By Richard Vary, Bird & Bird

Takeaway 

Must owners of IP rights (IPR) that may be essential to standards offer licenses of their essential IPR to each component maker? Or do they satisfy their commitment to Standards Development Organisations (SDOs) by offering licenses to the end user product maker, which provide “have made” rights or other assurances for the component maker? Do they need to offer license terms in circumstances where they do not seek to enforce their patents?

We look at the wording of the IPR policies of various SDOs. We find that none of the IPR policies requires the IPR holder to offer licenses of its essential IPR to component makers. All of the policies would appear to permit holders of essential IPR to adopt a policy of licensing to the end-user product manufacturer. Many of the policies contain wording that could only be relevant or applicable when licensing to parties who buy in components from component makers, such as end-user product manufacturers.

Background 

High-technology products today usually contain or rely on standardised technology. Cellular wireless connectivity is one example; Wi-Fi and video encoding are others. These technologies are developed not by a single company, but by groups of companies active in the field that collaborate through standards-development organisations. Representative engineers from the companies meet and agree on technological solutions to be adopted in the standard. Often these meetings identify technical problems to be overcome. The solutions identified by the participating engineers to these problems may be novel, and inventive, and therefore eligible for patent protection. 

SDOs typically have Intellectual Property Right (IPR) policies under which participants in the standards–development activities may voluntarily commit to offer licenses to IPRs that they hold which may be or may become essential to the standard (“standards-essential patents” or “SEPs”) to future implementers of the standards, according to the terms of the IPR policy. In drawing up an IPR policy, SDOs seek to balance the multiplicity of interests affected by the standard under development. They take into account patent law and licensing norms in the industry. They seek to impose the bare minimum of requirements necessary to ensure that the standard may be implemented: actual commercial licensing terms are to be negotiated outside the aegis of the SDO. 

High technology products today also typically result from a supply chain, with tiers of component makers supplying components, including software, to the manufacturer of the end user product. Connected cars are an example of this: the connectivity technology which goes into the vehicle is supplied by a multi-tiered supply chain to the automotive manufacturer.

In recent patent cases involving connected cars, the question arises of whether a SEP owner: 

1. must offer licenses of their SEP portfolio to component makers, or 
2. may offer licenses to the end user product maker, which provide “have made” rights for the automotive manufacturer’s component makers^[1], or 
3. may offer licenses to the end user product maker, and refraining from asserting patents at the component level^[2].

“Have-made” rights arise in licensing agreements, where a patent owner grants the right to “make, have made, use, keep, sell etc.” The extent of “have made” rights depends on the contract. In the mobile phone industry they permit handset makers to buy in cellular radio components, modules or even entire finished devices from third party suppliers, without those suppliers infringing the licensed patents.

The first of these models is referred to by the shorthand of “license to all” because each person in the supply chain may request and receive a written license agreement. The latter two models are known as “access for all”, because each person in the chain has access to the technology (for example, to manufacture and supply products incorporating it), even though they may not themselves be party to a written license agreement.

There are competition or antitrust law arguments about both models. In the United States these have been addressed in some detail in the FTC v Qualcomm United States Court of Appeals Ninth Circuit decision, which is discussed further below. In Germany, courts in Munich, Mannheim and Dusseldorf^[3] have considered whether SEP owners abuse their dominant position by offering “value chain” licenses for the connected vehicle,^[4] or whether they can be compelled to offer “exhaustive”^[5] patent licenses to component makers. At the time of writing, the answers have been that SEP owners are not obliged by competition or anti-trust law to offer licenses to component makers, exhaustive or otherwise.

Anti-trust and competition law is, however, not the only relevant issue. When IPR holders give commitments to Standards-development organisations to  license on FRAND terms, this can create a contractual commitment. The UK’s Supreme Court in Unwired Planet reminded the industry that the wording, and the meaning, of those commitments is decisive. 

In a paper dated 8 June 2020^[6], the Fair Standards Alliance sets out selected quotes from a number of SDO IPR policies^[7] which are said to support the assertion that an essential IPR Owner must offer licenses of their SEP portfolio to each component maker. 

This paper analyses the various SDO IPR policy wordings to determine whether they do in fact require licensing at the component level or whether they permit the IPR owner to adopt an “access for all” model of licensing. We find that, as a matter of wording, the SDO policies looked at do not require licensing to take place at the component level.

ETSI RULES OF PROCEDURE, 4 December 2019

ETSI remains the primary SDO for cellular radio technology. It is one of seven Organizational Partners of 3GPP.^[8]ETSI transposes 3GPP work into ETSI “technical specifications”, to which its members may identify potentially essential patents. ETSI has the largest database of telecommunications patents identified in the world, and most 3GPP participants identify their potentially essential patents to the ETSI organization. 

ETSI has an IPR Policy, which takes effect as a contract between the IPR holders who make licensing declarations to ETSI, and ETSI itself. Article 6.1 envisages that IPR owners would be prepared to offer licenses to cover the following acts:

• Manufacture, including the right to make or have made customized components and sub-systems to the licensee’s own design for use in Manufacture; 
• sell, lease, or otherwise dispose of Equipment so manufactured; 
• repair, use, or operate Equipment; and 
• use Methods. 

The article defines “Manufacture” as “the production of Equipment”.  “Equipment” is “any system, or device fully conforming to a [standard]”. “Methods” are “any method or operation fully conforming to a [standard]”.

The ETSI IPR Policy therefore only requires the licensing of devices fully conforming to its standards. Those standards are made up of technical specifications, many of which describe interoperability between user equipment and a network: how a device should respond given certain inputs. They require the device to function in the network environment. 

It is typically only a finished end-user product which can implement an ETSI standard. A component may enable some, or even most, of this functionality. But a component by itself cannot transmit a specified response to an incoming radio signal. The same is true of the definition of methods: a component by itself typically cannot carry out a method conforming to the standard. 

A patent owner’s claims which described the interaction of a device within a network will generally only constitute “Essential” IPR when they are directly infringed or indirectly infringed. They may only be directly infringed once implemented in a network. They may only be indirectly infringed by an end-user product manufacturer when it supplies a product that it knows will put the patented invention into effect.  

The ETSI IPR policy also expressly mentions “have made” rights. In other words, it envisages a license to a product maker which allows him to source components from a component maker, and the consent to use the patents would cover the manufacture and supply of those components. That is consistent with a model of licensing at the end-user product manufacturer level. It is not consistent with a model where each component maker would have his own separate license. 

Finally, it should be noted that the ETSI IPR policy does not mandate that SEP owners grant licenses. It does not require a SEP owner to offer licenses. SEP Owners merely confirm in a licensing declaration whether they “are prepared to” offer. Why were those words chosen? One reasonable interpretation of these words is that the policy envisaged that SEP owners may choose not to set up a licensing program at all. In practise many SEP owners, including some major contributors to the standards, have chosen not to set up a licensing program, and do not offer licenses to their SEPs, on FRAND terms or otherwise, to anyone. Nor have they given any binding commitment not to enforce their patents. In over 20 years that the policy has existed, that behaviour has not been challenged by ETSI, by regulators or by other members: it is, at least implicitly, accepted. Against that background it would be difficult to construe the ETSI IPR policy to argue that these SEP owners are obliged to offer licenses to any particular class of licensees, and are in breach of contract because they have not done so.  

In conclusion, the ETSI IPR policy does not oblige SEP owners to offer licenses. SEP owners declare whether they are prepared to offer licenses in respect of functioning products, and industry practice has been for SEP owners who do operate a licensing program to offer licenses at the end user product maker level. 

The policy therefore permits an IPR owner to adopt an “access for all” model, and contains wording that is inconsistent with an obligation to license at the component level. 

ATIS 

The Alliance for Telecommunications Industry Solutions (“ATIS”) is the U.S. Organizational Partner of 3GPP, which promulgates the UMTS and LTE standards. ATIS’s IPR Policy expresses its intent in the first paragraph:

“it is the intention of ATIS and its Forums to benefit the public while respecting the legitimate rights of intellectual property owners.”^[9]

ATIS’s IPR Policy envisages that a holder of essential patent claims may give an

 “assurance that a license to such essential patent claim(s) will be made available to applicants desiring to utilize the license for the purpose of implementing the standard . . . under reasonable terms and conditions that are demonstrably free of any unfair discrimination.”

Section 3 of the ATIS Declaration uses the following wording: 

“The license assurance is provided for [… ] the claim(s) of the patent(s) […] covering an invention the use of which is required for compliance with the American National Standard”.

The wording restricts the class of potential licensees to applicants desiring to “use” the invention for “compliance with the standard” or “utilize the license for the purpose of implementing the standard”. ATIS deals with the same technical specifications as ETSI. Many of these technical specifications specify how a device must react in a network environment, for example what the device must transmit in reaction to an incoming signal, or specify how it should initiate an exchange of radio transmissions. It is typically only a functioning end user device which can implement these. 

For a component manufacturer to fall within the definition, it would be necessary to read “for the purpose of implementing” as covering implementation by a downstream customer of the applicant, rather than implementation by the applicant himself. It would be necessary to read “use” in section 3 as covering “use by a downstream customer of the applicant” rather than use by the applicant himself. Although that is not an inconceivable interpretation, it would stretch the wording beyond its literal meaning.

Two factors would suggest that this was not what was intended:

• There is a need for compatibility among IPR Policies of cellular SDOs. As a 3GPP Organizational Partner, ATIS is the US equivalent of ETSI. Both promulgate essentially the same standard, developed in the 3GPP framework. The 3GPP Working Procedures require that IPR policies of members of 3GPP are compatible^[10]. It is also desirable that IPR policies are aligned because this strengthens the 3GPP partnership, it promotes the consistent application of telecommunication standards, and because differences in the IPR policies may cause SDO forum shopping. So there are good reasons to prefer a reading of the ATIS policy which is consistent with ETSI. 
• ATIS is accredited by the American National Standards Institute (“ANSI”). The ATIS IPR Policy is based on the ANSI patent policy. In 2018 the ANSI Executive Standards Council Appeals Panel issued a formal decision interpreting the requirement to license applicants “for the purpose of implementing the standard”^[11] . The Panel stated that this wording “does not mean that ANSI’s Patent Policy requires licensing at the component level.”

Judge Koh, in an interim decision in FTC v Qualcomm^[12], took a different view. Judge Koh used extrinsic evidence of requirements of antitrust law to guide the contractual interpretation of the ATIS policy. To paraphrase the court’s reasoning: if the policy on a literal reading produced a result which violated anti-trust law, then the policy must be read in a different way, a way which avoided that violation. 

Judge Koh’s antitrust analysis was based around Intel being a direct competitor of Qualcomm in the market for cellular chips. The judge found that, in those factual circumstances, Qualcomm’s refusal to offer licenses to direct competitor Intel would be anti-competitive and discriminatory. 

There are two obvious problems with that finding. First, Judge Koh’s antitrust analysis assumed that even though Qualcomm had licensed Intel’s customers, Qualcomm could still use its patents to seek to exclude Intel from the market. It does not consider whether, under an “access to all” model, Intel might nevertheless have been free to operate under other mechanisms. 

Secondly, that antitrust analysis depended on the fact that Qualcomm and Intel were competitors in the chip market. It would not apply to SEP holders who were not also chip makers. Had the SEP owner in that case been a manufacturer of infrastructure equipment or handsets, then Judge Koh’s analysis would have led to a different conclusion as to the meaning and effect of the ATIS policy. Equally, if the person seeking a license had been at another level of the supply chain, Judge Koh’s analysis would have led to a different conclusion. This would lead to the somewhat undesirable result that the contractual wording of the ATIS policy means different things for different parties. 

In any event, Judge Koh’s decision in relation to antitrust law was overturned on appeal. As a matter of procedural law, this removes any precedential value the decision may have had. However, the overturning of the decision in relation to antitrust also entirely removes the reasoning behind Judge Koh’s contractual interpretation of the ATIS policy. If the literal meaning of the ATIS policy does not, in fact, lead to a result which violates antitrust law, then there is no longer any reason to depart from the policy’s literal meaning. Judge Koh’s contractual interpretation of the ATIS policy falls away. 

Consequently, the wording of the ATIS policy does not mandate component-level licensing. 

TIA

The Telecommunications Industry Association (“TIA”) is the U.S. Organizational Partner of 3GPP2, which promulgates CDMA2000.

The TIA IPR policy provides that SEPs “will be made available under reasonable terms and conditions that are demonstrably free of any unfair discrimination to applicants only and to the extent necessary for the practise of” a standard adopted by TIA^[13].

The same arguments apply as in respect of the ATIS policy. It is limited to applicants who intend to practise the standard. It should be read in a way which is consistent with the ETSI policy. Like ATIS, TIA is also accredited by ANSI, so ANSI’s decision that it does not require licensing at the component level applies equally to the TIA policy. But significantly, the TIA policy is arguably narrower than ATIS in two respects:

• The TIA policy uses the words “only and to the extent necessary for the practise of”. “Only and to the extent necessary” restrains any commitment to license to a greater degree than the wording of the ATIS policy. 
• “Practise” is narrower than “implement” in that it is more direct: it is a stretch to suggest that a person who supplies a component to another person, who makes a device which will practise the standard, is himself “practising” the standard. To use an analogy: a supplier of pharmaceutical drugs to the medical profession does not himself “practise” medicine, even where those drugs will be used in the practise of medicine.

The TIA policy also does not require the licensing of patents. It requires the licensing of individual patent claims, and only “to the extent necessary” to practise the standard. Patents typically contain equipment claims, and method claims. Even where a component may fall within an equipment claim, it will not fall within a method claim. The method claim only comes into play when the end user device is operating within a network, or when the end user device is sold to a customer who intends to use it. So, even if we could stretch the meaning of “practise” to cover the supply of components, only those claims actually infringed by the supply of components would be covered by the licensing obligation. That would not include the method claims. The manufacturer of the end user product, which would infringe those method claims, would therefore also itself need a license.

In conclusion, the TIA IPR policy does not require or oblige licenses to be granted at the component level. Even if it did, the end user product maker would still need a separate license for those patent claims not infringed at the component level. 

IEEE

The IEEE (Institute of Electrical and Electronic Engineers) has a standards-development role in several areas. Most relevant are the family of radio communication standards known as Wi-Fi. 

The IEEE uses the term Letter of Assurance to describe a document stating a Submitter’s commitment to license an Essential Patent Claim. The requirements for a Letter of Assurance are defined in clause 6.1 of the IEEE-SA Standards Board Bylaws. 

These bylaws were revised in 2015. The revised byelaws include a paragraph on the use of the “smallest saleable patent practising unit” (“SSPPU”) as a baseline for determining royalties: they do not mandate this, but “recommend” it. It is often asserted that these revised bylaws require (or recommend) licensing at the component level. 

The new proposed wording of a Letter of Assurance includes a statement that the Submitter will make available a license for Essential Patent Claims to an unrestricted number of Applicants, on a worldwide basis to make, have made, use, sell, offer to sell, or import any Compliant Implementation that practises the Essential Patent Claims for use in conforming with the IEEE Standard. “Compliant Implementation“ means any product (e.g., component, sub-assembly, or end-product) or service that conforms to any mandatory or optional portion of a normative clause of an IEEE Standard.

In a paper in IAM on 12 October 2020 the respected University of Utah law professor Jorge L Contreras notes the reference to components, sub-assemblies. This explicit reference, he argues, clearly establishes a license-to-all obligation for SEP holders.

But, to fall within the wording it is not enough that a device be a component or sub assembly. It also has to “conform” to the standard. As with 3GPP standards, the IEEE’s family of Wi-Fi standards do not specify implementations: they do not specify how a device or a component is to be made. They specify interoperability: for example how one device in a radio network should respond to a signal from another device, or how it should initiate a communication. In order to “conform” to such clauses in a standard, the device must be capable of receiving wireless signals and transmitting wireless signals. So, although the policy expressly mentions components and sub-assemblies as potentially falling within the definition of a “compliant implementation”, there is a limit to the level of those components which could still be compliant. A Wi-Fi module may conceivably “conform” to an mandatory or optional portion of a normative clause of an IEEE Standard. But it is hard to see how any component which is unable to receive or transmit, or provide a specified response to a specified signal, could itself be said to “conform”. 

The license to be offered is also on a claim by claim basis, for those claims that are practised by the product. Although it is conceivable that a component may practise a device claim, it would not practise a method claim. Method claims would need to be licensed by the seller of the end user device. 

Like ETSI, the IEEE policy envisages that licenses would include “have made” rights. This does not seem to be consistent with a model which mandates component level licensing. It is more consistent with a model in which a license is granted to a manufacturer, who purchases components from component makers who do not themselves have a direct license. 

The revised bylaws incluude a paragraph on the use of the “smallest saleable patent practising unit” (“SSPPU”) as a baseline for determining royalties. We do not agree that this wording should be read as a recommendation of component level licensing for two reasons:

• There is a distinction between royalty base, which is a factor in determining the amount of the royalty, and licensing level, which concerns with whom the license agreement is entered. Even if the use of SSPPU were mandated by the policy, that is not the same as a statement that licenses must be granted to component makers.
• If anything, the inclusion of SSPPU language in the policy points to the opposite conclusion: that the authors of the policy had in mind that licensing would take place at the level of an end user product maker. Had the authors been envisaging licensing at the level of the component, it would not have been necessary to recommend the use of the SSPPU as a royalty base, because there would be no distinction between the price of the component and the price of the SSPPU.  

The US Department of Justice^[14] notes that since the IEEE’s 2015 policy (including this SSPPU wording) took effect, negative assurances—those in which a technology contributor declines to give a RAND assurance—increased significantly, with the result that in 2019 the American National Standards Institute declined to approve two proposed IEEE standards amending the 802.11 Wi-Fi standard. The Department suggests that this wording should be reconsidered or amended, and that “parties should be given flexibility to fashion licenses that reward and encourage innovation”. 

We conclude that the IEEE policy, even in its 2015 form, does not mandate licensing at the component level, and does not preclude the IPR holder from adopting an “access for all” model. Further, its wording is in two areas actively inconsistent with component level licensing. 

ITU

The International Telecommunication Union is a specialized agency of the United Nations responsible for all matters related to information and communication technologies. It has three sectors, of which two are relevant to this paper. ITU-R manages radio frequency allocation and communications globally. ITU-T operates as a standards-development organisation for telecommunications. ITU is most relevant as the SDO for video compression standards such as H.264. 

ITU-T and ITU-R share a common IPR policy with the ISO and IEC^[15].  The Policy is supplemented by Guidelines^[16]which are “intended to clarify and facilitate implementation of the Patent Policy.” 

The objective of the Policy is to make “a patent embodied fully or partly in a Recommendation | Deliverable […] accessible to everybody without undue constraints”. 

The objective of the policy is therefore to ensure access to the technology described within a patent. The Policy is neutral as to how that access is to be achieved, whether by licensing or other means, stating:. “[…] The detailed arrangements arising from patents (licensing, royalties, etc.) are left to the parties concerned, as these arrangements might differ from case to case.” 

The patent statement and licensing declaration form for ITU-T or ITU-R recommendations contains the following wording:

“The Patent Holder is prepared to grant a license to an unrestricted number of applicants on a worldwide, non-discriminatory basis and on reasonable terms and conditions to make, use and sell implementations of the above document.

Negotiations are left to the parties concerned and are performed outside the ITU-T, ITU-R, ISO, or IEC.”

The word “Patent” on the ITU form means “those claims contained in and identified by patents, utility models and other similar statutory rights based on inventions (including applications for any of these) solely to the extent that any such claims are essential to the implementation of the same above document. Essential patents are patents that would be required to implement a specific Recommendation or Deliverable”.

As is usual, the ITU’s family of video compression standards do not specify implementations: they do not specify how a device is to be made. The H.264 standard describes the format of the encoded data and how the data is decoded, but it does not specify algorithms for encoding video – that is left open as a matter for encoder designers to select for themselves, and a wide variety of encoding schemes have been developed. Nor does it specify how those algorithms are to be implemented, which may be done in a hardware device, or in software running on a generic processor.

In order to “conform”, a device must be capable of receiving a compressed video signal and decoding it. As with the cellular standards above, there is a limit to the level of those components which could still be compliant. If the implementation of the receiving portion is done in software, the hardware components would clearly not conform. If implementation of the receiving portion is done in hardware, a stand-alone H.264 module might “conform” to the H.264 Standard. But it is hard to see how any individual component which is unable by itself to receive or decode could be said to “conform”. 

The license to be offered is also on a claim by claim basis, for those claims that are practised by the product. Although it is conceivable that a component may practise a device claim, it would not practise a method claim. Method claims would need to be licensed by the seller of the end user device. 

For these reasons, we do not conclude that the ITU licensing declaration mandates licensing to component makers. The objective of the policy is to give access to the technology, which could include by licensing, but also by other means. On the transmission side, there does not appear to be any commitment to license at all. On the receiving side, any licensing commitment appears limited to devices, together with any necessary software, which are capable of receiving and decoding. For components, there is no licensing obligation. 

Broadband forum IPR Policy April 20, 2016

The Broadband Forum is a consortium for developing broadband network specifications. Members include telecommunications networking and service provider companies, and broadband device and equipment vendors.

The IPR policy requests SEP owners to either (a) offer to license Necessary Claim(s) to any Implementer, or (b) agree not to assert patents against implementers. 

The IPR policy also envisages that licenses would include “have made” rights. This would be inconsistent with a model of component level licensing. It is consistent with licensing a manufacturer who purchases components from unlicensed suppliers. Provided that this would not result in discrimination, the policy would appear to allow a combination: a SEP owner would comply with the policy by offering licenses to end user product makers and agreeing not to assert patents against component makers. 

The IPR Policy of the Broadband forum therefore does not require component level licensing: it expressly allows for an “access for all” model, and contains wording inconsistent with mandatory component level licensing.

Internet Engineering Task Force May 2017

The Internet Engineering Task Force develops and promotes voluntary Internet standards, in particular the standards that comprise the Internet protocol suite. It has no formal membership, but for parties who participate in working groups the IETF Secretariat may request written licensing assurances. The assurances allow the patent owner to offer royalty-bearing licenses, royalty free licenses, or agree not to assert its patents. It may choose different options for different persons. 

Whilst the IETF’s standards are less obviously hardware-related than other standards, an SEP owner could comply with the IETF’s IPR policy by offering licenses to end user product makers and refraining from asserting patents against component makers.

OASIS

Oasis is a standards body encompassing cybersecurity, blockchain, privacy, and other technologies. A party to an Oasis Technical Committees may license or provide under non-assertion covenants for any claims under its patents or patent applications that become Essential Claims. 

The license permits an implementer to make, have made, use or sell Licensed Products that implement such OASIS Standards. Licensed Products are portions of a Licensee’s products (hardware, software or combinations thereof) that implement and are compliant with all Normative Portions of an OASIS Standards Final Deliverable that must be implemented to comply with such deliverable.

The licensing obligation would therefore appear to apply only to a product which brings together all elements of the relevant standard. Whether a license should be granted would depend on whether that product can, of itself, implement the standard, or whether it instead forms part of a larger product which does so. 

The reference to “have made” rights again suggests that the policy does not envisage a license being granted to each component maker, but rather allows for a  license to be offered to a mnufacturer, and acquire components from unlicensed component makers. 

The OASIS IPR Policy therefore does not require component level licensing: it expressly allows for an “access for all” model, and contains wording inconsistent with mandatory component level licensing.

W3C

W3C publishes “Recommendations”, which are standards that define web technologies for application development. HTML5 is an example. 

Participants in Working Groups may agree to make available essential patent claims related to the work of that particular Working Group. The terms of the license allow an implementer to make, have made, use, sell, etc. implementations of the Recommendation. The license:

• shall be available to all, worldwide, whether or not they are W3C Members;
• covers essential patent claims; and
• may be limited to implementations of the Recommendation, and to what is required by the Recommendation.

Although the licensing obligation expresses that it is “available to all”, as with other policies it would appear to apply only to a product which is an implementation of the Recommendation. Whether a license should be offered in respect of a product would again depend on whether that product of itself implements the Recommendation, or whether it merely forms part of a larger product which does so. 

The reference to “have made” rights again suggests that the policy does not envisage a license being granted to each component maker, but rather allows for a manufacturer to take the license, and acquire components from unlicensed component makers. 

The W3C IPR Policy therefore does not require component level licensing, and contains wording inconsistent with mandatory component level licensing.

Conclusion

In its 8 June 2020 paper, the Fair Standards Alliance has sets out quotes from a number of SDO IPR policies which are said to support the assertion that a SEP Owner must adopt a policy of offering licenses of their SEP portfolio to each component maker. In these, the Fair Standards Alliance places emphasis on words such as “all applicants” in support of the assertion that the SEP Owner must adopt a policy of offering licenses to component makers. 

The words “all applicants”, however, must be read in context, since each of the policies goes on to set out further restrictions on the licenses to be granted. Where IPR Policies have adopted wording such as “all applicants”, they may be seeking to make clear that licenses must be offered to persons whether or not they are members of the SDO, for example. The requirement to offer licenses to “all” does not override the other provisions of the policy. To use an analogy: a local government requirement that a swimming pool be open to “all” does not permit its use by windsurfers or kayakers. 

We find that the SDO policies looked at do not require a SEP owner to offer licenses to the manufacturer of components, even where they state that licenses must be offered to “all applicants”, or similar. There are, broadly, five reasons for this. 

1. Some policies contain wording that is inconsistent with component level licensing. 
2. Some policies give the express option of either a) licensing, or b) non-assertion^[17]. A SEP owner may choose to license to some classes of implementers, and “not assert” against others^[18].
3. All of the policies restrict the purpose for which the SEPs may be used. This purpose restriction inherently limits to whom licensing may occur: a SEP owner need not offer a license to a component maker who does not operate within the purpose restriction.  
4. Some policies require licensing on a claim by claim (as opposed to a patent by patent) basis. Many SEPs have both equipment and method claims. Even if a component satisfies all integers of an equipment claim, it will typically not of itself fall within a method claim.  There would be no obligation to offer a license to those method claims to such a component maker.
5. None of the policies actually require that licenses be granted. All give the SEP holder the option of refusing to offer licenses. This would include limiting to whom it offers licenses. 

We also note that, at the time of writing, case law appears to support this interpretation. 

• The FTC v Qualcomm^[19] decision, initially finding that the ATIS policy required licensing at the component level, has been overturned. 
• Two decisions of the German courts^[20] have dismissed an argument that the IPR holder is obliged to offer licenses to component makers. 
• The English Court in Unwired Planet^[21], the District Court for the Central District of California in TCL v Ericsson^[22] and the District Court for the Western District of Washington in Microsoft v Motorola^[23] have each imposed SEP licenses at the end-user product manufacturer level^[24].
• In similar circumstances, courts from China, India, Brazil and across Europe have upheld license offers made by, or to, end-user product makers as being FRAND. They have not found that such licensing should be taking place at a component level. 

There would therefore appear to be a general global consensus that SDO IPR policies do not mandate licensing at the component level, and that they permit IPR holders to adopt an “access for all” IPR licensing model.  

---

^[1] For example Nokia v Daimler, Case No. 2 O 34/19 (Mannheim District Court), in which Nokia offered a license covering the vehicle, with “have made” rights covering Daimler’s suppliers.

^[2] See, for example, FTC v Qualcomm 19-16122, U.S. Court of Appeals for the Ninth Circuit (San Francisco), in which the court heard evidence from the major SEP holders in the cellular industry that they do not assert SEPs against component makers.

^[3] Nokia v Daimler Case No. 2 O 34/19 (Mannheim District Court), Sharp v Daimler: judgments not published at the time of writing.

^[4] A “value chain” license covers the vehicle, and, through “have made” rights, the components which go into it. It accords to model b, above. 

^[5] An “exhaustive” patent license is a license to the SEP portfolio which covers both the component maker’s activities and those of his downstream customers. This is not the same as a true “license to all” model, under which the downstream customers would also enter a license. 

^[6] https://fair-standards.org/wp-content/uploads/2020/06/200605-FSA-Position-Paper-Supporting-Legal-and-Policy-References-License-to-All.pdf

^[7] ITU, IEEE, TIA, ATIS, 3GPP and ETSI

^[8] https://portal.etsi.org/new3g/Management/OP.htm

^[9] https://www.atis.org/wp-content/uploads/01_legal/docs/OP.pdf

^[10] art 6 of the 3GPP Working procedures https://www.3gpp.org/ftp/Information/Working_Procedures/3GPP_WP.htm#Article_6

^[11] Decision of the ANSI Executive Standards Council Appeals Panel 23 February 2018

^[12] No. 17-CV-00220-LHK (N.D.Cal. Nov. 06, 2018.   This decision was subsequently overturned by the Ninth circuit [19-16122, U.S. Court of Appeals for the Ninth Circuit (San Francisco)]

^[13] It was amended in 2005 “to make it explicit that a patent that was necessarily infringed by the practise of an ‘optional’ element of a standard was nonetheless still an ‘essential’” (Koninklijke Philips NV v. Asustek Computer Inc., [2016] EWHC 2220 (Pat) 

^[14] Letter dated 10 September 2020 to the IEEE

^[15] The text of the Policy  is at https://www.itu.int/en/ITU-T/ipr/Pages/policy.aspx . 

^[16] Guidelines are at https://www.itu.int/dms_pub/itu-t/oth/04/04/T04040000010005PDFE.pdf  

^[17] In some countries, courts have found that a formal written “non assert” can be equivalent to a license. Where an IPR policy draws a distinction between a license and a non-assertion, it seems likely that the intention of the draftsman was to give effect to the distinction.

^[18] Provided that this does not result in discrimination between direct competitors. 

^[19] 19-16122, U.S. Court of Appeals for the Ninth Circuit (San Francisco),

^[20] Nokia v Daimler and Sharp v Daimler, ibid. 

^[21][2017] EWHC 711 (Pat)

^[22] TCL Communications v Ericsson, Case No. 8:14-cv-00341-JVS-DFM (C.D. Cal. Dec. 21, 2017).

^[23] Case 2:10-cv-01823-JLR

^[24] In each case the manufacturers of the end-user products used Qualcomm or Mediatek chips.