Written by Stephen C. Tupper

On May 13, 2009, the European Commission (the “Commission”) published a summary of a decision1 confirming its view that Intel Corporation was guilty of serious anti-competitive conduct during a period from 2002-2007 in an important microprocessor market and, consequently, the Commission imposed a massive €1.06 billion fine on Intel. (A non-confidential version of the decision was published on September 21, 2009).

The Commission’s decision is the culmination of nine years of work and two separate, but related, investigations. It is a notable event for two reasons: (i) the eye-catching size of the fine, the largest ever imposed by the Commission on a single company for an antitrust violation; and (ii) because it is the result of a major Commission Article 82 investigation, the rarely used — at least when compared to its Article 81 anti-cartel counterpart — abuse of dominant position, or anti-monopoly, provision in the EC Treaty. It also constitutes a continuation of the Commission’s unofficial policy of reserving the use of Article 82 for “big fish” in “high profile” markets.

Intel’s reaction to the decision has been understandably aggressive. On July 22, 2009, Intel appealed to the Court of First Instance, seeking an annulment of the Commission’s decision.2 From the public pronouncements of both sides, it is clear that the ensuing judicial encounter is going to be as hard fought and intense as anything before it.

This Alert examines the Commission’s decision. In turn it: (i) describes the alleged practices the Commission identified as being anti-competitive; (ii) summarizes the Commission’s legal rationale; and (iii) outlines the battle lines for the upcoming judicial review.

Alleged Anti-competitive Practices

The product market at issue is the sale of central processing units (CPUs) of what is referred to as the “x86 architecture.” CPUs are colloquially considered the “brains” of a computer and are of crucial importance to the manufacturing of data processing equipment. The term “x86 architecture” harks back to the time when IBM effectively set the technical standards for personal computers (PCs). It is industry jargon to describe the CPU of choice for PCs and similar-technology. Sales to several original equipment manufacturers (OEMs), such as Dell and Lenovo, and to one retailer, a German-based company called Media-Saturn-Holding GmbH, were scrutinized. Of specific interest to the Commission was Intel’s pricing policies and the manner in which Intel discounted from what might be considered its standard list prices.

The Commission grouped Intel’s allegedly illicit conduct under two categories: conditional rebates and so-called “naked restrictions.” In the first category, the Commission took issue with the manner in which Intel allegedly rewarded its customers for buying most, or all, of their needs from Intel. Those rewards came in the form of price discounts. The level of rebate granted, so argues the Commission, was de facto conditional on customers using Intel CPUs exclusively (e.g., as in the claimed case of Dell, Intel’s largest customer).

Under the second heading, the Commission focused on payments allegedly made by Intel to some of its major customers, e.g., Acer, Lenovo — to reward them for not launching, or postponing the launch of, products that contained CPUs supplied by Intel’s principal competitor, Advanced Micro Devices (AMD). In these instances, according to the Commission, Intel adopted a more targeted approach, singling out specific products or sales channels. For example, the Commission referred to a desktop product for which the manufacturer intended to have a dual-source purchase strategy (both AMD and Intel). According to the Commission, on hearing of the manufacturer’s plans, Intel intervened and forced the manufacturer to agree not to sell the AMD version of the desktop to mainstream business customers, and to delay an AMD European launch by six months.

The Commission’s Legal Rationale

In simple terms, to make its Article 82 case stick, the Commission has to be able to: identify the relevant market; show that Intel was dominant in that market during the relevant time period; and demonstrate that what Intel did, while so dominant, constituted an abuse of its dominant position.

The Commission took a relatively relaxed view of the finer points of market definition, concluding that although CPUs for x86 architecture was a separate market from CPUs for non-x86 architecture, the issue of whether the former market split into separate markets for desktops, laptops and servers could be left open without it doing mischief to the legal foundations of its final decision. The Commission’s confidence in that regard was based in part on the fact that, from a review of the available data, Intel appeared to have an 80 percent share of the overall x86 CPU market, and market shares in excess of 70 percent for the various segments.

The Commission used its market share data to conclude that Intel was dominant in the relevant market during the entirety of the relevant period. In doing so, the Commission also made reference to high barriers to entry (e.g., the nature and size of sunk investments, in areas such as R&D, coupled with capacity restraints), the strength of Intel as a brand that makes its products “must stock” for most OEMs. It dismissed Intel’s submissions relating to falling prices and countervailing buyer power.

As to the abuse requirement, the Commission relied on the foreclosure effect of Intel’s alleged fidelity rebates and “naked restrictions.” Contending that the provision of certain types of rebates can constitute an “abuse” is relatively uncontroversial. In its decision, the Commission relied on long-standing case law3 that prohibits dominant suppliers from entering into requirements contracts with customers or gearing prices in

such a way that customers are rewarded for taking 100 percent, or close to 100 percent, of their needs from the dominant supplier. Where the Commission and Intel are at loggerheads in this regard is whether the Commission is required to demonstrate that the rebates actually led to market foreclosure. Intel has argued that its rebates did not foreclose the market to AMD, pointing to data that showed that AMD’s market share grew during the relevant period. The Commission rejected this argument, contending that the case law does not require evidence of total foreclosure for a violation. The Commission took note of evidence that suggested that AMD’s product was technically more “desirable,” and concluded, with the aid of econometric modeling, that Intel’s conduct in fact had resulted in significant foreclosure of an “efficient” competitor.

The Commission’s “naked restrictions” line of reasoning was quite novel. Although it sought to rely on existing case law4 as precedent, the Commission clearly attempted to break new ground. Although the term “abuse” is not a static concept that can be reduced to a closed list of anti-competitive conduct under Article 82, the Commission does not normally stray far from established legal/economic logic. Here it stated that Intel’s payments to the OEMs resulted in consumers being denied choice, or at least had their right to choose artificially retarded by Intel’s conduct. According to the Commission, this in turn prevented AMD from gaining traction in the market, and damaged its ability to compete effectively to the detriment of consumers.

Finally, as fall-back reasoning, the Commission suggested that the abuses should not be viewed in isolation. Each, it argued, was simply an example of abusive conduct that formed part of a single course of conduct by Intel aimed at foreclosing AMD’s access to the relevant market.

Issues for Judicial Review

It appears that the key battle grounds on appeal will be both evidential and procedural. Intel likely will argue that the Commission got the wrong end of the stick and, somewhat more controversially, deliberately so. The facts, Intel has contended, speak for themselves. This is a highly competitive market in which prices fell by 36 percent, and AMD increased its market share from 8 to 22 percent over the period of the Commission’s investigation. Whatever the Commission thinks was going on, so Intel likely will contend, is a mirage and a figment of the Commission’s imagination. For its part, the Commission doubtless will contend that simply because Intel’s foreclosure strategy failed in its objective to eliminate AMD totally does not mean that Intel is innocent.

The Intel case may be too market specific and data intensive to have significant precedent-setting potential. That being said, however, Article 82 cases before the Court of First Instance do not come along often, and they usually are worthy of note when they do. In this case, Intel’s first line of defense is that it was trying to compete “normally” in a tough market. If, ultimately, the Court opts to rule against Intel, what it then goes on to say about how dominant companies should behave in circumstances such as those that faced Intel will doubtless be compulsory reading for industry leaders around the world.

1Summary of Commission Decision of 13 May 2009 relating to a proceeding under Article 82 of he EC Treaty and Article 54 of the EEA Agreement (Case COMP/C-3/37.990 – Intel), OJ C227/13, 22.9.2009. A full non-confidential version of the infringement decision is available on the Commission’s website. up
2 See OJ C220/41, 12.9.2009. up
3See e.g. Case 85/76 Hoffmann-La Roche [1979] ECR 461. up
4See e.g. Case T-228/97 Irish Sugar [1999] ECR II-2969. up