The devil is in the detail: General Court’s renvoi judgment in Intel demands rigorous effects analysis in abuse of dominance cases

In a 26 January 2022 judgment, the General Court has partially overturned the Commission’s 2009 decision and entirely overturned the associated €1.06bn fine imposed on Intel. In its 2009 decision, the Commission found that Intel had abused its dominant position in x86 CPUs by granting exclusivity or quasi-exclusivity rebates to four computer manufacturers and one retailer, and by imposing so-called “naked restrictions” by making payments to three computer manufacturers to delay the launch of products containing AMD’s x86 CPUs.

The Commission must analyse effects of anti-competitive conduct

This judgment tells us that the Commission cannot rely on an oversimplified view of loyalty rebates. The effects analysis is not a formality when considering certain by-object restrictions – such as loyalty rebates – it is a crucial aspect of the assessment. While by-object restrictions can be presumed to be capable of restricting competition, it does not follow that they constitute a per se infringement. The presumption is rebuttable. If the dominant company has evidence casting doubt on whether the effects of a by-object restriction do actually result in a restriction of competition, the Commission must prove those effects.

In its decision to refer the case back to the General Court (GC), the Court of Justice noted that, although – per Hoffmann-La Roche – loyalty rebates are presumed to be an abuse of a dominant position under Article 102 TFEU, the Commission must take into account evidence presented by the dominant company showing that its conduct was not capable of causing the alleged anti-competitive foreclosure. In the GC’s view, the Commission failed to do so in its investigation of Intel.

The Court of Justice set out the applicable criteria in its 2017 judgment (see our brief here), which is consistent with the Commission's own guidelines on Article 102 TFEU. The Commission must consider:

1. The extent of a company’s dominant position;
2. The share of the market covered by the practice;
3. The conditions and arrangements for granting the rebates;
4. Their duration and amount; and
5. Whether there was an exclusionary strategy in place aiming to exclude from the market competitors that are at least as efficient as the dominant undertaking.

In the GC’s view, on re-examining the facts, these standards were clearly not met in the Commission’s original 2009 decision. The Commission failed to determine the share of the market covered by Intel’s rebates scheme and the duration. Moreover, although the Commission carried out an as-efficient-competitor (AEC) analysis and took the results into account to show that the behaviour resulted in negative effects, it did not take the results into account in concluding that these practices were in fact abusive. The GC therefore concluded that the Commission made an error of law in concluding that because the rebates were a by-object infringement, it did not need to consider their effects to find an infringement. As a result, the part of the fine related to exclusivity/quasi-exclusivity rebates was annulled – however, because the GC could not identify the portion of the fine that related to the naked restrictions (which was not annulled), effectively the entire fine was annulled.

Burden and standard of proof

The Commission’s AEC analysis (or lack thereof) was a key element of the case. Although the Commission carried out an AEC analysis, it claimed it was not necessary and did not rely on it in its findings on whether Intel abused a dominant position. But Intel did conduct its own AEC analysis that showed that the AEC test ran by the Commission was flawed and that its conduct was not preventing other as-efficient companies from competing.

In this respect, the GC stated that any doubt in the mind of the court must operate to the advantage of the dominant undertaking and against the Commission. The court cannot conclude that the Commission has established the infringement if it still entertains any doubts. By providing its own AEC analysis, Intel demonstrated that the Commission’s AEC test was riddled with flaws and gaps, which did not make it possible to establish to the requisite legal standard that the rebates at issue were capable of having anticompetitive effects.

Implications and looking forward

This judgment shows the high standard of proof the Commission must meet when finding infringements of Article 102 TFEU. In a departure from previous case law such as Tomra, the GC has restated the principles set out in the Commission’s 2009 Guidelines that Article 102 TFEU does not allow for a black and white assessment of by-object infringements. Assessing whether a dominant company’s conduct did in fact result in restrictions of competition remains a nuanced assessment for which the Commission must take into account all relevant evidence. It also stands to reason that this approach would extend to other forms of abusive behaviour, not just rebates – although the GC has made clear it will not apply to ‘naked restrictions’ for which there is no plausible justification.

This judgment stands somewhat in contrast with the approach proposed in the new Digital Markets Act (DMA). The DMA proposal sets out quantitative thresholds based on turnover or market value and user reach that the Commission can use as a basis to identify presumed gatekeepers – companies with substantial market power active in certain digital sectors (e.g., online marketplaces, search engines, social networks, video-sharing platform services, communication services, and others who have substantial market presence in the EU). Whereas the DMA would – at least in the first instance – grant the Commission wide-ranging power to identify gatekeepers and oversee their compliance with associated obligations, the GC’s Intel judgment ensures that, when it comes to ultimately enforcing competition law rules against gatekeepers or dominant companies in other sectors, the Commission is held to a high standard of proof.