In the last decade, cases such as Microsoft and GE/Honeywell have seen the European Commission (Commission) and the US antitrust authorities reaching  substantially different outcomes on the same or similar factual bases. But the US leg of Intel’s global competition law saga suggests that US antitrust enforcement policy as to unilateral conduct might be moving closer to the Commission’s position. Is this an exception, or a sign of things to come?

“New” US Antitrust Policy—Much Ado about Nothing?

During his election campaign, President Obama vowed to “reinvigorate antitrust enforcement, which is how we ensure that capitalism works for consumers.1  He promised “an antitrust division in the Justice Department that actually believes in antitrust law. We haven’t had that for the last seven, eight years.”2  Christine Varney, President Obama’s choice as Assistant Attorney General in charge of the Antitrust Division of the US Department of Justice (DOJ), echoed this rhetoric when she withdrew the prior administration’s policy statement on Section 2 of the Sherman Act.  Announcing “a shift in philosophy,” Varney said that “the Antitrust Division will be aggressively pursuing cases where monopolists try to use their dominance in the marketplace to stifle competition and harm consumers.”3

These assertions of vigorous enforcement indicated that faith in the markets’ ability to self-correct was no longer a guidepost for US antitrust policy. The New York Times quickly recognized the new policy as being more closely aligned with that of the Commission,4 so did a Wall Street Journal commentator, who called the realignment “a huge mistake.”5  Yet, more than one year later, the DOJ has yet to bring a case demonstrating its break with the previous administration’s Section 2 policy. The situation is different, however, at the Federal Trade Commission (FTC). 

In suing Intel last December exclusively under Section 5 of the FTC Act (Section 5), which prohibits “unfair methods of competition,” the FTC moved beyond what DOJ could ever do under Section 2 (which requires plaintiffs to show conduct fundamentally inconsistent with competition on the merits). As Intel’s discounting presumably profited its computer manufacturer customers, some US courts would have been skeptical of a Section 2 challenge.  Freed from the Sherman Act, the FTC adopted what amounts to a Commission-style abuse-of-dominance theory.

The Intel Saga

By its own account, Intel has sold between 70 and 85 percent of the x86 microprocessors (also called central processing units or CPUs) for use in computer systems.  Intel has competed aggressively, in particular with its principal competitor, Advanced Micro Devices (AMD).  The result has been a series of enforcement actions by several competition authorities and private plaintiffs even before the FTC filed suit:

  • In 2005, after the Japan Fair Trade Commission ruled that Intel had abused its monopoly power, Intel accepted a “cease and desist” order.
  • Also in 2005, AMD sued Intel in US federal and Japanese courts; the case settled in November 2009 with Intel paying AMD $1.25 billion. 
  • In 2008, the South Korea Fair Trade Commission fined Intel $26 million for offering rebates to personal computer makers in return for not buying competitors’ CPUs. 
  • In May 2009, the European Union (EU) slapped Intel with a €1.06 billion fine for abuse of dominance, the largest abuse-of-dominance fine handed out to date under EU competition law.
  • Just as the AMD case was settling, the New York State Attorney General beat the FTC to the punch, suing Intel in US federal court for violations of Section 2 of the Sherman Act and the corresponding New York State antitrust statute, characterizing Intel’s discounting as “bribery” of its original equipment manufacturing (OEM) customers.
  • The following month, the FTC filed its administrative suit (the Complaint).

After this global series of cases that seemingly addressed the same basic concern, what did the FTC’s case add? 

In the EU case, Intel received a hefty fine and was also required to cease the following specific practices:

  • Rebates given to computer manufacturers on the condition that they bought all or almost all, of their CPUs from Intel
  • Payments to a major retailer for stocking only computers with Intel CPUs
  • Direct payments made to computer manufacturers to halt or delay the launch of products containing a competitor’s CPUs.

The fact that AMD was still able to compete and innovate was insufficient to negate Intel’s so-called abuse of dominance: from an EU perspective, Intel’s rebate practices led to less choice for consumers and prevented AMD from competing on a level playing field.  This led to a determination that Intel had abused its dominant position.

Given that the Commission and the FTC kept “each other regularly and closely informed on the state of play of their respective Intel investigations...and shar[ed] experiences on issues of common interest,”6 it is not surprising that many of the FTC’s allegations were reminiscent of the European case.  But the scope of the FTC’s case exceeded that of the European Union.  The FTC’s complaint reached beyond CPUs; the FTC staff determined that Intel had also sought to derail competition from makers of graphics processing units (GPUs). Also, the range of conduct the FTC found “unfair” extended well beyond what the Commission identified.  Albeit in language less vivid than the New York Attorney General’s, the FTC complaint enumerated a broad range of allegedly illegitimate tactics meant to keep Intel’s competitors’ CPUs and GPUs from reaching end-users. These included:

  • Using market-share discounts that prevented customers from buying more than a set percentage of their CPUs from Intel’s rivals;
  • Using volume discounts and bundled discounts (discounts on one product predicated on purchase of another product) that, in the FTC’s view, amounted to below-cost pricing;
  • Threatening customers with the loss of benefits such as discounts, technical support, guaranteed supply and patent liability indemnification if they bought any CPUs from competitors;
  • Inducing computer manufacturers that bought competitive CPUs to agree to use suboptimal means of distribution for the computers containing those CPUs;
  • Designing its software compiler to generate object codes that ran more slowly on competitors’ CPUs, attributing the performance difference to the competitors’ CPUs rather than its compiler, and allowing industry benchmarks to be developed based on the compilers’ work, unfairly damaging the competitiveness of rival CPUs;
  • Encouraging Nvidia, the leading GPU maker, to develop GPUs compatible with Intel CPUs, and then curtailing interoperability once Intel saw the GPUs as potential substitutes for the CPUs themselves; and
  • Delaying standards development in order to skew the standards in its favor.

The FTC framed this conduct in Section 2 terms, alleging that Intel maintained monopoly power in the relevant x86 CPU market and that it attempted to monopolize a GPU market. But what the FTC complaint describes more than anything else is a rough, high-stakes competition between Intel and AMD, whose technological advances and aggressive marketing forced Intel to respond in kind.  This could have meant trouble for a claim guided by Sherman Act principles.  But Section 5 allows the FTC to “consider…public values beyond simply those enshrined in the letter or encompassed in the spirit of the antitrust laws.”7 

The FTC’s pure Section 5 gambit, untethered to Sherman Act standards, has attracted criticism.  As one commentator put it, reliance on Section 5 evaded “the strict requirements of proof of competitive harm embedded into Section 2 of the Sherman Act.8 It was not lost on observers that AMD’s survival and, indeed, its own aggressive competitive responses to Intel (including continuing innovation), made it difficult to discern the consumer harm that Section 2 would have demanded. More generally, the notion of unfairness is in tension with traditional US antitrust law theory. In contrast, unfairness appears explicitly in EU competition law:  in Article 102 of the Treaty on the Functioning of the EU (TFEU), “unfair” describes a type of prohibited conduct.  Moreover, in Michelin II, the Commission expressly characterized Michelin’s rebate system as unfair to dealers.9 

Liberated from the Sherman Act’s constraints, the FTC sought even more extensive relief than the Commission had obtained, ranging from prohibitions of the alleged misconduct to affirmative mandates on interactions with customers and competitors. Notably, though, the FTC did not seek to block pure volume-based discounts; its core objection evidently was to the use of commitments that locked Intel’s customers into limits on the chips that they would buy from Intel’s rivals.

This carve-out for volume discounts is a key element of the relief that Intel agreed to entering a Consent Order to settle the FTC case. The Consent Order also explicitly allows Intel to win all of a customer’s business if the customer has asked Intel to bid for it, and to enter into exclusive agreements with customers with which it has invested significantly in joint product development.  The prohibitions FTC sought on “near exclusiv[ity]” are gone as well.  Even so, the Consent Order does contain much of the relief that the FTC sought.  Intel may not:

  • Condition discounts and other customer benefits on exclusivity or on limitations on purchases of competitors’ chips.
  • Use bundled discounts or retroactive discounts that would yield below-cost pricing under the test the Ninth Circuit adopted in Cascade Health Solutions v. PeaceHealth.10 
  • Change any of its products so as to degrade competitive products without an improvement in the Intel product.

 In addition, Intel must:

  • Take extensive steps to remedy the compiler issue the FTC identified, including reimbursement of compiler customers for remedial modifications of their software
  • Maintain interoperability for six years through a Standard PCI bus interface, and provide an annual “interface roadmap” to Nvidia, its GPU competitor
  • Assist its competitors by (i) amending its licenses to allow disclosure of certain license rights to third-party foundries and customers, and (ii) restraining its ability to enforce patent rights against them after a change in control.  

These affirmative duties are highly unusual in the United States:  they appear to dictate conduct that would seem to undercut Intel’s own incentives to invest in innovation.  Ordinarily, for example, one would not expect a firm to face antitrust liability for terminating its licenses to competitors that merge with customers to which the firm has disclosed competitively sensitive information.  But in forcing Intel to modify its license terms, the FTC Decree seems to contemplate that it is better for competitively sensitive information to fall into the hands of Intel’s competitors, despite the obviously anticompetitive potential, than to allow Intel to terminate a licensee. 

Similarly, a requirement that a firm maintain existing interfaces for a period of time appears to discourage innovation. The FTC, by its own account, “is concerned that Intel’s past conduct has weakened AMD and Via [a Taiwan-based x86 producer].”11 One might have expected the Commission to obtain relief like this; instead, it is the FTC that has gone further—favoring open competitive access to Intel’s products over continued Intel innovation.  But as the FTC itself has said, the Consent Order’s terms “do not necessarily reflect the applicable legal standards under the Sherman Act, Clayton Act, or the FTC Act; indeed, the legal standards applicable to some of these practices remain unsettled by the Supreme Court and the federal courts of appeal.”12

Whether these legal standards “remain unsettled,” or perhaps simply remain at odds with the FTC’s policy views, the FTC-Intel settlement means that, for now, the US courts will not put the FTC’s theories to the test.  But the FTC is no doubt looking to develop more cases in winner-take-all, high-tech markets.  If it succeeds, the FTC is virtually certain to again utilize Section 5 rather than be limited, as the DOJ must, by the Sherman Act’s rigors.  Eventually, then, the FTC may force the issue as to whether US antitrust law, and not just one enforcement agency, is converging with EU competition law.  In the meantime, it remains to be seen if the AMD and FTC settlements will have any bearing on Intel’s plans for its appeal against last year’s EU decision.


1. See
2. “Obama eyes media with promise of antitrust push,” Reuters (May 18, 2009), available at
3. US Department of Justice Press Release, “Justice Department Withdraws Report on Antitrust Monopoly Law” (May 11, 2009), available at
4. Stephen Labaton, “Obama Takes Tougher Antitrust Line,” the New York Times (May 12, 2009), available at
5. George Priest, “The Justice Department’s Antitrust Bomb,” the Wall Street Journal (June 2, 2009), available at
6. European Commission Press Release, “Antitrust: Commission imposes fine of 1.06 billion euros on Intel for abuse of dominant position; orders Intel to cease illegal practices—questions and answers” (May 13, 2009), available at
7. FTC v. Sperry & Hutchinson Co. 405 U.S. 233, 244 (1972).
8. Josh Wright, “The Case Against the Section 5 Case Against Intel” (January 7, 2010), available at
9. Commission Decision 2002/405, 2001 O.J. (L 143) 1.  The General Court upheld the Commission’s decision. Case T-203/01, Manufacture Française des pneumatiques Michelin v. Commission, 2003 E.C.R. II-04071.    
10. 515 F.3d 883, 903 (9th Cir. 2008).
11. Analysis of Proposed Consent Order to Aid Public Comment, In re Intel, No. 9341 at 8 (F.T.C. August 4, 2010) (Intel Analysis), available at
12. Intel Analysis at 2.