on the lifetime extension, which allows them to be considered as one single intervention. Engie acknowledged that it had prerequisites regarding its risk exposure and the management of spent fuel and nuclear waste, which led to an overall and comprehensive agreement going beyond simply putting in place a remuneration mechanism for the extended operation of the nuclear units. However, without questioning this proposed qualification of single intervention, Engie pointed out that some of the sub-measures of the LTO Project do not constitute State aid since they do not confer a selective advantage on Electrabel, in particular:

(295)

Engie emphasised that the incentive effect, which the Commission deemed plausible in its assessment in recitals 229 and 239 of the Opening Decision, is not hypothetical, but that there is no doubt that Electrabel would not have continued operations of the LTO Units absent the LTO Project, including the three components. Engie submitted in this respect that:

(296)

In addition, Engie stressed that the incentive for Electrabel to continue operating the LTO Units stems from the whole package of support measures provided by the LTO Project, which has been adequately calibrated for this purpose. None of the sub-measures taken in isolation could have provided sufficient incentives. Likewise, removing any element of the package would have led Engie to refuse to proceed with the LTO Project. Engie concludes that, therefore, the measures referred to in recital 235 of the Opening Decision should not be regarded as individually providing a specific (or further) incentive but should be considered altogether.

(297)

Engie considered that the creation of a joint structure in which the Belgian State holds a 50 % stake is necessary to achieve the objectives of the LTO Project, i.e., the timely restart of the LTO Units by November 2025. Engie submitted that it requested a risk-sharing mechanism on a 50/50 basis with the Belgian State before adopting the agreement on the LTO Project.

(298)

According to Engie, having only the remuneration mechanisms in place was unfit for this purpose, for several reasons: (i) the RA leaves residual risks for Engie, (ii) without the JV, Engie would have had to temporarily bear 100 % of the losses pending the indemnification/compensation otherwise provided by the agreements, and (iii) Engie was in the process of limiting its exposure to nuclear energy and did not want to invest alone the required CAPEX of more than EUR [2-2,5] billion for the LTO Project, and support alone all potential losses due to unexpected unavailabilities and cost overruns.

(299)

In addition, Engie argued that, through the financial involvement of the Belgian State in the profits and losses derived from the LTO Units (in particular dividends), Engie anticipated that the Belgian State would be further incentivised to take account of the LTO Units in its future political decisions. Engie therefore concluded that the creation of the joint structure aligns the interests of the Belgian State and Engie.

(300)

Engie also submitted that the JV is appropriate and proportionate since Engie, a private operator, will be able to counterbalance any risk of the Belgian State acting in a partial manner through its joint control over the strategic decisions concerning BE-NUC, and the Belgian State and Engie will exercise their rights in the governance of BE-NUC on an equal footing.

(301)

Regarding the need for the CfD, Engie submitted, given the market failures related to the nuclear industry and the high volatility of the electricity market in the coming years, that the LTO Project is exposed to a strong risk of a funding gap, making the CfD a necessary instrument to achieve an objective of common interest. Engie also referred to the short duration of the lifetime extension (10 years) compared to the average investments in the nuclear sector, the significant costs, and the fact that the LTO Units will not run at 100 % capacity during the first three years, since simultaneously LTO extension works will be completed.

(302)

Engie submitted that the CfD design and the MPRA are appropriate and proportionate to preserve the incentives of the LTO Units to react efficiently to market signals. In this regard, Engie hereby to the same arguments of the Belgian State regarding the LTO Units’ technical, regulatory and economic constraints (see section 2.1). Engie argued that, even today, absent the CfD and with a fleet of five reactors allowing more diversification strategy, Engie has very limited flexibility to adjust nuclear electricity production in response to short- and mid-term market signals. According to Engie, within this limited room for manoeuvre, the CfD has been designed to fully incentivise BE-NUC to react to price signals, to the extent technically feasible.

(303)

Finally, regarding the parameters of the CfD design, Engie submitted that the DAM price is the appropriate MRP.

(304)

Engie emphasised that the need for the MOCP and the SDC Loans is primarily linked to the particular context of the LTO Project, as a consequences of: (i) the late decision to extend the lifetime of Doel 4 and Tihange 3 and the restricted timing to prepare and implement the necessary LTO works, (ii) the ageing assets and increased systemic risk by having two units of the same technology, (iii) the simultaneous restarting and operating of the LTO Units during the first three years, and (iv) the levelised cost of electricity of the nuclear extension which is more O&M-weighted than a nuclear greenfield asset.

(305)

Engie further submitted that both the MOCP and the SDC Loans are strictly necessary since: (i) the payments due by the Belgian State under the RA does not cover important unavailability events, caused by technical issues or due to international regulatory contingencies or nuclear emergencies; (ii) contrary to the statement in recital 269(b) of the Opening Decision, such unexpected events are not ‘extreme, low-probability’ scenarios, but are likely to occur in the next ten years; and (iii) only events above a certain threshold are covered.

(306)

Engie submitted that the MOCP is not an unlimited grant (contrary to what the Commission suggested in its Opening Decision) in light of the termination rights for the RA Counterparty as set out in the RA that would enable to limit its exposure.

(307)

Engie submitted that the principle of the extension of the SDC Loans to cover the portion of costs beyond the restart date has been granted by the Belgian State as a result of the advancing of the LTO restart date (from 2026 to 2025), at its request, implying that the LTO works must be undertaken while the units are in operation, which will lead to the generation of lesser revenues during the first three years of operation. It followed the key principle agreed between Engie and the Belgian State that the shareholders are exposed to any costs stemming from the LTO extension, while BE-NUC finances O&M costs with its own revenues.

(308)

Regarding the proportionality of the financial mechanisms provided by the LTO Project, Engie argued that they do not shelter BE-NUC from any operational risk, as suggested by the Commission in the Opening Decision. BE-NUC is partially exposed, amongst others, to availability risk throughout the lifetime extension of the LTO Units, as well as operating cost overruns after the True-Up Date, which may prevent its cash generation and negatively impact BE-NUC’s rate of return. Those costs pertain, amongst others, to increases in personnel costs (retention programmes and additional recruitment/training costs for the continued operations), O&M costs (increases in the price of material, shortages in the supply chain for essential spare parts and components, requirements from the ANFC/FANC (all project modifications are subject to extensive review and challenge by AFCN/FANC prior to approval), unplanned major equipment breakdowns before restart, etc.), and fuel costs (contracts are partially linked to market evolution).

(309)

In addition, Engie argued that Electrabel, as the operator of the LTO Units, has strong incentives to maximise the availability of the LTO Units under normal operation conditions to avoid paying liquidated damages if availability drops below [90-100] % in any contract year.

(310)

Engie submitted a simulation showing the impact on the LTO profitability of events that would not be covered by the MOCP: (i) reduced availability events during the Run Phase as an illustrative purpose (increased Forced Outage Rate (‘FOR’)) and (ii) increases in some operating costs after the True-Up Date as a result of unforeseen costs, as well as a combination of both events. Table 16 shows that these events could have a significant negative effect on BE-NUC’s IRR.

Table 16

Impact on the LTO profitability (100 %) of events not covered by the MOCP

(311)

Engie also argued that the interest rate on the SDC Loans is proportionate and appropriate as shown by an independent analysis (127). Engie added that the SDC Loans are intended to be repaid, subject to sufficient cash generation by BE-NUC, plus interests, at the end of the LTO period.

(312)

Engie clarified that the interest rate on the Shareholder Loans and the WCF will be an arm’s length rate determined by the board of BE-NUC, by reference to prevailing market rates and any comparable third-party debt financing which may be available at the relevant time.

(313)

Engie also clarified that it has prepared a term sheet describing the methodology followed to set the interest rates. Engie explained that this methodology is consistent with Engie’s transfer pricing loan policies and is in line with the OECD BEPS principle ensuring that the interest rate it set at an arm’s length level (see recital 63).

(314)

Engie therefore concluded that the interest rate of both the WCF and the Shareholder Loans are proportionate.

(315)

First, Engie emphasised that the entity of the Engie Group (‘GEMS’) that might be selected as the EMSA partner as a result of the tender procedure, or as a temporary fallback solution, will be fully independent from the business unit Nuclear. Engie clarified that, since January 2024, the appropriate tender procedures, on the one hand, and firewalls within the Engie Group, on the other hand, have been put in place to avoid any involvement of Electrabel in the selection of the EMSA partner.

(316)

Second, Engie submitted that there is no risk of market foreclosure arising from the potential selection of GEMS in the EMSA tender. Engie argued that GEMS will act as a party under a precise mandate specifying the BIS Strategy (from which it cannot derogate), remunerated through a fee, and the ownership of the electricity will remain with BE-NUC until its sale. Therefore, since the corresponding sales from BE-NUC could not be attributed to GEMS/the Engie Group, there could be no reinforcement of Electrabel’s position on the Belgian market should GEMS be appointed as the EMSA partner.

(317)

Third, the tender procedure to select the EMSA partner complies with the requirements of competitiveness, transparency, and non-discriminatory and unconditional character set out in the Commission’s practice. It has been published at the EU level and will specify the qualification criteria (particularly concerning the applicant's financial strength and operational experience). Therefore, if GEMS would be selected through the tender procedure, this appointment of GEMS as EMSA partner can be considered as proportionate. Engie also argued that, in case the tender procedure is not successful, and GEMS is appointed as the EMSA partner as a fallback solution, the proportionality criterion would nevertheless be met: GEMS will be remunerated on market conditions, at a price determined by an independent expert in the event of disagreement between the parties, so that this remuneration would not go beyond what is strictly necessary to achieve the purpose of the EMSA. In addition, Engie explained that this fallback solution would only be temporary, and the Belgian State would immediately restart the organisation of a new competitive tender.

(318)

Engie submitted that the aggregate amount of the Waste Cap is sufficient to cover any additional risks that would not have been considered in the calculation of the base amount. Engie further notes that the Waste Cap is set to transfer significantly less risks to the State than was the case in the German precedent validated by the Commission. In particular, Engie argued the following:

(319)

Engie therefore concluded that the amount paid by Electrabel under the waste deal will cover adequately the uncertainties taken over by BEGOV and will therefore be proportionate.

(320)

Regarding, the decommissioning liabilities, Engie submitted that Electrabel, as the nuclear operator of the seven Belgian nuclear units, will remain liable for its decommissioning obligations. Engie acknowledged however that Belgium will cover the (increase in) decommissioning costs directly resulting from the LTO Project, by way of a one-shot (full and final) lump sum payment to Electrabel.