Skip to main content

Greater Than the Sum of Its Parts: the Jurisdictional Reach of EU Competition Law After Intel

Related Lawyers: David Lawne, Antonio Delussu
Related Practice Areas: Competition Disputes

Speaking at a recent event in Washington, D.C.,[1] the EU Commission antitrust chief, Margrethe Vestager, hailed the recent Intel decision by the European Court of Justice (“ECJ”)[2] as a ‘clear win’ for the territorial reach of EU competition law. In its decision, the ECJ confirmed the “qualified effects” principle, making conduct outside the EU subject to the jurisdiction of EU competition authorities when it is liable to have an effect within the EU.

Does this ‘win’ signal a move towards a more robust extra-territorial enforcement of EU law? Assessing jurisdiction depends on the specific facts of each case, and this article makes the point that some of the dominant company abuse lessons from Intel are not easily transferable to different contexts, such as cartel conduct. Nevertheless, Intel is a significant step in this direction. Anti-competitive conduct is often global in nature, and this decision underlines that firms cannot escape the reach of EU competition authorities simply because the relevant conduct took place outside the EU.

The background

In 2009, the European Commission (“Commission”) imposed on Intel a fine of more than €1 billion for abuse of dominant position under Article 102 of the Treaty on the Functioning of the European Union (“Article 102”). The Commission found that Intel was dominant in the market for a specific type of central processing unit (“CPU”), sanctioning Intel for its conduct relating to five major original equipment manufacturers (“OEM”) aimed at excluding Intel’s only significant competitor, AMD. Intel implemented this strategy by: (1) granting the OEMs ‘exclusivity’ rebates, conditional on the purchase of all, or nearly all, of their CPUs from Intel; and (2) imposing so-called “naked restrictions,” payments to the OEMs so that they would delay, cancel or restrict the marketing of products equipped with rival CPUs. According to the Commission, both types of conduct were part of an “overall strategy”[3] intended to weaken AMD globally, including within the EU.[4]

Intel appealed the decision to the European General Court[5] (“EGC”), but the appeal was dismissed in 2014[6]. Intel then appealed the EGC’s decision before the ECJ, the EU’s highest court. On 6 September 2017, the ECJ allowed Intel’s appeal and referred the case back to the EGC. The ECJ found that the EGC should have addressed Intel’s arguments about the economic impact of its conduct. But Intel also contended that the Commission did not establish its territorial jurisdiction to apply EU competition law in respect of agreements entered into with one of the OEMs, Lenovo, a Chinese company. On this specific ground Intel failed, the ECJ siding with the lower court.

The implementation test

EU competition law does not have unlimited territorial scope. In line with the rules of public international law, when conduct occurs outside the EU, or involves entities not domiciled within it, a sufficiently close link with the EU must be established. Two tests have developed in EU case law to help assess the strength of this link: the ‘implementation’ test and the ‘qualified effects’ test.

According to the implementation test, first set out in the Woodpulp decision,[7] and based on the principle of territoriality, conduct will be captured only if it is implemented within the EU. Considering the place of implementation was offered as a better alternative than merely considering the place of formation of an anticompetitive agreement[8]:

If the applicability of prohibitions […] were made to depend on the place where the agreement […] was formed, the result would obviously be to give undertakings an easy means of evading those prohibitions. The decisive factor is therefore the place where it is implemented.

While the EGC discussed at length the implementation test in its Intel decision,[9] it only did so “for the sake of completeness,”[10] and thus the ECJ did not deem it necessary to comment on the test, focusing instead on the effects test. Nonetheless, the implementation test remains an established feature of the jurisdictional assessment of EU law.

The qualified effects test: an alternative

The second approach to jurisdiction is based on the effects of anticompetitive conduct in the EU. Commentators have long debated whether the effects doctrine is an alternative rather than simply a corollary to the implementation test. The ECJ settled the argument in its Intel decision, confirming that satisfying either test is sufficient to establish jurisdiction.[11] This opens up the possibility of EU competition law being applied to conduct formed and implemented outside the EU, when it is liable to have an anticompetitive effect in the EU market. An example would be a cartel of Japanese widget producers refusing to supply customers within the EU. It is hard to say that the cartel was implemented in the EU, but it is easy to see how such conduct could have anticompetitive effects within it.[12] In practice, however, it is likely that the Commission (and claimants relying on EU competition law) will err on the side of caution, and continue to buttress their cases by applying both tests.

The 1999 Gencor decision[13] set the parameters for the effects test: the effects must be foreseeable, immediate and substantial. In Intel, the ECJ confirmed that there is no requirement to show actual anticompetitive effects, but only that such effects be “probable.”[14] The reason for this is public policy. As the EGC put it: the Commission “cannot be expected to adopt a passive position” when competition is threatened, and must therefore intervene, where possible, before a threat materializes in actual effects.[15].

Applying the parameters is a highly fact-sensitive exercise, so it is worth summarizing, first, how Intel and Lenovo operated. Intel is based in the US and Lenovo is a Chinese company. Intel sold Lenovo the CPUs in Asia subject to the exclusivity rebates. Lenovo then incorporated them into its notebooks, which it manufactured in China and sold, amongst other places, in the EU.

Was it foreseeable that Intel’s conduct would have an effect on competition within the EU? Both the EGC and the ECJ concluded that Intel not only could foresee, but also intended an anticompetitive effect on competition globally—including in the EU.

The second requirement is that the anticompetitive effects must be immediate. The ECJ confirmed the EGC’s findings, approving its focus on the overarching nature of Intel’s conduct. It is true that Intel did not sell its CPUs to Lenovo in the EU (it only did so in Asia) and, similarly, that no rebate payments were made in the EU. However, that was beside the point. The focus was not on sales of Intel CPUs to the OEMs, nor, for that matter, on the sales of AMD-based computers which were delayed or cancelled. The effects under scrutiny were those of the agreements between the OEMs and Intel that formed, together, the “overall strategy” to weaken AMD globally, including in the EU.[16]

Similar considerations apply to the requirement that the anticompetitive effects within the EU must be substantial. In both appeals,[17] Intel argued that the number of products affected in the EU was limited – a fact acknowledged by the EGC.[18] Therefore, Intel reasoned, its conduct vis-à-vis Lenovo could not have a substantially adverse impact in the EU. However, the ECJ ruled that the analysis should not be limited to dealings between Intel and Lenovo. Instead, “the conduct of the undertaking or undertakings in question, viewed as a whole” should be considered[19]. Endorsing the EGC’s judgment, the ECJ commented:

It suffices, in that respect, to note that the General Court held that Intel’s conduct vis-à-vis Lenovo formed part of an overall strategy aimed at foreclosing AMD’s access to the most important sales channels […..].[20]

What should be made of the reference to conduct “viewed as a whole”? Referring back to the EGC decision,[21] the ECJ took account of the totality of the agreements with the OEMs, emphasising that Intel cannot consider the agreements with Lenovo in isolation. To do so, the ECJ ruled, would lead to an “artificial fragmentation of comprehensive anticompetitive conduct” – which has the potential to substantially affect the EU – “into a collection of separate forms of conduct,” which might escape the EU’s jurisdiction.[22] The ECJ’s view, therefore, was grounded in the overarching public policy concern of protecting the structure of competition within the EU.[23]

Ms Vestager was correct in calling the ECJ Intel decision a “win” for the Commission. It confirms that the qualified effects test is applicable to abuse of dominance cases and clarifies how the test should be applied.


Ultimately, establishing jurisdiction will remain a highly fact specific question. In the context of cartels, this means consideration of sales and the extent to which it can be shown that pricing of products or services within the EU was affected by the cartel. This can be a complex exercise, given the global nature of many cartels and the potential for long supply chains crossing several jurisdictions.

While the ECJ Intel decision does not directly assist with this exercise, it is nevertheless a helpful decision as it underlines that, where anticompetitive conduct is global in nature, it should be viewed as a whole when assessing its effects within the EU. It should therefore be easier for claimants in cartel damages cases to establish jurisdiction when it can be demonstrated that there was an overall global strategy with probable anticompetitive effects within the relevant jurisdiction.


[1] Speech before the American Enterprise Institute, Washington, DC, 18 September 2017. Reported by MLex in Newman, M. and Crofts, L. (2017, September 18). Intel ruling is ‘clear win’ for reach of EU law, Vestager says. Retrieved from on November 14, 2017.

[2] C-413/14 P Intel Corporation Inc v Commission (Intel). When reference is made to Intel, with no qualification, the reference is to this decision. To avoid confusion, when reference is instead made to the Intel General Court decision, specific reference to that court or to the ‘lower court’ is made.

[3] Intel, para. 52.

[4] The Intel decision refers to the ‘European Economic Area’ (‘EEA’) rather than the ‘EU’. The EEA comprises the EU Member States plus Norway, Iceland, and Liechtenstein. The Commission can seize jurisdiction, as it did in this case, over the EEA, and essentially the same provision against anti-competitive conduct apply both in the EU and the EEA.

[5] The European General Court (previously known as the Court of First Instance) is the first tier of the European Union’s courts. It is bound by prior rulings of the ECJ.

[6] T-286/09 Intel Corporation Inc v Commission [2014] ECR I-0000 (Intel EGC decision).

[7] C-89/85 A. Ahlström Osakeyhtiö and others v Commission [1988] ECR 1-5913 (Woodpulp).

[8] Woodpulp was decided in the context of Article 101 of the TFEU, which deals with anticompetitive agreements (cartel conduct), but the implementation test has also been used in the context of abuse of dominance cases under Article 102.

[9] Intel EGC decision, paras. 298-314.

[10] Ibid., para. 298.

[11] Intel, para. 46.

[12] This example was suggested in Whish and Bailey, Competition Law, (8th edn, 2015), pp. 521 and 528.

[13] T-102/96 Gencor Ltd v Commission [1999] ECR II-753 (Gencor). Gencor was a merger case, but its reasoning on jurisdiction remains valid in relation to Articles 101 and 102.

[14] Intel, paras. 51-52. The concept is expressed even more clearly in the GC decision, Intel GC decision, paras. 251-252.

[15] Intel EGC decision, para. 252.

[16] Intel, para. 52

[17] See Intel EGC decision, paras. 285-7 and Intel, para. 48.

[18] Intel EGC decision, para. 290, to be read in conjunction with paras. 267-272. The EGC’s reasoning on the relation between the concept of a single and continuous infringement and the qualified effects doctrine is further explained at paras. 267-270.

[19] Intel, para. 50.

[20] Ibid., para. 55.

[21] Intel EGC decision, para. 268-270.

[22] Intel, para. 57. The same idea is expressed in Intel EGC decision, para. 270.

[23] This echoes the reasoning behind the genesis of the implementation test in Woodpulp. In that case the Commission was concerned with depriving wrongdoers of “easy means” to evade the EC’s jurisdiction. 

*David Lawne is a senior associate and Antonio Delussu is an intern in the London office. 

Related Lawyers