An extract from The Intellectual Property and Antitrust Review, 5th Edition
Standard-essential patents
Collaboratively setting standards for technologies is important to a variety of industries, as the practice facilitates the creation of follow-on inventions while promoting future innovation and interoperability of products. However, standard setting may also raise antitrust issues, such as when a patent gains additional market power when the patented technology becomes adopted as part of an industry standard and the patent becomes a SEP. In such an event, the patent's value may increase because it can be used to block implementation of the standard, and the patent owner may attempt to exploit this newly acquired leverage by refusing to license the standard-essential technology unless licensees agree to excessively high royalties (a practice commonly referred to as 'patent holdup'). In essence, the adoption of patented technology into a standard may confer market power on the patent holder that would otherwise not have existed, and that market power may then be abused if not constrained. If many holders of SEPs engage in patent holdup, the aggregate royalties for patents essential to a given standard may actually be greater than the value of the actual feature, or indeed the product incorporating the feature.
In response to these issues, and to minimise exploitative licensing, many SSOs ask that patent holders who wish to have their technology considered for incorporation into the industry standard voluntarily pledge to license their patents on FRAND terms. FRAND commitments serve several salutary purposes: linking patent holders with those who incorporate and implement patented technologies; ensuring royalties more closely reflect the actual value of the patent by minimising the surplus attributable solely to adoption of the patented technology into the standard; and incentivising patent holders to put forth their best technology to be standardised. Of course, SSO participation is voluntary and SSO FRAND principles address only the constraints on participating patentees. Other doctrines, including the fact that the patent statute allows for damages in an amount of a 'reasonable' royalty, have been used to curtail royalty demands based on factors other than the actual value of the patented technology to the allegedly infringing device.
SSOs typically do not 'set' a FRAND rate or other terms. In most cases, the SEP holder and licensee attempt to agree on what would constitute FRAND terms and then execute a licence agreement reflecting those terms after successful negotiations. However, the FRAND commitment to the SSO generally is considered a binding commitment that runs with the patent, and may be enforced by a prospective licensee as a third-party beneficiary of the SSO commitment. A breach of a commitment to license on FRAND terms may constitute grounds for an antitrust suit by one injured as a result of the breach. When it comes to policing compliance with FRAND commitments, US antitrust law operates more as a backstop than a primary check on SEP licensing, with the Agencies primarily acting as competition advocates that assist SSOs through a variety of Agency programmes to craft clear guidance for patent holders. Indeed, the Agencies generally refrain from intervention in private licence negotiations unless a patent holder's failure to honour its FRAND commitment is deliberate and rises to the level of an antitrust violation under Section 2 of the Sherman Act. Indeed, former Assistant Attorney General Bill Baer in a September 2015 speech at the 19th Annual International Bar Association Competition Conference emphasised that the Agencies, in line with their European counterparts, will not intervene in basic commercial disputes over royalty rates in the absence of bad conduct by the patent holder or improper use of market power.
Exactly what 'fair, reasonable and non-discriminatory' means in the licensing context, however, remains sometimes controversial and largely unsettled. Although jurists and agencies in Europe have been more active in defining both what constitutes a FRAND rate and the process parties should engage in to get there, currently there is less guidance in the United States. While several district courts – most recently in the 2018 and 2019 TCL decisions and the 2019 Ericsson decision – and one appellate court have directly addressed the issue, determinations of FRAND terms are highly fact-specific and may vary widely depending on the patents, standards and products at issue. Thus, the Agencies have been mostly silent on what specifically constitutes FRAND licensing, and the courts have avoided applying bright-line tests. Instead, US courts defining FRAND have considered the rates in 'comparable' licences for similar technologies and patents, the rates for patent pools with respect to standards, the objective value of the patent to the standard, the availability of alternatives and their quality, whether the patent covers the core features of the standard, the possibility of royalty stacking, and the number of essential patents that may read on a particular standard.
In 2015, the US Court of Appeals for the Federal Circuit decided Commonwealth Science and Industry Research Organisation (CSIRO ) v. Cisco Systems, a case involving certain technologies incorporated into the IEEE 802.11 Wi-Fi standard. The Federal Circuit held that, while the smallest saleable patent practising unit is one appropriate base for calculating damages, it is not the exclusive method for apportioning the value of a patent to the allegedly infringing product, especially if the record indicates that the parties actually negotiated a different base in reality. The CSIRO decision also emphasised that, in determining whether a given royalty is excessive, the actual cumulative royalty paid by the implementer must be proven and the implementer may not rely on abstract recitations alleging royalty stacking or qualitative assertions of the value of inventions without some quantitative evidence. The CSIRO decision's focus on actual quantitative evidence is thus consistent with the Federal Circuit's 2014 decision in Ericsson v. D-Link. D-Link involved the adjudication of whether Ericsson had complied with its obligation to license a patent essential to the IEEE 802.11 Wi-Fi standard on FRAND terms, and held that in determining FRAND damages, concerns about patent holdup must be proven using specific facts from the case, rather than on the basis of theory or general probability. In January 2019, the US District Court for the Eastern District of Texas held that, as a matter of French law, the FRAND commitment does not require a FRAND licence to be based on the smallest saleable patent practising unit. The final judgment in this case is currently on appeal in the Fifth Circuit. In contrast, the FTC v. Qualcomm court granted the FTC's motion for partial summary judgment in November 2018, holding that FRAND licensing obligations required Qualcomm to license its SEPs to competing modem chip suppliers. The same Court later found, in May 2019, that basing royalties on handset value, rather than on the value of modem chips, was contrary to the Federal Circuit's approach to royalty apportionment and that modem chips were the smallest saleable patent practising unit in cellular handsets. This case is currently on appeal in the Ninth Circuit.
