OPINION OF ADVOCATE GENERAL

MEDINA

delivered on 18 January 2024(1)

Case C‑240/22 P

European Commission

v

Intel Corporation Inc.

(Appeal – Competition – Abuse of dominant position – Microprocessors market – Decision finding an infringement of Article 102 TFEU and Article 54 of the EEA Agreement – Loyalty rebates – Characterisation as abuse – As-efficient-competitor analysis – Overall strategy – Single and continuous infringement – Complex economic assessment – Extrapolation of economic data – Rebates awarded in the form of non-cash advantages)

I.      Introduction

1.        This Opinion concerns the appeal brought by the European Commission seeking to have the judgment of 26 January 2022, Intel Corporation v Commission (T‑286/09 RENV), (2) set aside. That judgment followed the judgment of the Court of Justice of 6 September 2017, Intel v Commission (C‑413/14 P), (3) which, itself, both annulled the judgment of 12 June 2014, Intel v Commission(T‑286/09) (4) and referred the case back to the General Court.

2.        By the judgment under appeal, the General Court concluded that Decision C(2009) 3726 final relating to a proceeding under Article [102 TFEU] and Article 54 of the EEA Agreement (5) should be annulled in part.

3.        More specifically, the General Court found that the Commission was not in a position to determine that the exclusivity rebates and payments granted by Intel to a number of original equipment manufacturers (‘OEMs’) and to a European distributor of desktop computers were capable of having or likely to have anticompetitive effects and that they therefore constituted an infringement of Article 102 TFEU. (6) According to the General Court, that was due to the errors vitiating the contested decision regarding, first, the assessment carried out by the Commission of the as-efficient-competitor test (‘AEC test’) and regarding, second, the assessment of the share of the market covered by Intel’s practice and its duration. (7)

4.        The Commission disputes the approach taken by the General Court and submits six grounds in support of its appeal. By those grounds of appeal, the Commission alleges, in essence, that the General Court failed to make an overall assessment of the capability of Intel’s practices to foreclose competition in the light of all the relevant circumstances, that it distorted evidence and made a number of errors of law in relation to the standard of review, misinterpreted the AEC test as applied in the contested decision, and infringed the Commission’s rights of defence.

5.        The Court has requested that two specific legal issues be examined, which concern the fourth and fifth grounds of appeal. Both grounds dispute the General Court’s assessment of the AEC test in relation to two of the OEMs benefiting from Intel’s practice. The present Opinion will accordingly focus on those grounds of appeal and, in particular, on the issues raised by them, relating, first, to the Commission’s margin of discretion when applying the AEC test to a specific conduct and, second, to the assessment of rebates awarded in the form of non-cash advantages.

II.    Facts and procedure

6.        For the purposes of the present Opinion, the facts and procedure of the proceedings may be summarised as follows. (8)

A.      Background to the dispute and administrative procedure

7.        Intel Corporation (‘Intel’) is a company based in the United States which designs, develops, manufactures, and markets central processing units (‘CPUs’), chipsets, and other semiconductor components, as well as platform solutions for data processing and communications devices.

8.        Following a formal complaint submitted on 18 October 2000 by Advanced Micro Devices, Inc. (‘AMD’), supplemented on 26 November 2003, the Commission launched investigations pursuant to Council Regulation (EC) No 1/2003. (9)

9.        On 26 July 2007, the Commission notified to Intel a statement of objections concerning its conduct vis-à-vis five major OEMs – namely Dell, Hewlett-Packard Company (‘HP’), Acer Inc. (‘Acer’), NEC Corp. (‘NEC’) and International Business Machines Corp. (‘IBM’).

10.      On 17 July 2008, the Commission issued to Intel a supplementary statement of objections concerning its conduct vis-à-vis MSH, a European distributor of microelectronic devices and the leading European desktop computer distributor. That statement of objections also covered Intel’s conduct vis-à-vis Lenovo Group Ltd (‘Lenovo’) and included new evidence relating to Intel’s conduct vis-à-vis certain of the OEMs mentioned above.

11.      After various procedural steps, the Commission adopted, on 13 May 2009, the contested decision, a summary of which appears in the Official Journal of the European Union (OJ 2009 C 227, p. 13).

B.      The contested decision

12.      According to the contested decision, Intel committed a single and continuous infringement of Article 102 TFEU and of Article 54 of the Agreement on the European Economic Area (EEA), from October 2002 until December 2007, by implementing a strategy aimed at foreclosing a competitor, AMD, from the market for x86 CPU microprocessors (‘x86 CPUs’).

1.      Relevant market

13.      The goods at issue in the contested decision are CPUs – namely x86 CPUs – which are key components of any computer, both in terms of overall performance and cost of the system. CPUs are often referred to as a computer’s ‘brain’ and the process for their manufacture requires expensive high-tech facilities. Prior to 2000, a number of manufacturers of x86 CPUs were present on the market. However, most of those manufacturers have since exited the market. The contested decision states that Intel and AMD have been essentially the only two companies still manufacturing x86 CPUs.

14.      The geographical market was defined as worldwide.

2.      Dominance

15.      On the basis, first, of a market share in excess of or around 70% held by Intel between 1997 and 2007 and, second, the significant barriers to entry and expansion in the relevant market – resulting from sunk investments in research and development, intellectual property and the necessary production facilities – the Commission concluded that, at least in the period covered by the contested decision, that is to say, from October 2002 to December 2007, Intel held a dominant position in that market.

3.      Types of conduct

16.      The contested decision described two types of Intel conduct vis-à-vis its trading partners, namely conditional rebates and naked restrictions.

17.      In the first place, according to the contested decision, Intel granted four OEMs – namely Dell, Lenovo, HP and NEC – rebates which were conditioned on those OEMs purchasing all or almost all of their x86 CPUs from Intel. Similarly, Intel awarded payments to MSH, which were conditioned on MSH selling exclusively computers containing Intel’s x86 CPUs.

18.      The contested decision concluded that the conditional rebates granted by Intel to the OEMs constituted fidelity rebates. With regard to MSH, the contested decision established that the economic mechanism of Intel’s conditional payments to these payments was equivalent to that of the conditional rebates granted to the OEMs.

19.      The contested decision also conducted an economic analysis of the capability of the rebates to foreclose a hypothetical competitor as efficient as Intel, albeit not dominant. More precisely, the analysis established at what price a competitor as efficient as Intel would have had to offer CPUs in order to compensate an OEM for the loss of an Intel rebate. The same kind of analysis was conducted for the Intel payments to MSH.

20.      The evidence gathered by the Commission led it to the conclusion that Intel’s conditional rebates and payments induced the loyalty of the key OEMs and of MSH. The effects of those practices were complementary, in that they significantly diminished competitors’ ability to compete on the merits of their x86 CPUs. Intel’s anticompetitive conduct thereby resulted in a reduction of consumer choice and in lower incentives to innovate.

21.      In the second place, with regard to naked restrictions, the Commission stated that Intel awarded three OEMs – namely HP, Acer and Lenovo – payments which were conditioned on those OEMs postponing or cancelling the launch of AMD CPU-based products and/or placing restrictions on the distribution of those products. The contested decision concluded that Intel’s conduct also directly harmed competition, and did not constitute normal competition on the merits.

4.      Abuse and fine

22.      The Commission concluded in the contested decision that, in each instance, Intel’s conduct vis-à-vis the OEMs and MSH constituted an abuse under Article 102 TFEU, but that each of those individual abuses was also part of a single strategy aimed at foreclosing AMD, Intel’s only significant competitor, from the market for x86 CPUs. Those individual abuses were therefore part of a single and continuous infringement of Article 102 TFEU, committed from October 2002 until December 2007. (10)

23.      By applying the Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation No 1/2003 (OJ 2006 C 210, p. 2), the Commission imposed on Intel a fine of EUR 1.06 billion. (11)

C.      The initial judgment

24.      On 22 July 2009, Intel brought an action for annulment of the contested decision. The Association for Competitive Technology (‘ACT’) was granted leave to intervene in those proceedings in support of Intel.

25.      By the initial judgment, delivered on 12 June 2014, the General Court dismissed the action in its entirety.

26.      In its judgment, the General Court held, in essence, that the rebates granted to the OEMs were exclusivity rebates, since they were conditional upon customers purchasing from Intel either all their x86 CPU requirement or most of their requirements. In addition, the General Court explained that the question whether such rebates could be characterised as abusive did not depend on an analysis of the circumstances of the case aimed at establishing the capacity of those rebates to restrict competition, nor on the demonstration of potential anticompetitive effects by means of an AEC test.

27.      For the sake of completeness, the General Court considered, in the context of an examination carried out in the alternative, that the Commission had demonstrated, to the requisite legal standard and on the basis of an analysis of the circumstances of the case, that the exclusivity rebates and payments granted by Intel to Dell, HP, NEC, Lenovo and MSH were capable of restricting competition. By contrast, the General Court deemed unnecessary an examination of whether the Commission had carried out the AEC test in accordance with the applicable rules and without committing any errors.

D.      The judgment on the first appeal

28.      On 26 August 2014, Intel appealed against the initial judgment.

29.      By the judgment on the first appeal, delivered on 6 September 2017, (12) the Court of Justice set aside the initial judgment and referred the case back to the General Court. (13)

30.      In particular, after rejecting the fifth and fourth grounds of appeal, alleging, respectively, the misapplication by the General Court of the criteria relating to the Commission’s jurisdiction with regard to the agreements concluded between Intel and Lenovo and a procedural irregularity, (14) the Court examined and upheld the first ground of appeal, which alleged an error of law on account of the failure to examine the rebates at issue in the light of all the relevant circumstances.

31.      In that regard, the Court recalled, first of all, that it is in no way the purpose of Article 102 TFEU to prevent an undertaking from acquiring, on its own merits, the dominant position on a market. Nor does that provision seek to ensure that competitors less efficient than the undertaking with the dominant position should remain on the market. (15) However, given the special responsibility of a dominant undertaking not to allow its behaviour to impair genuine, undistorted competition on the internal market, it cannot, among other things, adopt pricing practices that have an exclusionary effect on competitors considered to be as efficient as it is itself. (16)

32.      On the basis of those considerations, the Court held that the principles laid down in the judgment of 13 February 1979, Hoffmann-La Roche v Commission, (17) had to be clarified, where the undertaking concerned submits, during the administrative procedure, on the basis of supporting evidence, that its conduct was not capable of restricting competition and, in particular, of producing the alleged foreclosure effects. (18) In that case the Commission is not only required to analyse, first, the extent of the undertaking’s dominant position on the relevant market and, second, the share of the market covered by the challenged practice, as well as the conditions and arrangements for granting the rebates in question, their duration and their amount; it is also required to assess the possible existence of a strategy aiming to exclude competitors that are at least as efficient as the dominant undertaking from the market. (19)

33.      Moreover, the Court found that, if, in a decision finding a rebate scheme abusive, the Commission carries out an AEC analysis, the General Court must examine all of the applicant’s arguments seeking to call into question the validity of the Commission’s findings concerning the foreclosure capability of the rebate concerned. (20)

34.      As to the contested decision, the Court observed that the AEC test played an important role in the Commission’s assessment of whether the rebate scheme at issue was capable of having foreclosure effects on as efficient competitors. (21) In those circumstances, the General Court was required, according to the Court, to examine all of Intel’s arguments concerning that test. (22) Since the General Court held that it was not necessary to examine the question whether the alternative calculations proposed by Intel had been carried out correctly, (23) it wrongly failed to take into consideration Intel’s line of argument seeking to expose alleged errors committed by the Commission in the AEC test. (24)

E.      The judgment under appeal

35.      Following the referral back, the case was allocated to the Fourth Chamber (Extended Composition) of the General Court.

36.      By the judgment under appeal, delivered on 26 January 2022, the General Court held that the contested decision had to be annulled in part.

37.      As a preliminary remark, the General Court referred to the subject matter of the proceedings following the referral back. (25) More specifically, the General Court concluded that the subject matter of the dispute concerned, in essence, the analysis of the capacity of the rebates at issue to restrict competition in the light of (i) the clarifications provided regarding the principles laid down in the judgment in Hoffmann-La Roche and (ii) the parties’ main and supplementary observations on the conclusions to be drawn from those clarifications. (26)

38.      Moreover, the General Court accepted the findings of the initial judgment on the legal characterisation of the naked restrictions and their unlawfulness under Article 102 TFEU. (27) It also accepted the assessment concerning the rebates at issue being characterised as ‘exclusivity rebates’. (28) The General Court held, however, that, in accordance with the judgment of the Court of Justice on the first appeal, that characterisation did not mean that it was not necessary to apply the AEC test for analysing their capacity to restrict competition. That characterisation was not either sufficient for those rebates to be considered as abusive under Article 102 TFEU. (29)

39.      Regarding the substance of the case, after recalling the method defined by the Court of Justice for assessing whether a rebate scheme has the capacity to restrict competition (30) and the principles arising from the judgment on the first appeal, (31) the General Court assessed the arguments raised by Intel and by ACT.

1.      Arguments concerning the legal analysis relied on by the Commission

40.      In the first place, the General Court examined the arguments put forward by Intel and ACT that the contested decision was based on an incorrect legal analysis. In that regard, the General Court concluded that, in taking as a starting point the premiss that the rebates at issue infringed Article 102 TFEU on the ground that they were by their very nature abusive, without necessarily having to take into account the capability of those rebates to restrict competition, the Commission vitiated the contested decision by an error of law. (32) However, given the important role played by the AEC test in the Commission’s assessment of whether the rebates at issue were capable of having foreclosure effects on as-efficient competitors, the General Court considered itself required, as a second step, to examine all of Intel’s arguments concerning that test. (33)

2.      Arguments concerning the errors vitiating the AEC analysis

41.      In the second place, the General Court devoted its analysis to the arguments concerning the AEC test. That analysis was divided into four sections.

42.      The first section concerned the scope of the General Court’s review, (34) which it considered to extend, in accordance with the Court’s case-law, to all the elements of the Commission’s decision, in law and in fact, in the light of the pleas raised by the applicant and taking into account all the relevant evidence submitted by the latter. The General Court recalled, however, that in reviewing the lawfulness of the act at issue, it could not substitute its own reasoning for that of the Commission.

43.      The second section contained general considerations concerning the AEC test, and was worded as follows: (35)

‘152.      The starting point for the AEC test … is that, given, in particular, the nature of its product, its brand image and its profile, Intel was an unavoidable trading partner and that the OEMs would always have purchased at least part of their CPU requirements from Intel, regardless of the quality of products offered by alternative suppliers. Therefore, customers were willing and able to switch only a certain share of their demand to that alternative supplier (‘the contestable share’). Intel’s power to use the non-contestable share as leverage to reduce the price on the contestable share of the market was the result of that status as an unavoidable trading partner.

153.      As the [General] Court observed in paragraph 141 of the [initial judgment], the AEC test carried out in the contested decision starts from the premiss that an as-efficient competitor, which seeks to obtain the contestable share of the orders hitherto satisfied by a dominant undertaking, must compensate the customer for the exclusivity rebate which it would lose if it purchased a smaller portion than that stipulated by the exclusivity or quasi-exclusivity condition. The AEC test is designed to determine whether the competitor which is as efficient as the undertaking in a dominant position, which faces the same costs as the latter, can still cover its costs in that case.

154.      The AEC test, as applied in the present case, establishes the price at which a competitor as efficient as Intel would have had to offer its x86 CPUs in order to compensate an OEM for the loss of any exclusivity payment granted by Intel. In the AEC test, that price is called the “effective price” or “EP”.

155.      In principle, the portion of the total rebates for which an as-efficient competitor must offer compensation comprises only the amount of the rebates which is subject to the exclusive supply condition, excluding quantity rebates (“the conditional portion of the rebates”). … In order to take into consideration only the conditional part of a payment, the AEC test refers, in the present case, to the average sales price (“the ASP”), namely the list price, after deduction of conditional discounts.

156.      The smaller the contestable share and, therefore, the smaller the quantity of products for which the alternative supplier may compete, the greater the likelihood that the exclusivity payment will be capable of foreclosing an as-efficient competitor. If the loss of payments granted by Intel to its customer must be spread over a small quantity of products offered by the alternative supplier in the contestable share, that leads to an appreciable reduction in the effective price. It is therefore more probable that the effective price will be below Intel’s measure of viable cost.

157.      The effective price must be compared with Intel’s measure of viable cost. Intel’s measure of viable cost adopted in the contested decision is that of the average avoidable cost (“AAC”).

158.      … It may be concluded that a system of exclusivity payments is capable of foreclosing market access for equally efficient competitors if the effective price is below Intel’s AAC. In that situation, the AEC test has a negative result. If, by contrast, the effective price is above the AAC, an as-efficient competitor is deemed to be able to cover its costs and to therefore be in a position to enter the market. In that situation, the AEC test has a positive result.

159.      It is in the light of those general considerations that the [General] Court must examine the substance of the applicant’s arguments that the AEC analysis is vitiated by numerous errors.’

44.      It is apparent, in particular, from the judgment under appeal (36) that, pursuant to the methodology adopted by the Commission, the result – positive or negative – of the AEC test, as defined in paragraph 158 of that judgment, is ultimately determined by means of a comparison between the contestable share and the required share, the latter being the proportion of the customer’s requirements which an as-efficient competitor must obtain in order for it to be able to enter the market without incurring losses. If the contestable share is higher than the required share, the AEC test result is positive for Intel. The opposite situation indicates a negative result and that the contested rebates are capable of foreclosing a competitor as efficient as Intel.

45.      The third section dealt with the burden of proof and the standard of proof required. (37)

46.      In the fourth section, the General Court examined the merits of Intel’s arguments that the decision at issue was vitiated by numerous errors as regards the AEC test. (38) That section contains five subsections, each of which deals with the arguments raised by Intel concerning the AEC analysis contained in the contested decision as regards the four OEMs concerned, namely Dell, HP, NEC and Lenovo, on the one hand, and MSH, on the other. In the light of its analysis, the General Court accepted Intel’s argument that the AEC analysis carried out by the Commission in the contested decision was vitiated by errors. (39)

3.      Arguments concerning the criteria referred to in paragraph 139 of the judgment on the first appeal

47.      In the third place, the General Court examined Intel’s and ACT’s arguments that the Commission had not properly analysed the criteria mentioned in paragraph 139 of the judgment on the first appeal. (40)

48.      In that regard, the General Court held that Intel was justified in claiming that the analysis carried out in the contested decision was vitiated by a number of errors, in that the Commission did not consider properly the criterion relating to the share of the market covered by the contested practice and did not analyse correctly the duration of the rebates. (41)

4.      Conclusion

49.      The General Court found that, in the light of the foregoing considerations, and given the errors vitiating the contested decision as regards, first, the assessment of the as-efficient-competitor test carried out by the Commission and as regards, second, the assessment of the share of the market covered by Intel’s practice and its duration, (42) the Commission was not in a position to determine that Intel’s exclusivity rebates were capable of having or being likely to have anticompetitive effects and that they therefore constituted an infringement of Article 102 TFEU. (43)

50.      Therefore, the General Court concluded that the grounds of the contested decision were not capable of serving as a basis for Article 1(a) to (e) of that decision as regards in particular Intel’s exclusivity rebates. That article was consequently annulled. (44) In addition, since the General Court did not consider itself in a position to identify the amount of the fine relating solely to the naked restrictions, which were deemed to be correctly established in the initial judgment, (45) Article 2 of the contested decision was also annulled. (46) Finally, Article 3 of the contested decision was annulled in so far as it concerned Intel’s exclusivity rebates. The General Court dismissed the action as to the remainder. (47)

III. Forms of order sought by the parties before the Court

51.      By its appeal, lodged with the Court on 5 April 2022, the Commission claims that the Court should:

–        set aside the judgment under appeal, except point 3 of its operative part;

–        refer the case back to the General Court;

–        reserve the costs.

52.      By decision of the President of the Court of Justice of 5 August 2022, the Federal Republic of Germany was granted leave to intervene in the present case in support of the forms of order sought by the Commission.

53.      Intel and ACT contend that the Court should dismiss the appeal and order the Commission to pay the costs.

54.      No oral hearing has been held in the present case.

IV.    Assessment

55.      In support of its appeal, the Commission relies on six grounds of appeal. In particular, the Commission alleges that:

–        the General Court ruled ultra petita, misapplied the judgment on the first appeal and failed to make any overall assessment of the capability of Intel’s practices to foreclose competition (first ground of appeal);

–        the General Court’s review of the AEC test infringed the Commission’s rights of defence (second ground of appeal);

–        the General Court erred with respect to the standard of review, infringed the Commission’s rights of defence and distorted the evidence in the context of the examination of the AEC test with regard to Dell (third ground of appeal);

–        the General Court erred in law and infringed the Commission’s rights of defence in the context of the examination of the AEC test with regard to HP (fourth ground of appeal);

–        the General Court erred in its interpretation of the AEC test and of Article 102 TFEU, distorted the evidence and infringed the Commission’s rights of defence in the context of the examination of that test with regard to Lenovo (fifth ground of appeal);

–        in so far as the judgment under appeal relies on the review of the AEC test for the purposes of annulling in part the contested decision, the General Court failed to consider properly the implications of its findings (sixth ground of appeal).

56.      In line with the request from the Court, my analysis will be devoted to the fourth and fifth grounds of appeal.

A.      The fourth ground of appeal, alleging a number of errors of law and the infringement of the Commission’s rights of defence in the context of the examination of the AEC test with regard to HP

57.      By its fourth ground of appeal, the Commission submits that the General Court erred in law in finding that the contested decision had failed to demonstrate the foreclosure effect of the rebates granted by Intel to HP for the entire infringement period.

58.      That ground of appeal is divided into four main parts, which allege, first, failure to have due regard to the Commission’s margin of discretion in complex economic matters; second, failure to take into account Intel’s implicit acknowledgment of the reference period during the administrative proceedings; third, infringement of the Commission’s right of defence; and, fourth, an error by the General Court in relation to the appropriate conclusion to be drawn in respect of the entire period of the practice at issue. (48)

59.      After recalling the findings of the General Court as regards the rebates granted by Intel to HP, I shall examine each of those arguments in turn.

1.      The findings of the General Court as regards the rebates granted by Intel to HP 

60.      It is apparent from the judgment under appeal that, according to the contested decision, Intel had concluded with HP, for the period between November 2002 and May 2005, two agreements concerning corporate desktop computers. (49)

61.      The first of those agreements (‘the HPA1 agreement’) covered a period from November 2002 to May 2004, and the second (‘the HPA2 agreement’) covered a period from June 2004 to May 2005. In both agreements, the grant of rebates by Intel was subject to the unwritten condition that HP would source from the former at least 95% of its x86 CPU requirements in order to equip its computers. According to the Commission, the AEC test demonstrated that those rebates were capable of having an anticompetitive foreclosure effect. (50)

62.      That conclusion is based on a comparison between HP’s required share and the contestable share (51) and on two reinforcing factors. (52)

63.      Regarding the calculation of HP’s required share, which is at the core of the present ground of appeal, the judgment under appeal refers, first, to Table 34 of the contested decision (‘Table 34’). (53) As the General Court indicates, that table sets out the parameters and the concrete figures used by the Commission to calculate HP’s required share, which it presents in eight rows, corresponding to each of the quarters covered from the fourth quarter of the 2003 tax year to the third quarter of the 2005 tax year. (54)

64.      Second, the judgment under appeal refers to Table 35 of the contested decision (‘Table 35’), (55) which presents the Commission’s global calculation of HP’s required share as regards the HPA1 agreement and the HPA2 agreement. As the General Court observes, that global calculation is the result of the arithmetic sum or average of the figures established in Table 34. (56) The judgment under appeal also notes that the Commission found that HP’s required share was consistently above the contestable share throughout the entirety of the infringement period. (57)

65.      Furthermore, it is apparent from the judgment under appeal that Intel claimed before the General Court that the contested decision contained a number of errors concerning, inter alia, the infringement period which was examined. (58)

66.      In that regard, the General Court pointed out that Table 34 did not incorporate any data relating to the initial period covered by the HPA1 agreement, namely to the months of November and December 2002, nor to the first three quarters of the 2003 tax year. (59) However, the global calculation of HP’s required share as regards the HPA1 agreement, as it appeared in Table 35, had been calculated by the Commission from the arithmetical sum or average of the figures appearing in Table 34, in particular from the first three lines thereof – referred to as Q4 FY03, Q1 FY04 and Q2 FY04. (60)

67.      Consequently, the General Court found that the months of November and December 2002 and the first three quarters of the 2003 tax year had not been taken into account by the Commission in the calculations leading to HP’s required share as regards the HPA1 agreement. (61) In that respect, it added, in essence, that the Commission had not maintained that the absence of the values for the missing three quarters in its calculations was the product of a coincidence, or that those values were identical to the values of the missing three quarters of the period covered by that agreement. (62)

68.      It is in those circumstances that the General Court concluded that the calculation of the required share did not cover the entire period between November 2002 and May 2005, in relation to which the Commission took the view that it could demonstrate that there was a foreclosure effect caused by the rebates which Intel granted to HP. (63)

69.      Moreover, the General Court dismissed the Commission’s arguments based on the fact that the result of a quarterly calculation did not differ fundamentally from the result of the overall calculation which was alleged to have been made. (64) It further considered that the additional calculations presented by the Commission in its rejoinder – namely in Annex D.17 thereto – were inadmissible and, in any event, could not serve as the basis for the findings of the contested decision. (65)

2.      The failure to have due regard to the Commission’s margin of discretion in complex economic matters

(a)    Arguments of the parties

70.      In the first place, the Commission, supported by the Federal Republic of Germany, argues that the judgment under appeal ignores the nature of the AEC test as a complex economic assessment and the margin of discretion which that institution is acknowledged to have available to it in such cases. That margin of discretion must necessarily encompass the calculation of an undertaking’s required share for the purposes of applying the AEC test and, in particular, the selection of the economic parameters and the reference period that the Commission uses for that calculation, which should be subject only to limited judicial review.

71.      Furthermore, the Commission criticises the General Court’s assessment that it had relied on incomplete figures to calculate HP’s required share. The Commission submits that it was justified in relying on the figures corresponding to the last three quarters covered by the HPA1 agreement because they were believed to be sufficiently representative for the entire period since the amount of the rebates during that period was stable from one quarter to the other.

72.      Finally, the Commission argues that the figures used to calculate the required share for the period covered by the HPA1 agreement, as found in Table 35 of the contested decision, were favourable to Intel. Had the contested decision used the available data for previous quarters, the required share for that period would have been even higher, indicating that Intel’s rebates were even more capable of foreclosing competition.

73.      Intel, supported by ACT, disputes those arguments. It submits that the Commission chose to perform a quarter-by-quarter AEC assessment in order to establish the infringement by Intel as regards HP and that the Commission then failed to incorporate into its calculations all the necessary data to demonstrate the foreclosure effects of the entire period covered by the HPA1 agreement. That failure would not be part of a complex economic matter, but rather of an assessment of facts. Moreover, Intel disputes that the averaged figures from the later quarters were sufficiently representative to find an infringement for the entire period. In that regard, it notes that that assessment simply does not result from the contested decision and that the additional calculations presented by the Commission during the judicial proceedings were declared inadmissible.

(b)    Analysis

74.      As a preliminary remark, I would point out that the argument put forward by the Commission, as regards the nature of the AEC test as an economic complex assessment, applies horizontally to other grounds of appeal submitted by it in the present appeal. Some of my reflections might therefore also be applicable to the analysis to be carried out by the Court regarding those other grounds of appeal which are not specifically covered by the present Opinion.

75.      According to well-established case-law, the level of review carried out by the EU Courts of the analyses made by the Commission on the basis of the competition rules of the Treaty must take into account the margin of discretion which underlies each decision under consideration and which is justified by the complexity of the application of those rules. Cases involving a limited review must be restricted to those in which the contested decision is based on a complex economic assessment. (66)

76.      That case-law, which sets out the standard of limited review in complex economic assessments, has been applied by the EU Courts to all areas of competition law (67) since its initial formulation in the judgment in Consten and Grundig v Commission. (68) It requires that, when facing complex economic matters, EU Courts must limit their review to verifying whether the relevant procedural rules have been complied with, whether the statement of the reasons for the decision is adequate, whether the facts have been accurately stated and whether there has been any manifest error of appraisal or a misuse of powers. (69)

77.      Nevertheless, the Court has consistently recalled that the margin of discretion recognised to the Commission when dealing with complex economic assessments, in particular in the field of competition law, is necessarily accompanied by certain guarantees which the EU Courts are obliged to verify. (70)

78.      Accordingly, in its judgment in Commission v Tetra Laval, (71) the Court established that, whilst the Court recognises that the Commission has a margin of discretion with regard to economic matters, that does not mean that the EU Courts must refrain from reviewing the Commission’s interpretation of information of an economic nature. Not only must the EU Courts, inter alia, establish whether the evidence relied on is factually accurate, reliable and consistent, they must also establish whether that evidence contains all the information which must be taken into account in order to assess a complex situation and whether that evidence is capable of substantiating the conclusions drawn from it. (72)

79.      Later, in the judgments in KME Germany and Others v Commission (73) and Chalkor v Commission, (74) which stem from the previous statement in the judgment in Tetra Laval, the Court emphasised that the EU Courts cannot use the Commission’s margin of discretion as a basis for dispensing with the conduct of an in-depth review of the law and of the facts. (75)

80.      In the present case, the Commission makes a general criticism of the standard of review applied by the General Court in the judgment under appeal. The Commission considers that the General Court, when examining the AEC test applied in the contested decision – including as regards the rebates granted by Intel to HP – exceeded the limits applicable to the judicial review set out in its case-law.

81.      It is important to note that the issue raised by the Commission is not whether that institution should benefit from a margin of discretion as regards the choice of the specific test for determining whether pricing practices adopted by an undertaking in a dominant position have the capacity to foreclose a competitor as efficient as that undertaking. (76) Neither the General Court in the judgment under appeal, nor the defendant in the present appeal, question the use of the AEC test by the Commission for assessing whether the rebates covered by the contested decision were capable of undermining competition under Article 102 TFEU.

82.      Moreover, contrary to the arguments of the Commission, it is not disputed that that institution should benefit of a margin of discretion with respect to the application of the AEC test to a specific conduct, in particular, in terms of the selection of the economic parameters and the reference period which must be used in order to apply that test to specific conduct.

83.      Indeed, it is apparent from the judgment under appeal, (77) without any challenge by the General Court, that the AEC test carried out by the Commission in the contested decision, with respect to Intel’s rebates to HP, was based on an econometric model that the Commission decided to apply on a quarterly basis within a period that it had previously defined from November 2002 to May 2005. That model required, on the one hand, the calculation of the required share and, on the other hand, the calculation of the contestable share. In particular, in order to calculate the required share, which is at issue in the present part of the ground of appeal, account was taken of the conditional portion of the rebates granted by the undertaking in a dominant position, the average sales price and the average avoidable cost of that undertaking. Ultimately, the outcome of the AEC test, whether positive or negative, was determined by comparing the figures obtained from the calculation of HP’s required share and the contestable share of the market. (78)

84.      I would point out, in line with the Commission’s arguments, that, inasmuch as the calculation of the required share encompasses broad methodological choices, it must be considered to be a complex economic assessment. Accordingly, the Commission must be able legitimately to define the parameters used to calculate that share, subject only to limited review by the EU Courts. That entails that, in the performance of their duties, the EU Courts must refrain from substituting their own assessment for that of the Commission, (79) unless the undertaking concerned argues that there has been a manifest error of assessment and also demonstrates such error. (80)

85.      That being said, the calculation of the required share cannot be immune to judicial review of calculation errors or selective or incomplete consideration of evidence. It is precisely in those cases that the case-law established in the judgment in Tetra Laval, as subsequently recast in the judgments in KME Germany and Chalkor, is applicable to its fullest extent. The evidence relied on by the Commission in order to assess a complex situation must indeed be factually accurate, reliable and consistent, that evidence must contain all the information needed for that assessment and must be capable of substantiating the conclusions drawn from it.

86.      In this instance, I am of the view that the General Court performed its assessment as regards the calculation by the Commission of HP’s required share in the manner required by the Court.

87.      Indeed, it is apparent from the judgment under appeal that the General Court first explained, in relation to the reference period set out by the Commission itself, the parameters used in Table 34 for the purposes of calculating HP’s required share and the quarter-by-quarter assessment also chosen by the Commission in order to carry out that calculation. The General Court next indicated the outcome of the global calculation of HP’s required share as regards the HPA1 agreement, presented in Table 35, which, as it also noted, resulted from the arithmetic sum or average of the figures set out in Table 34. The General Court then signalled the absence of any figure corresponding to the first part of the period covered by that agreement, in particular to the months of November and December 2002 and to the first three quarters of the 2003 tax year. That led it to consider that the calculation of HP’s required share could not be regarded as having been demonstrated for the entirety of the infringement period. (81)

88.      It follows that, as Intel and ACT argue, whilst the Commission chose to perform a quarter-by-quarter AEC assessment in order to determine HP’s required share during the reference period which had been set out in advance, it subsequently failed to incorporate into its calculations, as the Court’s case-law requires, all the relevant and necessary data. In those circumstances, the General Court was entitled to conclude, without encroaching on the Commission’s margin of discretion, that the evidence contained in the contested decision was not capable of substantiating its conclusions regarding the foreclosure effects of Intel’s rebates for the entire period covered by the HPA1 agreement.

89.      Consequently, the Commission’s criticism that the General Court imposed a standard of review exceeding the limits of the Court’s case-law ought, in my view, to be rejected.

90.      Furthermore, in so far as the Commission submits that it was justified in relying on the figures corresponding to the last three quarters covered by the HPA1 agreement because those figures were sufficiently representative for the entire period, that argument invites the Court to examine whether, in the present case, in the absence of specific evidence, it can be justified to make an inference as to the date for the beginning of that period by means of extrapolation.

91.      As is well known, extrapolating involves estimating an unknown value from the extension of a known sequence of values. To that effect, it entails an element of presumption, (82) which is a recurrent mechanism used for the purpose of alleviating the evidential burden (normally) borne by the Commission when establishing an infringement – or an element thereof – of the competition rules of the Treaty. (83) Moreover, inasmuch as extrapolations intend to infer the unknown from the known, they must be based on a concrete pattern. That pattern is usually drawn from a trend resulting from the particular identified sequence of values or, at least, from normal experience or common sense. It is important to note that, except in cases where it is self-evident, the pattern which allows an extrapolation to be made must be defined and made explicit by the party bearing the burden of proof.

92.      I could agree that defining the pattern capable of acting as a basis for an extrapolation of data can fall in certain cases within the margin of discretion available to the Commission when its definition entails an economic complexity. However, that question does not even arise in the present case, since the contested decision makes no reference to a representative or regular pattern which would demonstrate that the data for the second part of the period of the HPA1 agreement could also apply to the first part of that period. I therefore find unpersuasive the Commission’s argument that the omission of economic data found by the General Court in respect of that period was, in fact, the result of a voluntary extrapolation carried out by the Commission in the exercise of its margin of discretion.

93.      In that regard, the General Court was correct in observing that, in the contested decision, the Commission did not maintain that the demonstration of HP’s required share for the HPA1 agreement, as it resulted from Table 35, was the product of a coincidence. Nor did it hold that the various values set out in that table were identical for the missing three quarters and for the following three quarters. (84) From that perspective, the General Court was fully entitled to consider, as is also apparent from the judgment under appeal, (85) that it mattered little whether the calculations of the Commission were made on a quarter-by-quarter basis or on an overall basis, since the months of November and December 2002 and the first three quarters of the 2003 tax year would have never been taken into account, whatever figures they might have involved.

94.      It follows that the Commission’s argument based on an extrapolation of the values of the last three quarters of the period covered by the HPA1 agreement to the initial part of that period should be also rejected.

95.      Finally, the Commission submits that the figures used to calculate the required share during the second part of the period covered by the HPA1 agreement were stable from one quarter to the other, and that, in any event, they were favourable to Intel. In that regard, I would briefly note, without prejudice to my analysis within the third part of the present ground of appeal, (86) that that assessment clearly does not result from the contested decision and nor is it self-evident. As I shall explain later, it cannot therefore be used, within the present proceedings, to justify the failure to apply all the necessary data in order to calculate HP’s required share for the entire period covered by the HPA1 agreement.

96.      In the light of the foregoing considerations, I find that the General Court cannot be reproached for having infringed the Commission’s margin of discretion in complex economic matters – neither for committing the other errors alleged in this part – by concluding that the Commission had not demonstrated a foreclosure effect caused by the rebates which Intel granted to HP over the entirety of the reference period.

97.      The first part of the fourth ground of appeal submitted by the Commission should be, in my view, rejected.

3.      Intel’s implicit acknowledgement during the administrative proceedings

(a)    Arguments of the parties

98.      In the second place, the Commission claims that the judgment under appeal errs in denying any evidential value to Intel’s implicit acknowledgment, during the administrative proceedings, of the reference period used in the AEC test in relation to HP.

99.      According to the Commission, first, Intel did not challenge, during those proceedings, the choice of the reference period used in Table 35 of the contested decision to calculate HP’s required share. Intel and its economic advisors relied on the figures proposed by the Commission for the purposes of their own calculations and did not provide a counter-calculation for the allegedly missing quarters. Second, the Commission points out that, since there remains a legal presumption that exclusivity rebates are unlawful, according to the judgment adopted by the Court of Justice on the first appeal, it was for Intel to submit supporting evidence during the investigation to demonstrate that its rebates were not capable of foreclosing competition.

100. Intel disputes those arguments. It submits that Table 35 of the contested decision, which includes data for the three quarters in question and which was used as the reference period to calculate HP’s required share for the entire period of the HPA1 agreement, was never provided by the Commission to Intel during the administrative proceedings. It was introduced for the first time in the contested decision, which means that Intel never acknowledged the Commission’s choice of the reference period during those administrative proceedings. In any event, Intel submits, in line with the judgment under appeal, that there is no requirement that the addressee of a statement of objections must challenge its content during the administrative procedure.

(b)    Analysis

101. According to the case-law of the Court, where an undertaking concerned by an investigation under the competition rules of the Treaty does not expressly acknowledge the facts, the Commission must prove those facts and that undertaking is free to put forward, at the appropriate time and in particular in the procedure before the Court, any plea in its defence which it deems appropriate. (87)

102. Moreover, the Court has declared, in the judgment in Knauf Gips v Commission, (88) that, as regards the application of Articles 101 and 102 TFEU, there is no requirement under EU law that the addressee of a statement of objections must challenge its various matters of fact or of law during the administrative procedure in order not to be barred from doing so later at the stage of the judicial proceedings.

103. Finally, it expressly results from the judgment of the Court on the first appeal that, if, in a decision finding a rebate scheme abusive, the Commission carries out an AEC analysis, the General Court must examine all of the applicant’s arguments seeking to call into question the validity of the Commission’s findings concerning the foreclosure capability of the rebate concerned. (89)

104. In the present case, I observe that, in the judgment under appeal, (90) the General Court correctly referred to the judgment in Knauf Gips as a response to the Commission’s argument that Intel had not challenged, during the administrative proceedings, the reference period used by the Commission for the calculations of HP’s required share. That case-law, which is consistent with the declarations of the Court of Justice in the judgment on the first appeal, is clear about the limited value granted by the EU Courts, in the context of an action for annulment, to the position that an undertaking may have taken during the administrative proceedings. (91)

105. Indeed, it results from the Court’s case-law that an undertaking’s express or implicit acknowledgement of matters of fact or of law during the administrative proceedings must not restrict the actual exercise of a natural or legal person’s right to bring proceedings before the General Court under the fourth paragraph of Article 263 TFEU. (92) The basis of that understanding ultimately relies on the fundamental principles of the rule of law and of respect for the rights of the defence, as well as on the rights to an effective remedy and of access to an impartial tribunal as guaranteed by Article 47 of the Charter of Fundamental Rights of the European Union. (93)

106. Therefore, judicial review involving Articles 101 and 102 TFEU cannot prevent undertakings from raising arguments and relying on facts which they did not mention or did not contest during the administrative procedure, and also from putting forward evidence which the Commission did not have when it adopted the contested decision. (94)

107. The Commission recalls, however, that, in the judgment in Knauf Gips, the Court also considered that an undertaking’s acknowledgement of matters of fact or of law during the administrative procedure before the Commission may constitute additional evidence when determining whether an action is well founded. (95)

108. In that respect, it is sufficient to point out that it does not result from the judgment under appeal, nor from the Commission’s observations in the present case, that, as regards the rebates granted by Intel to HP, the Commission had adduced additional evidence during the administrative proceedings – including evidence relating to Intel’s alleged implicit acknowledgment of the reference period used in the AEC test in relation to HP – which was capable of corroborating HP’s required share for the entire period covered by the HP1A agreement.

109. It follows that that the statement of the Court’s judgment in Knauf Gips cited by the Commission is not relevant for the purposes of the present case.

110. Finally, inasmuch as the Commission argues, within the present part of the ground of appeal, that it was, in any event, for Intel to submit supporting evidence during the investigation to demonstrate that its rebates were not capable of foreclosing competition during the missing months and quarters of the reference period, that argument bears, in my view, little relation to its actual claim concerning Intel’s implicit acknowledgment of that period for the calculation of HP’s required share.

111. In any event, that argument must be rejected in the light of the case-law cited in point 103 above, which stems, as already stated, from the Court’s judgment on the first appeal.(96) That case-law also suffices to establish that the General Court also did not err in its response to the argument that Intel used the reference period provided by the Commission for its own calculations during the administrative proceedings. As is apparent from the judgment under appeal, (97) that period formed part of the reasoning of the contested decision and, therefore, it was capable of being challenged by the applicant before the General Court.

112. In the light of the foregoing, I consider that the General Court did not err in its assessment of the evidential value of Intel’s position during the administrative proceedings with respect to the reference period used in the AEC test to calculate HP’s required share.

113. The second part of the fourth ground should, in my view, be rejected.

4.      The infringement of the Commission’s right of defence

(a)    Arguments of the parties

114. In the third place, the Commission argues that the General Court erred in refusing to take into account the additional calculations which it presented in the context of the proceedings before that court, in particular in Annex D.17 to the rejoinder, for the purposes of disputing Intel’s arguments concerning the period of infringement covered by the HPA1 agreement.

115. The Commission points out that that annex demonstrated that Intel’s arguments were not capable of vitiating the outcome of the AEC test as carried out in the contested decision in relation to HP. Furthermore, the Commission considers that, while the General Court allowed Intel to adduce new analyses to challenge the reference period of the AEC test applied in respect of HP, it denied the Commission the right to respond. In that regard, the Commission mainly relies on the judgment in Dole Food and Dole Fresh Fruit Europe v Commission. (98)

116. Intel disputes those arguments. In essence, it submits that the Commission’s claim that it should be allowed to submit new AEC analysis for the first time in the judicial proceedings, in particular at the stage of the rejoinder, is manifestly inconsistent with the settled case-law of the Court.

(b)    Analysis

117. It is important to recall that, by Annex D.17 to its rejoinder, the Commission submitted for the first time before the General Court additional calculations based on a figure supplied by HP for two of the three missing quarters of the period covered by the HPA1 agreement, namely the second and third quarters of the 2003 tax year. (99) As already mentioned, (100) the purpose of that submission was to demonstrate that, because the amount of the rebates granted by Intel to HP were stable during the period covered by the HPA1 agreement, the results of HP’s required share would have been the same had the missing months and quarters been taken into account. Annex D.17 to the rejoinder also demonstrated, according the Commission, that the results of the calculation incorporating the data for the missing months and quarters were less favourable to Intel than the averaged results on which the contested decision was based.

118. In the judgment under appeal, the General Court rejected the request to take into account those additional calculations by declaring, first, that they were not apparent from the contested decision and were submitted for the first time during the judicial proceedings. In the General Court’s view, to take account of those calculations would have entailed, in clear contradiction with the Court’s case-law, substituting the reasoning set out by the Commission in the contested decision for its own assessment. (101) Second, be that as it may, the General Court concluded, in essence, that the amount of the rebates granted by Intel to HP was only one of the parameters needed for the calculation of HP’s required share and that information regarding the other parameters – namely the volume of HP’s purchases and the average sales price – was still lacking. In that context, the General Court stated that there was no guarantee that the data for the months and quarters not taken into account for the AEC test did not differ from those identified for the quarterly periods analysed. (102)

119. I must point out, from the outset, that, inasmuch as, in the present appeal, the Commission does not submit that the General Court erred in law in its assessment as regards the additional calculations presented in Annex D.17 to the rejoinder, the argument of the Commission based on the infringement of its rights of defence ought to be dismissed as being of no consequence. (103) In any event, from the point of view of admissibility, I consider that the General Court was correct in refusing to consider the content of that annex.

120. In that regard, I would recall that the refusal to accept further grounds from the Commission during the judicial proceedings for the purpose of sustaining an infringement of the competition rules of the Treaty can be approached from various angles, which concern both the form (external legality) and substance (internal legality) of the decision adopted by that institution.

121. Indeed, according to the Court’s case-law, even if the Commission may, in the proceedings before the General Court, explain in more detail in its defence the reasoning of a contested decision, the Commission may not introduce completely new reasons in those proceedings. That prohibition relies on the fact that the original lack of reasons cannot be cured at the stage of proceedings before the EU Courts by enabling the person concerned to learn of those reasons only during such proceedings. (104) As pointed out by Advocate General Kokott, that limit on introducing new reasons is of critical importance in criminal and quasi-criminal proceedings such as proceedings pursuant to Article 102 TFEU. (105)

122. Further, that same prohibition can be founded on the obligation of the Commission to base its decisions only on objections on which the parties concerned have been able to comment. (106) Indeed, the Commission must afford an undertaking the opportunity, during the administrative procedure, to make known its views on the truth and relevance of the evidence put forward by the Commission to support its claim that there has been an infringement of the Treaty. (107) That obligation concerns, among other essential elements of the infringement, its actual duration.

123. Finally, as the General Court correctly found in the judgment under appeal, the Court has declared that the amendment of an omission in the reasoning of a contested decision by taking into account additional grounds not present in that decision leads that court to substitute its assessment for that of the contested decision and, therefore, to err in law. (108)

124. I believe that any of the previous approaches, based on the Court’s case-law, was sufficient to allow the General Court to declare that Annex D.17 to the rejoinder was inadmissible, including the approach which it specifically adopted in its judgment, regarding the prohibition on substituting its assessment for that of the contested decision. In fact, to make a finding to the contrary would have led the General Court, in my view, to err in law for failing to find an infringement of the obligation to state reasons by the Commission, for failing to find an infringement of Intel’s rights of defence during the administrative proceedings or for substituting its own assessment for that of the contested decision.

125. In response to the above findings, the Commission relies on the judgment in Dole Food. In that judgment, the Court declared that, once an issue is raised for the first time in the application initiating the proceedings, the Commission may, without infringing the prohibition on introducing new reasons, defend its assessment in the contested decision by reference to any information provided in the course of the proceedings. (109)

126. To my mind, nevertheless, the factual context in the Dole Food case and the present situation differ. In Dole Food, the Court endorsed the General Court’s approach that the clarifications made in the proceedings before it served, in essence, merely to elucidate the statement of reasons already contained in the decision at issue. (110) By contrast, in the present case, the General Court found that the contested decision failed to establish that HP’s required share would have been the same for the initial period of the HPA1 agreement had the missing months and quarters been taken into account. Therefore, the additional calculations contained in Annex D.17 to the rejoinder, which were provided ex novo by the Commission at the judicial stage, were at no point anchored to the reasoning already present in the contested decision.

127. Besides, it appears from the Opinion of Advocate General Kokott in the Dole Food case that the evidence addressed in that case by the Commission before the General Court was contained in the investigation file. (111) That means that, contrary to the present instance, the undertaking concerned had acquaint itself with that information during the administrative proceedings.

128. Consequently, I am of the opinion that the judgment in Dole Food, as cited by the Commission, is not a valid precedent to support its arguments.

129. In the light of the foregoing, I consider that the General Court did not infringe the rights of defence of the Commission by refusing to take into account the additional calculations provided in Annex D.17 to the rejoinder.

130. The third part of the fourth ground of appeal submitted by the Commission should therefore, in my view, be dismissed.

5.      The error by the General Court in relation to the appropriate conclusion to be drawn in respect of the entire period of the practice at issue

(a)    Arguments of the parties

131. The Commission alleges that, even assuming that the General Court in the judgment under appeal was entitled to find that the Commission had not demonstrated foreclosure effects for the period between November 2002 and September 2003, that would not vitiate the conclusion that Intel’s rebates to HP were capable of having foreclosure effects at least from October 2003 to May 2005.

132. Intel disputes that argument. In particular, it argues that the judgment under appeal correctly establishes that, since the Commission did not consider properly the criterion relating to the share of the market covered by the contested practice and did not correctly analyse the duration of the rebates, the Commission had not demonstrated the foreclosure effects of the rebates granted to HP for the entire reference period.

(b)    Analysis

133. As already pointed out, the General Court found in the judgment under appeal, first, that the Commission erred in finding that its calculation of HP’s required share allowed it to draw conclusions concerning the foreclosure effect of those rebates for the entire period between November 2002 and May 2005. In the view of the General Court, the Commission had failed to demonstrate that that effect was present for the period between November 2002 and September 2003. (112)

134. Second, within the General Court’s examination of the criteria mentioned in paragraph 139 of the judgment on the first appeal, that court also found that the Commission had not considered properly, in the contested decision, the criterion relating to the share of the market covered by the Intel’s practice and to its duration. (113) 

135. It is on the basis of those considerations that the General Court further concluded, as regards the rebates granted by Intel to HP, that, even if it were necessary to infer that the AEC test could be regarded as conclusive for part of the infringement period, the errors in examining the criteria mentioned in paragraph 139 of the judgment on the first appeal meant that the foreclosure effect of the rebates was not demonstrated to the requisite legal standard. (114)

136. In my view, criticism of the reasoning of the General Court is not justified. That reasoning is actually consistent with paragraphs 138, 139 and 141 of the judgment on the first appeal, which the judgment under appeal correctly recalls as well. (115) That case-law requires that, where an undertaking in a dominant position submits, during the administrative procedure, on the basis of supporting evidence, that its conduct was not capable of restricting competition and, in particular, was not capable of producing the alleged foreclosure effects, the Commission must analyse the foreclosure capability of the system of rebates by applying the five criteria listed in paragraph 139 of the judgment on the first appeal. In addition, where the Commission has carried out an AEC test, that test is one of the factors which the Commission must take into account in assessing whether the system of rebates is capable of restricting competition.

137. In the present case, it is sufficient to observe that, in so far as the General Court had considered that the criterion regarding, inter alia, the market coverage of the contested practice had not been correctly assessed by the Commission, that court was entitled to conclude that the foreclosure effects of Intel’s rebates to HP could not be demonstrated even for the period between October 2003 to May 2005. (116)

138. It follows that, contrary to the Commission’s claim, the General Court did not commit any error in relation to the appropriate conclusion to be drawn in respect of the entire period of the infringement as regards the rebates granted by Intel to HP.

139. The fourth part of the fourth ground of appeal ought to be, in my view, rejected.

6.      Intermediate conclusion

140. Having regard to the foregoing, I must conclude that none of the analysed parts of the fourth ground of appeal is capable, in my view, of calling into question the conclusion of the General Court that the contested decision failed to demonstrate the foreclosure effect of the rebates granted by Intel to HP for the entire period between November 2002 to May 2005.

141. The fourth ground of appeal should be therefore dismissed. (117)

B.      The fifth ground of appeal, alleging a number of errors of law, the distortion of evidence and the infringement of the Commission’s rights of defence in the context of the examination of the AEC test with respect to Lenovo

142. By its fifth ground of appeal, the Commission contests the assessment made by the General Court of the AEC test applied in the contested decision with respect to Lenovo. In particular, the Commission criticises that assessment regarding the quantification of two non-cash advantages awarded by Intel in exchange of Lenovo’s exclusivity obligation – namely the extension of Intel’s standard one‑year warranty and the improved use of a hub facility situated in Shenzhen (China).

143. The findings of the General Court on that specific issue can be summarised as follows.

1.      The findings of the General Court as regards the non-cash advantages granted by Intel to Lenovo

144. It results from the judgment under appeal that the General Court considered it to have been established in the initial judgment that Intel and Lenovo had concluded a declaration of intent, the 2007 Memorandum of Understanding (‘2007 MoU’), which was subject to an unwritten exclusivity condition. (118) It also follows from the judgment under appeal that, according to the Commission, the amount of the rebates granted by Intel to Lenovo was set out in that memorandum, which provided for a financial support of 180 million United States dollars (USD) for 2007, in the form of quarterly payments. (119)

145. Moreover, in the judgment under appeal, (120) the General Court pointed out that, during the administrative proceedings before the Commission, Intel had put forward the argument that the relevant value for the size of the rebates was only USD 138 million. That was explained by the fact that, out of the financial support for Lenovo of USD 180 million provided for in the 2007 MoU, only USD 135 million was awarded in the form of cash. The remainder of the financial support was granted in the form of non-cash advantages, namely the extension of Intel’s standard one-year warranty and the offer of improved use of an Intel hub facility situated in China. The Commission stated that Intel had submitted that, whilst the value of those two non-cash contributions for Lenovo was USD 20 million and USD 24 million respectively, their cost for Intel was much lower, namely USD 1.7 million and USD 1.3 million respectively. According to Intel, for the purposes of the as-efficient-competitor analysis, those elements had to be assessed not by reference to their value to Lenovo, but by reference to their economic cost to Intel. Intel reached the amount of USD 138 million by adding the USD 1.7 million and USD 1.3 million costs to the USD 135 million cash funding.

146. The General Court also observed that, in the contested decision, the Commission rejected Intel’s argument, considering that it was based on a misunderstanding of the principles of the as-efficient-competitor analysis. (121) In the Commission’s view, that analysis assessed the price at which a competitor which is as efficient as the dominant company – but which is not dominant – would have to offer its products in order to compensate the customer for the loss of the conditional benefits granted by the dominant company and which would result from that customer’s switching the contestable share of its supply needs away from the dominant company to the hypothetical as-efficient competitor. (122) In the view of the Commission, it was clear then that the relevant measure was the loss for the customer, as that was the loss which the as-efficient competitor would have to compensate, and not the economic cost to the dominant undertaking in the event that the two figures diverged. (123)

147. Contrary to the view of the Commission, and in line with Intel’s main argument, the General Court found, in the judgment under appeal, that that institution proceeded on the basis of an assumption which was contrary to the foundation of the AEC test set out in the contested decision, which was based on the principle that the hypothetical competitor is as efficient as Intel, in particular from the perspective of the costs of extending a hub or a warranty. (124)

2.      The misinterpretation of the AEC test applied in the contested decision and of Article 102 TFEU

(a)    Arguments of the parties 

148. The Commission submits that the judgment under appeal errs in law in defining the nature of the AEC test applied in the contested decision, which ultimately amounts to a misapplication of Article 102 TFEU. It considers that the General Court incorrectly assessed, in cost-price terms, whether a competitor is as-efficient as a dominant undertaking in the context of rebates granted in the form of non-cash advantages.

149. The Commission states, at the outset, that the AEC test is an analytical tool which involves a number of technical methodological choices. As such, the findings in the contested decision regarding the rebates granted by Intel to Lenovo could only be set aside in case of a manifest error, which the General Court did not find.

150. Moreover, according to the Commission, the General Court’s error consisted of concluding that the basis for assessing the value of the non-cash advantages granted by Intel was the costs of that undertaking in providing them, rather than their value to Lenovo. In that regard, the Commission reiterates, in essence, the answer provided in the contested decision to Intel’s argument during the administrative proceedings, as described in point 146 above.

151. Indeed, even assuming that the non-cash advantages granted by Intel had to be examined from the perspective of the costs to that undertaking in providing them, the General Court incorrectly failed to consider, in the view of the Commission that a competitor smaller than Intel would not have hub facilities such as that offered by Intel to Lenovo and would not be capable of offering a similar extended warranty on its products. That competitor would have therefore had to offer a cash compensation for Lenovo’s loss of Intel’s non-cash advantages.

152. Finally, the Commission argues that, should the Court reject the above submissions, the General Court’s refusal to consider Annex D.39 to the rejoinder in order to rebut Intel’s claims infringed its rights of defence.

153. Intel dispute those arguments. In its view, the General Court’s approach is consistent with the Court’s settled case-law that the focus of Article 102 TFEU is on as efficient – not less efficient – competitors and, in particular, that assessments of pricing practices must be based on the costs and the strategy of the dominant undertaking itself. In fact, the Commission’s own guidance on Article 102 TFEU also affirms that, where available, the Commission will use information on the costs of the dominant undertaking. A different approach would penalise a dominant undertaking for operating more efficiently than its rivals and thus delivering a product to a customer at a lower cost.

(b)    Analysis

154. As a preliminary remark, I would point out, first, that there is no dispute between the parties concerning whether non-cash advantages such as those at issue in the present case had to be taken into account for the purposes of the application of the AEC test in relation to Lenovo. Rather, what is disputed is how to calculate those advantages for the purposes of applying that test. The Commission considers, in fact, that the General Court misunderstood the principles underlying the AEC test as set out in the contested decision and erred in its interpretation of how to apply that test in cost-price terms. It should be noted that, according to the Court’s case-law, the AEC test proves useful even where the elements of the practice in question are not expressed in cash terms, but can still be quantified. (125)

155. Second, contrary to the arguments of the Commission, I do not think that the General Court’s assessment in the judgment under appeal called into question the configuration of the AEC test or, as the Commission puts it, ‘the methodological choices’ made by that institution when setting that test out. Instead, it appears from the judgment under appeal that the General Court verified whether the application of the AEC test to the non-cash advantages awarded by Intel to Lenovo was consistent with the foundation underlying that test specifically defined by the Commission in the contested decision. From that perspective, the General Court should not be criticised, in my view, for encroaching on the Commission’s margin of discretion when framing the AEC test. The General Court’s analysis actually refers to the internal consistency of the Commission’s application of the AEC test in this particular case.

156. As regards whether the General Court erred in setting aside the approach adopted by the Commission in relation to the non-cash advantages granted by Intel to Lenovo, it follows correctly from the judgment under appeal (126) that the logic inherent in the AEC test as applied in the contested decision was to examine whether Intel itself, in view of its own costs and the effects of the rebates, would be able to enter the market at a more limited scale without incurring losses. The General Court further pointed out, without any error being committed, that the AEC analysis was configured in the contested decision as a purely hypothetical exercise, in the sense that it attempted to analyse whether a competitor as efficient as Intel, in terms of producing and delivering x86 CPUs which provide the same value to customers as Intel, but which would not have as broad a sales base as Intel, would be foreclosed from entering. (127)

157. I observe that, in its case-law, the Court of Justice has declared that the assessment of pricing practices must be based ‘on the costs and the strategy of the dominant undertaking itself’ (128) and, in particular, that the AEC test must be carried out by taking into account a hypothetical competitor with a cost structure similar to that of the undertaking in a dominant position. (129) Moreover, the Court has declared that the validity of such an approach is reinforced by the fact that it conforms to the general principle of legal certainty, since taking into account the costs and prices of the dominant undertaking enables that undertaking to assess the lawfulness of its own conduct, which is consistent with its special responsibility under Article 102 TFEU. (130) The General Court’s interpretation of the foundation of the AEC test in the contested decision appears, therefore, to be consistent with the Court’s case-law.

158. Furthermore, that interpretation is also consistent with the manner in which the different parameters needed for the application of the AEC test must be calculated according to the contested decision. As the General Court states in the judgment under appeal, (131) the AEC test is designed in that decision to determine whether the competitor which is as efficient as the undertaking in a dominant position, which faces the same costs as the latter, can still cover its costs on a minor scale. Therefore, in every calculation step leading to the application of the AEC test, as described for instance in points 43 and 83 of the present Opinion, the costs of the dominant undertaking are at their main centre of the analysis.

159. It is true that, where an exclusivity rebate is granted in cash, its value is objective and is the same for both the dominant undertaking and the beneficiary of the rebate. By contrast, where that rebate is granted in a non-cash form, the value for that undertaking and for the beneficiary might differ, which raises the question of how is it to be assessed. That being said, the logic of the calculation of that value must respect the same foundation in both cases. Otherwise, the assessment of whether a rebate is capable of having a foreclosure effect would not be made by reference to the relevant economic operator. In the present case, there is no doubt, as the General Court correctly indicates in the judgment under appeal, (132) that that relevant economic operator should be, in principle, a competitor as efficient as Intel offering non-cash advantages to Lenovo on the same terms as that dominant undertaking.

160. Finally, it cannot be ruled out that the latter value may require a certain adjustment to allow account to be taken of the fact that the as-efficient competitor does not hold a dominant position and that that competitor might operate on a minor scale. In the contested decision, the Commission illustrated that situation by referring, in essence, to the case where the as-efficient competitor to the dominant undertaking would not be capable of providing access to a similar supply hub or warranty extension. (133)

161. However, even assuming that possibility, that does not justify that a rebate in a non-cash form be assessed, for the purposes of an AEC analysis, at the level of the value which that rebate represents for its recipient. The cost for a competitor as efficient as Intel of providing access to a supply hub should have been calculated. Consequently, the General Court was right to consider that the quantified assessment made by the Commission of the non-cash advantages offered by Intel to Leonovo was itself vitiated by an error. (134)

162. In the light of those considerations, I consider that the criticism submitted by the Commission as regards the approach adopted by the General Court should be rejected. The General Court did not commit, in my view, any error by concluding that the Commission, when assessing the non-cash advantages granted by Intel to Lenovo, had proceeded on the basis of an assumption which was contrary to the foundation of the AEC test set out in the contested decision.

163. For the remainder, inasmuch as the Commission argues that the General Court infringed its rights of defence due to its refusal to take into account Annex D.39 to the rejoinder, that claim should be rejected on grounds similar to those set out in detail in my analysis of the third part of the fourth ground of appeal. (135)

164. In particular, Annex D.39 to the rejoinder demonstrated, according to the Commission, that, even if it were accepted that a competitor as-efficient as Intel had a supply hub in China, the cost it would incurr in making that hub available to Lenovo would have been significantly higher than the cost for Intel of providing that advantage.

165. In that regard, I would merely point out that that analysis corresponds to my observation in point 160 above, whereby I accept that a certain adjustment might prove necessary on the calculation of the production cost of the required share to take into consideration the fact that an as-efficient competitor as Intel operates on a minor scale.

166. However, inasmuch as the Commission only provided a response to that question at the stage of the rejoinder, as the General Court also correctly points out, (136) it was entitled to refuse to take into account the additional analysis provided in Annex D.39 to the rejoinder without infringing the rights of defence of the Commission.

3.      Intermediate conclusion

167. Having regard to the foregoing, I conclude that the Commission’s criticisms of the General Court’s assessment of the AEC test as applied to Lenovo in the contested decision, in particular as regards two non-cash advantages granted by Intel in exchange for its exclusivity obligation, is not well founded and ought to be rejected.

168. The fifth ground of appeal ought therefore to be dismissed.

V.      Conclusion

169. In the light of the analysis of the present Opinion, I propose that the Court should dismiss the appeal as regards the fourth ground of appeal and as regards the fifth ground of appeal.

170. I express no opinion on the dismissal of the appeal as regards the remainder of the grounds of appeal submitted by the Commission, nor on which party ought to be ordered to pay the costs pursuant to Article 184(2) of the Rules of Procedure of the Court of Justice.

1      Original language: English.

2      EU:T:2022:19 (‘the judgment under appeal’).

3      EU:C:2017:632 (‘the judgment on the first appeal’).

4      EU:T:2014:547 (‘the initial judgment’).

5      Case COMP/C‑3/37.990 – Intel (‘the contested decision’).

6      The judgment under appeal, paragraph 526.

7      The judgment under appeal, paragraph 524.

8      See, for further details, the judgment under appeal, paragraphs 1 to 61.

9      Council Regulation of 16 December 2002 on the implementation of the rules on competition laid down in Articles [101] and [102 TFEU] (OJ 2003 L 1, p. 1).

10      The contested decision, Article 1.

11      The contested decision, Article 2.

12      As rectified by orders of 19 September and 24 October 2017.

13      The judgment on the first appeal, paragraphs 149 and 150.

14      The judgment on the first appeal, paragraphs 65 and 107.

15      The judgment on the first appeal, paragraphs 133 and 134, where the Court refers to the judgment of 27 March 2012, Post Danmark (C‑209/10, EU:C:2012:172, paragraph 22).

16      The judgment on the first appeal, paragraphs 135 to 137.

17      85/76, EU:C:1979:36 (‘the judgment in Hoffmann-La Roche’).

18      The judgment on the first appeal, paragraph 138.

19      See the judgment on the first appeal, paragraph 139 (‘the criteria referred to in paragraph 139 of the judgment on the first appeal’).

20      The judgment on the first appeal, paragraph 141.

21      The judgment on the first appeal, paragraph 143.

22      The judgment on the first appeal, paragraph 144.

23      The judgment on the first appeal, paragraph 145.

24      The judgment on the first appeal, paragraph 147.

25      The judgment under appeal, paragraphs 74 to 102.

26      The judgment under appeal, paragraph 85.

27      The judgment under appeal, paragraph 96.

28      The judgment under appeal, paragraphs 97 and 98.

29      The judgment under appeal, paragraph 101.

30      The judgment under appeal, paragraphs 116 to 122.

31      The judgment under appeal, paragraphs 123 to 127.

32      The judgment under appeal, paragraph 145.

33      The judgment under appeal, paragraph 149.

34      The judgment under appeal, paragraphs 150 and 151.

35      The judgment under appeal, paragraphs 152 to 159.

36      The judgment under appeal, paragraphs 175, 258, 260, 283, 285, 286, 297 to 299 and 334.

37      The judgment under appeal, paragraphs 160 to 166.

38      The judgment under appeal, paragraphs 167 to 481.

39      The judgment under appeal, paragraph 482.

40      The judgment under appeal, paragraphs 483 to 520.

41      The judgment under appeal, paragraph 521.

42      The judgment under appeal, paragraph 524.

43      The judgment under appeal, paragraph 526.

44      The judgment under appeal, paragraph 527 and point 1 of the operative part.

45      See point 38 of the present Opinion.

46      The judgment under appeal, paragraph 529.

47      The judgment under appeal, paragraph 527 to 531 and point 3 of the operative part.

48      In its application, the Commission additionally alleges, in a fifth part, that the General Court erred in law in finding that the contested decision failed to state reasons with regard to the reinforcing factors which, according to that decision, exacerbated the results of the AEC test as carried out for HP. That part not being subject of the request of the Court, I shall exclude it from my analysis.

49      The judgment under appeal, paragraphs 288 and 289, which refer to the contested decision, recitals 338, 341, 413 and 1296.

50      The judgment under appeal, paragraph 288, which refers to recital 1406 of the contested decision.

51      The judgment under appeal, paragraphs 297 to 299, 303 and 304, which refer to recitals 1334 to 1337 and 1385 to 1389 of the contested decision.

52      The judgment under appeal, paragraph 321, which refers to recitals 1390 to 1395 of the contested decision. The reinforcing factors were, first, that the Commission had used figures which were most favourable to Intel and, second, that, if HP were to switch its purchases to AMD, Intel could in turn transfer the rebates initially intended for HP to another competitor using its x86 CPUs, such as Dell.

53      The judgment under appeal, paragraph 303, which indicates that Table 34 is to be found in recital 1334 of the contested decision.

54      Those quarters are indicated by a sequence of abbreviations that goes from Q4 FY03 to Q3 FY05, ‘Q’ standing for ‘quarter’ and ‘FY’ standing for ‘fiscal (tax) year’.

55      The judgment under appeal, paragraph 292, which indicates that Table 35 is to be found in recital 1337 of the contested decision.

56      The judgment under appeal, paragraph 304.

57      The judgment under appeal, paragraphs 298 and 299, which refer to recitals 1385 to 1389 and 1406 of the contested decision.

58      The judgment under appeal, paragraph 291.

59      The judgment under appeal, paragraph 303.

60      The judgment under appeal, paragraphs 304 and 305.

61      The judgment under appeal, paragraph 307.

62      The judgment under appeal, paragraph 306.

63      The judgment under appeal, paragraph 307.

64      The judgment under appeal, paragraph 308 to 310.

65      The judgment under appeal, paragraphs 316 to 318.

66      See, to that effect, judgment of 10 July 2014, Telefónica and Telefónica de España v Commission (C‑295/12 P, EU:C:2014:2062, paragraph 54).

67      Regarding the evolution of the standard of limited review for complex economic assessments, see Jaeger, M., ‘The standard of review in competition cases involving complex economic assessments: towards marginalisation of the marginal review?’, Oxford Journal of European Competition Law & Practice, Vol. 2, Issue 4, 2011, pp. 295 et seq.; and Da Cruz Vilaça, J. L., ‘The intensity of judicial review in complex economic matters – recent competition law judgments of the Court of Justice of the EU’, Journal of Antitrust Enforcement, Vol. 6, Issue 2, 2018, pp. 173 et seq.

68      Judgment of 13 July 1966 (56/64 and 58/64, EU:C:1966:41, p. 347).

69      Judgment of 11 July 1985, Remia and Others v Commission (42/84, EU:C:1985:327, paragraph 34).

70      See judgment of 21 November 1991, Technische Universität München (C‑269/90, EU:C:1991:438, paragraph 14).

71      Judgment of 15 February 2005 (C‑12/03 P, EU:C:2005:87, ‘the judgment in Tetra Laval’, paragraph 39).

72      The test formulated in the judgment in Tetra Laval is mentioned in relation to Article 102 TFEU in judgment of 10 July 2014, Telefónica and Telefónica de España v Commission (C‑295/12 P, EU:C:2014:2062, paragraph 54), as well as in judgments of the General Court of 10 April 2008, Deutsche Telekom v Commission (T‑271/03, EU:T:2008:101, paragraphs 185); of 30 January 2007, France Télécom v Commission (T‑340/03, EU:T:2007:22, paragraphs 163 and 165); of 17 September 2007, Microsoft v Commission (T‑201/04, EU:T:2007:289, paragraphs 379 to 381); and of 1 July 2010, AstraZeneca v Commission (T‑321/05, EU:T:2010:266, paragraph 32).

73      Judgment of 8 December 2011 (C‑272/09 P, EU:C:2011:810, ‘the judgment in KME Germany’).

74      Judgment of 8 December 2011  (C‑386/10 P, EU:C:2011:815, ‘the judgment in Chalkor’).

75      See the judgment in KME Germany, paragraph 102, and the judgment in Chalkor, paragraph 62.

76      See judgment of 19 January 2023, Unilever Italia Mkt. Operations (C‑680/20, EU:C:2023:33, paragraphs 56 and 57), in which the Court held, as regards the AEC test, that ‘that concept refers to various tests which have in common the aim of assessing the ability of a practice to produce anti-competitive exclusionary effects’, and that the AEC test is ‘only one of a number of methods’ for assessing those effects.

77      See point 43 of the present Opinion.

78      See point 44 of the present Opinion.

79      See, in that regard, Opinion of Advocate General Kokott in Commission v Alrosa (C‑441/07 P, EU:C:2009:555, point 90), which cites judgment of 10 July 2008, Bertelsmann and Sony Corporation of America v Impala (C‑413/06 P, EU:C:2008:392, paragraph 145).

80      An illustration is provided by judgment of 10 April 2008, Deutsche Telekom v Commission (T‑271/03, EU:T:2008:101, paragraph 183 et seq.), upheld by judgment of 14 October 2010, Deutsche Telekom v Commission (C‑280/08 P, EU:C:2010:603, paragraph 143).

81      See point 68 of the present Opinion.

82      Ritter, C., ‘Presumptions in EU competition law’, Journal of Antitrust Enforcement, 2018, Vol. 6, p. 193.

83      A classic example in competition law is the parental liability presumption. See judgment of 10 September 2009, Akzo Nobel and Others v Commission (C‑97/08 P, EU:C:2009:536). An exhaustive list of presumptions can be found in Ritter, C., op. cit., pp. 189 to 212, and Bailey, D., ‘Presumptions in EU competition law’, European Competition Law Review, Vol. 9, Issue 20, 2010, p. 20.

84      The judgment under appeal, paragraph 306.

85      The judgment under appeal, paragraph 310.

86      See points 117 to 129 of the present Opinion.

87      Judgment of 16 November 2000, SCA Holding v Commission (C‑297/98 P, EU:C:2000:633, paragraph 37).

88      Judgment of 1 July 2010 (C‑407/08 P, EU:C:2010:389, ‘the judgment in Knauf Gips’, paragraph 89).

89      The judgment on the first appeal, paragraph 141.

90      The judgment under appeal, paragraphs 300 to 302.

91      See also judgment of 11 July 2013, Ziegler v Commission (C‑439/11 P, EU:C:2013:513, paragraphs 57 and 58).

92      The judgment in Knauf Gips, paragraph 90.

93      The judgment in Knauf Gips, paragraph 91.

94      See, in that regard, van der Woude, M., ‘Judicial control in complex economic matters’, Journal of European Competition Law & Practice, Vol. 10, n. 7, 2019, p. 421.

95      The judgment in Knauf Gips, paragraph 90.

96      See also the judgment on the first appeal, paragraph 144.

97      The judgment under appeal, paragraph 302.

98      Judgment of 19 March 2015 (C‑286/13 P, EU:C:2015:184, ‘the judgment in Dole Food’).

99      The judgment under appeal, paragraphs 316 and 317.

100      See point 95 of the present Opinion.

101      The judgment under appeal, paragraph 317, which also refers to paragraph 150 of that judgment.

102      The judgment under appeal, paragraph 318.

103      That observation is not called into question by the succinct argument put forward by the Commission in a footnote to its appeal, submitting, in essence, that the amount of the rebates is ‘by far’ the most important parameter in the calculation of the required share and that the additional calculations provided to the General Court on the basis of that parameter were therefore sufficient to calculate HP’s required share for the missing quarters. The well-established case-law requiring arguments brought before the Court to be properly developed applies. See, to that effect, order of 26 March 2020, Magnan v Commission (C‑860/19 P, EU:C:2020:227, paragraph 27 and the case-law cited).

104      See, inter alia, judgment of 28 June 2005, Dansk Rørindustri and Others v Commission (C‑189/02 P, C‑202/02 P, C‑205/02 P to C‑208/02 P and C‑213/02 P, EU:C:2005:408, paragraph 463).

105      See Opinion of Advocate General Kokott in Dole Food and Dole Fresh Fruit Europe v Commission (C‑286/13 P, EU:C:2014:2437, point 26).

106      See the second sentence of Article 27(1) of Regulation No 1/2003

107      See, to that effect, judgment of 14 May 2020, NKT Verwaltungs and NKT v Commission (C‑607/18 P, EU:C:2020:385, paragraph 50 and the case-law cited).

108      See judgments of 24 January 2013, Frucona Košice v Commission (C‑73/11 P, EU:C:2013:32, paragraphs 89 and 90), and of 21 January 2016, Galp Energía España and Others v Commission (C‑603/13 P, EU:C:2016:38, paragraphs 73, 78 and 79).

109      The judgment in Dole Food, paragraph 38.

110      See, to that effect, the judgment in Dole Food, paragraphs 34 to 38, read in conjunction with judgment of 14 March 2013, Dole Food and Dole Germany v Commission (T‑588/08, EU:T:2013:130, paragraphs 46 and 47).

111      See Opinion of Advocate General Kokott in Dole Food and Dole Fresh Fruit Europe v Commission(C‑286/13 P, EU:C:2014:2437, point 22).

112      The judgment under appeal, paragraphs 319 and 334.

113      The judgment under appeal, paragraphs 500 and 520.

114      The judgment under appeal, paragraph 525.

115      The judgment under appeal, paragraph 522, under the heading ‘Conclusion on the application for annulment of the contested decision’. See also points 32 and 33 of the present Opinion.

116      Nevertheless, it is apparent from the appeal in the present case that, by its first ground of appeal, the Commission seeks to question the validity of the General Court’s finding regarding the assessment according to the criteria mentioned in paragraph 139 of the judgment on the first appeal. Should the Court uphold that first ground of appeal, the present part of the fourth ground of appeal would require assessment in the light of that consideration.

117      That intermediate conclusion would remain correct even if the Court were to consider that the fifth part of the fourth ground of appeal – which, as already noted, is not covered by the present Opinion – is well founded. That part concerns the reinforcing factors that exacerbated the results of the AEC test carried out as regards HP, which, by definition, cannot be sufficient to compensate the errors correctly found by the General Court in the application of that test according to my analysis.

118      The judgment under appeal, paragraph 98, which refers to paragraphs 1045 to 1208 of the initial judgment.

119      The judgment under appeal, paragraph 415, which refers to recital 1461 of the contested decision.

120      The judgment under appeal, paragraph 417, which refers to recital 1463 of the contested decision.

121      The judgment under appeal, paragraph 419.

122      The judgment under appeal, paragraph 420, which refers to recital 1466 of the contested decision.

123      The judgment under appeal, paragraph 421, which refers to recital 1467 of the contested decision.

124      The judgment under appeal, paragraphs 433 to 439, which refer to recitals 1003 and 1004 of the contested decision.

125      See, to that effect, judgment of 19 January 2023, Unilever Italia Mkt. Operations (C‑680/20, EU:C:2023:33, paragraph 59).

126      The judgment under appeal, paragraph 434, which refers to recital 1003 of the contested decision.

127      The judgment under appeal, paragraph 435, which refers to recital 1004 of the contested decision.

128      Judgment of 14 October 2010, Deutsche Telekom v Commission (C‑280/08 P, EU:C:2010:603, paragraph 198).

129      Judgment of 19 January 2023, Unilever Italia Mkt. Operations (C‑680/20, EU:C:2023:33, paragraph 59).

130      See judgment of 14 October 2010, Deutsche Telekom v Commission (C‑280/08 P, EU:C:2010:603, paragraph 202).

131      The judgment under appeal, paragraphs 152, 154, 157 and 158. See also point 43 of the present Opinion.

132      The judgment under appeal, paragraph 437 and 439.

133      The judgment under appeal, paragraph 438.

134      See the judgment under appeal, paragraph 455.

135      See points 117 to 129 of the present Opinion.

136      The judgment under appeal, paragraph 441.