INTEL V. COMMISSION : LANDMARK VICTORY FOR COMPANIES, SIGNIFICANT LOSS FOR COMMISSION

On 26 January 2022 the General Court (GC) issued its latest judgment in the Intel case, faulting fundamental aspects of the Commission’s original decision almost 13 years prior. The judgment’s findings are promising for pending cases appealing similar issues, including the application of the “as efficient competitor” (AEC) test and the economic analysis necessary to establish the abusive nature of a rebates system. The judgment could also have serious ramifications for the Commission – both in terms of its substantive infringement finding procedures and its duty to repay hefty annulled fines.

Background

The Commission issued its original decision in May 2009, finding that Intel had committed a single and continuous infringement of Article 102 TFEU from October 2002 until December 2007, by implementing a strategy aimed at foreclosing its competitor, Advanced Micro Devices Inc. (AMD), from the market for x86 central processing unit (CPU) microprocessors.

This strategy encompassed two aspects – conditional rebates and so-called “naked” restrictions. The former involved Intel granting four original equipment manufacturers (OEMs) rebates conditional on such OEMs purchasing all or almost all of their x86 CPUs from Intel, as well as payments made to a desktop computer distributor which were conditional on such distributor exclusively selling computers containing Intel’s x86 CPUs. The latter involved Intel awarding three OEMs payments conditional on such OEMs postponing or cancelling the launch of products with x86 CPUs from Intel’s competitor, AMD, or otherwise placing restrictions on the distribution of such products.

In its decision, the Commission concluded that Intel’s conditional rebates and payments induced the loyalty of key OEMs and the distributor, significantly diminishing competitors’ ability to compete on the merits of their x86 CPUs. The naked restrictions were deemed to directly harm competition and not constitute normal competition on the merits. Both sets of conduct were held by the Commission to form part of a single strategy aimed at foreclosing AMD from the market for x86 CPUs, contrary to Article 102.

The Commission imposed a fine of €1.06 billion on Intel – at that time, the highest sanction of its kind. Although initially upheld on appeal to the GC in 2014, the Court of Justice (ECJ) in 2017 subsequently found that the Commission failed to adequately assess whether the conditional rebates were capable of restricting competition on the basis of the AEC test, and referred the case back to the GC. The GC’s latest judgment is the outcome of such referral.

Conditional Rebates

The key question before the GC was whether the conditional rebates Intel had granted OEMs had the capacity to restrict competition in light of the analysis set out in the ECJ’s 2017 judgment,^[1] which required that the Commission consider four factors in making such a finding: (1) the extent of the undertaking’s dominant position on the relevant market, (2) the share of the market covered by the contested practice, (3) the conditions and arrangements for granting the rebates in question and (4) their duration and their amount. The ECJ added that the Commission was 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 (i.e. the AEC test).^[2]

The GC reiterated the test set out by the ECJ in its 2017 judgment, and confirmed this was, “as a minimum,” what the Commission was required to examine in order to assess the foreclosure capability of a rebates system.^[3] In conclusion, the GC sided with the applicants, finding that the Commission was wrong to find that the rebates infringed Article 102 on the basis that they were abusive by their nature, without conducting a more detailed analysis of their capability to foreclose.^[4]

In regard to the specific factors for consideration, the GC found that:

• The Commission failed to adequately determine the share of the market covered by the conduct, by inter alia (1) looking only at two OEM’s market shares and excluding other OEMs and (2) considering only the period from Q1 2003 to Q4 2005. This was deemed not only contrary to the requirements set out in the 2017 judgment, but also to the Commission’s own guidelines on the analysis of Article 102 cases.^[5]
• The Commission’s review of the duration of the rebates within the context of applying the AEC test was insufficient to meet the second condition for assessing foreclosure capability as set out in the 2017 judgment. Specifically, it was inadequate for the Commission to examine such duration in “a haphazard and limited manner” without performing “a thorough and exhaustive examination for all OEMs of those aspects.”^[6]

In light of the above, the GC concludes in its 2022 judgment that the Commission did not adequately consider the relevant criteria to establish foreclosure capability, without having to assess the applicants’ arguments in regard to the first and third conditions from the 2017 judgment.

AEC Test

The GC also looked at the application of the AEC test, which 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 short, its application involves calculating: (1) the share of customers’ demand for x86 CPUs which they would be willing and able to obtain from suppliers other than Intel (Contestable Share), (2) the portion of the rebates for which an AEC must offer compensation (i.e., the portion subject to the exclusive supply condition, excluding any quantity rebates (Conditional Portion)) and (3) Intel’s costs (Costs).

The GC reviewed a number of the Commission’s applications of the AEC test in detail. Although much of these were specific to each of the relevant rebates systems, some generic findings based on the GC’s approach can be summarised as follows:

• The Commission must be prepared to review all relevant evidence in calculating the correct values on application of the AEC test. For instance, with regards to the calculation of the Contestable Share in a rebate granted to an OEM, the Commission had incorrectly relied on just an internal spreadsheet of such OEM and not additional materials which illustrated a higher value (namely, statements made by the OEM executives during the course of a US private litigation between Intel and AMD and other internal emails).^[7]
• The Commission should also adduce evidence encompassing the entire period of the alleged infringements. For another OEM rebate, the Commission invalidly sought to allege a foreclosure effect for the entire period between November 2002 and May 2005, without adducing sufficient evidence demonstrating such effect for the period between November 2002 and September 2003.^[8] The GC also held that the Commission was unable to draw conclusions in regard to Intel’s ability to foreclose an AEC between Q4 2002 and November 2005 by relying on data from Q4 2002 alone.^[9]

Conclusion

The latest Intel judgment is a landmark .It conducts a detailed review of the Commission’s assessment, clarifying that a per se infringement finding under Article 102 is not feasible – for rebates at least – and the Commission should always comply with its obligation to examine competitive effects in such cases, specifically where, during the administrative procedure, the applicant adduces evidence that its conduct was not capable of restricting competition.^[10] Equally, the AEC test is not something the Commission can apply lightly or cut corners with – it requires robust substantive evidence and analysis.

Aside from picking holes in the details of the Commission’s evidential analysis, the GC also annulled the €1.06 billion fine in full. It claimed it was unable to determine the fine amount relating solely to the upheld portion of the infringement regarding the naked restrictions, based on the Commission’s response to the Court’s question on the severability of such restrictions from the rebates – to this, the Commission replied in its usual fashion, stating the conduct had been assessed “as a whole” and both infringements “complemented and mutually reinforced one another.”^[11] This confidence in its single and continuous infringement analysis has ultimately been the Commission’s downfall in this case, leading to an annulment of one of its highest fines on record and a harsh critique of its substantive decision making process.

While the Commission will no doubt assess whether the GC’s findings are worth appealing to the ECJ, the judgment itself is potentially promising to the other dominance cases pending before the Court which similarly challenge the application of the AEC test, namely the Qualcomm and Google Android cases.^[12] Here, it remains to be seen whether the Commission has evolved sufficiently in its economic analysis to validly establish Article 102 infringements.

More wide-ranging impacts are also possible – though much of the facts are limited to the assessment of rebates within the Article 102 context, the overarching assessment of the Commission’s high standard of proof may well be carried through to future cases with the Commission having to provide increasingly substantiated infringement decisions in future. This is all in addition to the potential dent in the Commission’s budget, arising not only out of the repayment of a hefty principal fine amount but also the default interest accrued, alluded to by the GC just last week in Deutsche Telekom.^[13]

Overall, the previous rubber-stamping practice of the EU Courts is evolving into what appears to be a relentless rap on the Commission’s knuckles for both its procedural and substantive inadequacies. This may well lead to a welcome shift in the Commission’s approach to cases, encouraging it to establish more robust infringement findings in future.