Summary Contents
The license grant is the heart of any intellectual property (IP) license. This chapter explores many of the issues that arise in defining what rights are granted under a license agreement.
6.1 Licensed Rights
One of the most fundamental things that every license agreement must define is the set of rights that are being licensed. This definition must answer two related questions: what type of rights are being licensed (e.g., patents, copyrights, trademarks, etc.), and which of those rights are being licensed (e.g., which of the licensor’s patents, copyrights, trademarks, etc.)? Though this exercise may sound straightforward, there are many ways that licensed rights can be identified, with significant ramifications for both the licensor and licensee. (Note that the definition of licensed rights is often tailored to the type of IP being licensed, so that a patent license agreement might refer to “Licensed Patents” instead of “Licensed Rights” and a trademark license may refer to “Licensed Trademarks,” “Licensed Marks,” “Licensed Brands” or some other variant.)
6.1.1 Enumerated Rights
One way to identify licensed rights is by enumerating those rights specifically and individually. Such an enumeration can refer to the governmental registrations for those rights, such as patent, trademark and copyright registrations. If there are too many rights to list conveniently in the text of a definition, a separate list can be attached as an exhibit to the agreement. Here is a simple example involving registered trademarks.
Single Registered Mark
“Licensed Mark” means U.S. Trademark Reg. No. 999,999 “SUPER-BEV”.
Multiple Registered Marks
“Licensed Marks” means those U.S. and foreign trademark registrations listed in Exhibit A to this Agreement.
Unregistered IP can also be enumerated, so long as it can be described in a manner that clearly identifies and distinguishes it. Thus, unregistered (common law) trademarks can be included as part of a license grant, as can unregistered copyrights and even inventions and trade secrets that are not (yet) subject to any patent application. Some examples include the following.
Enumerated Rights Including Unregistered IP
“Licensed Marks” means those Marks that are listed in Exhibit A to this Agreement.
“Marks” means trademarks, service marks and designs, whether or not registered.
“Licensed Rights” means all Authorship Rights throughout the world subsisting in the work THE GREAT AMERICAN NOVEL by Author.
“Authorship Rights” means copyrights and related rights of authors, including moral rights.
“Licensed Rights” means all Know-How in Licensor’s proprietary method for curing rubber utilizing heat modulation calibrated using the Arrhenius equation, as described in the confidential specification delivered by Licensor to Licensee on October 31, 2020.
“Know-How” means all know-how, trade secrets, discoveries, inventions, data, specifications and other information [, including biological, chemical, pharmacological, toxicological, pharmaceutical, analytical, safety, manufacturing and quality control data and information, study designs, protocols, assays and clinical data], whether or not confidential, proprietary or patentable and whether in written, electronic or any other form.Footnote 5
The above definitions include both a generic definition of the category of IP, as well as a definition of the licensed IP that incorporates the generic category. Using this modular structure in all but the simplest licensing agreements is advisable, as the generic IP category may be referred to elsewhere in the agreement (e.g., in the indemnification section) and it is best to use consistent terminology throughout.
Patents pose some additional issues. Like trademarks, patents are registered (and there are no common law patent rights analogous to common law trademarks). Yet many different patents may relate to the same basic invention. That is, during the patent prosecution process, patent applications may be subdivided, amended, continued and extended through a variety of different procedural mechanisms. Thus, one invention can end up being claimed by a dozen different patents that are issued for years following the issuance of the initial patent. Foreign patent applications can also be filed in multiple countries under the Patent Cooperation Treaty (PCT) based on an original application in one country. These groups of related patents are often called a “patent family,” and licensing is often conducted at the family level rather than the level of individual patents. The unifying trait of a patent family is often the ability to trace the origin of a patent to a single “ancestor” application filed on a date that is known as the “priority date” for the family and its other members. Figure 6.1 illustrates the different members of such a patent family. An example of a patent family definition follows.
Graphical representation of a patent “family.”
“Licensed Patents” means U.S. Patent No. x,xxx,xxx entitled “Improved Method for Slicing Bread and Apparatus Therefor” (September 9, 1999), together with all Patents claiming the same priority as such patent. [1]
“Patents” means (a) patents and patent applications [4], and all divisional, continuation, and continuation-in-part applications of any such patent applications; (b) all patents issuing from any of the foregoing applications; and (c) all reissues, reexaminations, extensions, foreign counterparts [2] and supplementary protection certificates [3] of any of the patents described in clauses (a) or (b).
[1] Priority – the “priority date” of a patent is the date on which the earliest utility patent application in the “family” of related applications was filed.
[2] Foreign counterparts – this term refers to foreign patents and patent applications, often filed under the Patent Cooperation Treaty, that derive from the same parent application.
[3] Supplementary protection certificates – these are European rights that protect certain pharmaceutical and other regulated compositions after their patent protection has expired.
[4] Applications – even though patent applications convey no enforceable rights, they can, and often are, licensed. Doing so is a convenient way to ensure that any patent rights that eventually emerge from such applications are licensed. The alternative would be to require the licensor to be extremely diligent in adding patents to the license grant as they are issued, a responsibility that benefits neither party.
The following case illustrates the importance of including the “right” rights in a license agreement.
829 F.2d 1075 (Fed. Cir. 1987)
BALDWIN, SENIOR CIRCUIT JUDGE
In 1983, Suessen, brought an action in the district court for infringement of two patents relating to improvements in the technology of open-end spinning devices, U.S. Patent No. 4,059,946 (the ’946 patent) and U.S. Patent No. 4,175,370 (the ’370 patent).
Schubert argues that it has an implied license under the ’946 patent. Its argument involves two agreements.
The first was a license agreement entered in 1982 between Schubert and Murata Machinery, Ltd. (Murata). That agreement, entered into before the filing of this suit in 1983, in pertinent part reads:
Murata hereby grants to Licensee [Schubert] a non-exclusive worldwide license under the Patents to make, use and sell the patented device only as part of the open end spinning machines of the License. The License hereby granted is a limited license, and Murata reserves all rights not expressly granted.
The “Patents” were defined [to] include U.S. Patent No. 4,022,011 (’011 patent) and other patents belonging to Murata in the name of Hironorai Hirai. Schubert asserts that, notwithstanding any infringement of ’946, its accused infringement is merely a practicing of the ’011 invention, which it is licensed to do under the 1982 agreement.
The second agreement, entered in 1984 after this lawsuit began, involved Suessen’s purchase of the ’011 and [other] patents from Murata. The agreement reads, in pertinent part:
Suessen has been advised by Murata that a non-exclusive license of the patents and patent applications mentioned under 1. above had been granted by Murata to [Schubert] (hereinafter called the Licensee). Suessen hereby agrees to purchase the patents and patent applications mentioned under 1. above together with the License Agreement as of 23rd/28th July, 1982, with the said Licensee and agrees that you and your business/license concerns will maintain the licensed rights of the Licensee under the License Agreement as stipulated during the life of the patents and patent applications mentioned under 1. above.
Schubert asserts that, per the 1984 agreement, Suessen “stepped in the shoes of Murata” [and] cannot – just as Murata cannot – sue under the ’946 or any other patent for infringement based on practicing the ’011 invention. To allow such a suit, Schubert argues, would unfairly take away what it paid for in 1982. Schubert labels its argument one of “legal estoppel.”
[Schubert] asserts an implied license based on its theory of legal estoppel. Though we recognize that theory in appropriate circumstances, it does not work for Schubert here. Legal estoppel is merely shorthand for saying that a grantor of a property right or interest cannot derogate from the right granted by his own subsequent acts. The rationale for that is to estop the grantor from taking back that for which he received consideration. Here, however, we have a suit by a third party, Suessen, under a patent owned by Suessen. The license by the grantor, Murata, did not purport to, and indeed could not, protect Schubert from a suit by Suessen under ’946. Hence, Suessen, by filing in 1983 and now maintaining its suit under ’946, does not derogate from the right given by Murata in the 1982 license agreement.
Ownership and license of ’011 and ’946 patents in Spindelfabrik.
Schubert nevertheless urges this three prong argument: (1) “legal estoppel” would prevent Murata from suing under the ’946 patent if it were to acquire it; (2) Suessen “stepped into” Murata’s shoes in 1984 when Suessen acquired the Hirai patents and committed to maintain Schubert’s licensed rights; and hence, (3) just as Murata could not, Suessen cannot sue under the ’946 patent. We reject that argument.
As a threshold matter, a patent license agreement is in essence nothing more than a promise by the licensor not to sue the licensee. Even if couched in terms of “[l]icensee is given the right to make, use, or sell X,” the agreement cannot convey that absolute right because not even the patentee of X is given that right. His right is merely one to exclude others from making, using or selling X. Indeed, the patentee of X and his licensee, when making, using, or selling X, can be subject to suit under other patents. In any event, patent license agreements can be written to convey different scopes of promises not to sue, e.g., a promise not to sue under a specific patent or, more broadly, a promise not to sue under any patent the licensor now has or may acquire in the future.
As stated previously, the first prong of Schubert’s three part “stepping in the shoes” argument is that legal estoppel would prevent Murata from suing Schubert under the ’946 patent if Murata were to acquire that patent. However, even assuming, arguendo, that such estoppel against Murata exists, the final two prongs of Schubert’s “stepping in the shoes” argument would fail. Given the assumption of estoppel against Murata, the 1982 license agreement would necessarily be a promise by Murata not to sue under any patent, including those acquired by Murata in the future. In the 1984 agreement, Suessen incurred what Murata promised in 1982. Thus, Suessen would be committed to forebear from suit under (1) the transferred patents and (2) any of Murata’s nontransferred patents (future and present). That commitment does not include a promise not to sue under Suessen’s own ’946 patent.
Schubert’s “standing in the shoes” argument, however, would add to Suessen’s commitment a promise not to sue under Suessen’s separate patents that Murata never owned. On the facts of this case, we cannot interpret the 1984 agreement so broadly, at least not with respect to the ’946 patent.
The district court correctly determined that there is nothing in the 1984 agreement about the ’946 or other Suessen patent rights. Schubert points to no extraneous evidence tending to show any understanding on the part of either contracting party that Suessen was to forego rights under the ’946 or any other patent then owned by Suessen. To the contrary, that a lawsuit under ’946 was ongoing but not mentioned in the 1984 agreement indicates strongly that there was no intent by the parties to have Suessen forfeit its rights under ’946. Furthermore, an implied promise by Suessen to forego its ’946 suit is inconsistent not only with Suessen maintaining its lawsuit after the 1984 agreement but, also, with the course of events leading up to the 1984 contract. In sum, we agree with the district court’s conclusion that the 1984 agreement did not impose on Suessen any obligation to stop its ongoing suit under the ’946 patent.
Schubert argues that not implying a license in this case is unfair because Schubert paid valuable consideration for the right to practice the ’011 invention but is in danger of losing that right as a result of doing no more than that for which it paid. We disagree. The right Schubert paid for in the 1982 agreement was freedom from suit by Murata, not Suessen. Indeed, when Schubert signed the 1982 agreement, it was aware of possible suit by Suessen, who had previously denied Schubert a license under the ’946 patent. Moreover, Schubert has not shown us that it has lost any obligation Murata may still owe it under the 1982 license agreement, e.g., not to sue under any patents Murata still has or may acquire. To rule that the Suessen acquisition of the ’011 patent somehow bestows on Schubert an absolute defense to a suit already filed by Suessen under ’946, would result in an unintended windfall to Schubert that makes no sense under the facts of this case.
AFFIRMED
Notes and Questions
1. Implied licenses. In Chapter 4 we saw several examples in which courts implied licenses based on the conduct of the parties. How is Spindelfabrik different than these cases? If you were the judge, would you have recognized an implied license from Suessen to Schubert?
2. Patent families. How might the patent family definition suggested above have helped the parties in the TransCore and Endo cases discussed in Section 4.3, Notes 3–4?
3. The importance of timing? In Spindelfabrik, under the 1982 agreement, Murata licensed the ’011 patent to Schubert. In 1984, Murata assigned the ’011 patent and Schubert’s license to Suessen. Prior to that, Suessen asserted the ’946 patent against Schubert. The court held that nothing about Suessen’s purchase of the ’011 patent and license committed it to license the ’946 patent to Schubert. But what if Suessen had purchased the ’011 patent and license before it asserted the ’946 patent against Schubert? Would this have changed the outcome? What if the ’946 patent had originally been owned by Murata, but not included in the 1982 agreement, and then assigned to Suessen at the same time as the ’011 patent? Would Schubert’s estoppel argument be stronger?
4. Products versus patents. The Spindelfabrik case is really about product versus rights licenses (see the box “Rights Licenses versus Product Licenses”). Schubert argues that because it licensed the ’011 patent, it had an absolute right to manufacture the product covered by the ’011 patent. But it did not. The licensee of a patent only has the right to operate under the licensed patent and no more. How might the license agreement have been written to achieve what Schubert hoped, or assumed, it had achieved?
6.1.2 Portfolio Rights
Defining licensed rights by reference to a specific registered (or unregistered) IP right and its associated family members is relatively precise and avoids ambiguities regarding what is and is not licensed. Yet enumerating individual licensed rights can be both an administrative burden and a trap for the unwary. Suppose that a licensee wishes to obtain a license not to one, but a thousand different patents covering a complex product such as a smartphone or a computer. If the licensor were required to list every one of the licensed patents, it is possible that one or more patents might be overlooked. And, given cases like Spindelfabrik, it is difficult to argue that a right that is not enumerated in a list of licensed IP should be included in a license.
To get around this problem, parties have developed language under which groups of IP rights can be licensed without enumerating every one of them. Below is an example of such a “portfolio.”
“Licensed Patents” means all Patents throughout the Territory that are Controlled by Licensor or any of its Affiliates at any time during the Term [1] and that (a) claim all or any part of Licensor’s Super-Slicer bread slicing device, or the use thereof [2], and (b) have a priority date earlier than January 1, 2021 [3].
“Control” means with respect to any intellectual property right, possession of the power and right to grant a license, sublicense, or other right to or under such intellectual property right as provided for in this Agreement without violating the terms of any agreement or other arrangement with any third party [4].
[1] Temporal portfolio constraint – this clause applies to every patent that is in the licensor’s portfolio during the term, including patents that the licensor acquires after the effective date of the agreement. If the parties wish to limit the portfolio to patents held as of the effective date, “during the Term” can be changed to “prior to the Effective Date.”
[2] Portfolio scope – the above definition is said to cover the licensor’s portfolio of patents pertaining to a particular device. If the licensor wishes instead to grant a license of its entire patent portfolio, then clause (a) would be eliminated.
[3] Cutoff date – clause (b) serves to exclude new inventions from a portfolio license. This approach can be useful if, for example, the license fee is paid in a lump sum (as it may be in a settlement agreement – see Section 11.6) based on the value of the licensor’s existing patent portfolio. Note that the cutoff date in clause (b) may be prior to or after the effective date of the agreement itself and would not exclude newly acquired patents so long as they meet the cutoff date.
[4] Third-party licenses – the definition of control is intended to encompass rights that the licensor owns or otherwise has the power to license. If it has already granted an exclusive license with respect to a right, then it cannot license it again (see Section 7.2.1), so such rights are not included in the license. Of course, a licensee that is concerned about such exclusions (e.g., the Swiss cheese effect) should insist that the licensor make representations and warranties (see Chapter 10.1) regarding the scope of the portfolio that is licensed and any exclusive licenses that could potentially remove necessary IP from the rights granted.
Notes and Questions
1. Which portfolio? In the patent portfolio definition set forth above, the licensed portfolio is defined by reference to a specific product sold by the licensor. Are there other ways that you could define a licensed portfolio? When might a licensor wish to grant a licensee a license with respect to its entire portfolio of patents?
2. Cutoff date. In the patent portfolio definition, there is a cutoff date beyond which patents controlled by the licensor are not included in the license. Would such a cutoff date ever be useful in a license in which the licensed rights are specifically enumerated? The cutoff date in clause (b) may be prior to or after the date of the agreement itself – when would it be useful to have a cutoff date that is after the date of the license agreement?
3. Control. In order to be licensed, patents (and other IP rights) must be owned or controlled by the licensor. This is the principal reason that Schubert’s claim failed in Spindelfabrik – Murata could not license the ’946 patent to Schubert because Murata did not own that patent. Accordingly, the patent portfolio definition set forth above defines licensed patents as those that are owned or controlled by the licensor or its corporate affiliates (see Note 4). Why is this language not needed when the licensed rights are enumerated specifically?
4. Affiliates. The term “Affiliates” is often used in licensing agreements to signify the other members of a party’s corporate “family” – parent, subsidiary and sibling entities. Including IP held by affiliates in definitions such as the licensed rights is important, as large multinational organizations often hold or exploit IP rights in various entities for tax and accounting purposes. It is common to define “Affiliates” using the definition provided under the Securities Exchange Act of 1934:
An “affiliate” of, or a person “affiliated” with, a specified person, is a person that directly, or indirectly through one or more intermediaries, controls, or is controlled by, or is under common control with, the person specified.
The term “control” (including the terms “controlling,” “controlled by” and “under common control with”) means the possession, direct or indirect, of the power to direct or cause the direction of the management and policies of a person, whether through the ownership of voting securities, by contract, or otherwise.
Under this definition, “control” typically means ownership of at least 50 percent of the voting securities or interests of an entity.Footnote 6 As such, majority-owned subsidiaries of an entity are included within the definition of “Affiliates.” Can you think of any reasons that the parties might prefer to define “control” as the ownership of 100 percent of the voting securities or interests of an entity (e.g., limited to wholly owned subsidiaries only)?
Affiliate relationships in a corporate “family.”
There are several contexts in which it is particularly important to pay careful attention to rights and IP held by affiliates:
Alpha grants Beta an exclusive license under “all of Alpha’s and its Affiliates’ IP covering technology x”; Alpha is then acquired by Gamma, a larger company that also works in technology x. Does the license now cover Gamma’s IP as well? What if the license was paid-up at the time of grant?
Instead, assume that Alpha’s subsidiary Delta also holds IP relating to technology x, and Alpha then sells Delta to Epsilon. Is Delta’s IP still licensed to Beta? Is Epsilon’s?
Now suppose that Beta has granted a license under its own IP to Alpha and its affiliates. Alpha sells Delta to Epsilon. Does Beta’s license to Delta continue once it is owned by Epsilon? Does Epsilon now get a license under Beta’s IP? Again, what if all of these licenses were paid-up at the time of grant?
Can you think of more scenarios in which the extent of corporate families can play an important role in IP transactions?
5. IP divestitures. As discussed in Chapter 3.5, licenses of IP rights generally continue even if the underlying IP right is sold by the licensor. Thus, when a licensor grants a portfolio license and then sells some of the IP rights in the portfolio, the licensee can generally rely on the continuation of that license, and the buyer (or even a new exclusive licensee) takes subject to the earlier-granted license. If this is the case, then in Spindelfabrik why did Murata assign the 1982 license agreement with Schubert to Suessen when Murata sold the ’011 patent to Suessen?
6. Copyright (and trademark) portfolios. Portfolio licenses are not limited to patents. In many cases, copyrights are licensed on a portfolio basis as well. For example, a television network may license all of its programming to a cable provider or online streaming service, and a performing rights organization such as ASCAP or BMI routinely licenses thousands of songs to its licensees for particular uses (see Chapter 16). Trademarks, however, are not typically licensed on a portfolio basis, but are usually enumerated, even if a large number of marks are being licensed. Why?
6.1.3 The Puzzle of “Know-How” Licensing
License agreements involving technology often include a grant of rights with respect to “know-how.” What is “know-how”? It is not a recognized form of IP. Though it may encompass trade secrets, know-how is generally understood to be broader than trade secrets alone. Noted organizational theorist Eric von Hippel defines know-how as “the accumulated practical skill or expertise which allows one to do something smoothly and efficiently.”Footnote 7 J. N. Behrman, who conducted some of the first empirical studies of IP licensing in the United States, defines it as:
whatever unpatented or unpatentable information the licensor has developed and which the licensee cannot readily obtain on his own and is willing to pay for under the agreement; such as techniques and processes, the trade secrets necessary to make and sell a patented (or other) item in the most efficient manner, designs, blueprints, plant layouts, engineering specifications, product mixes, secret formulas, etc.Footnote 8
Unlike know-how, trade secrets are recognized IP rights protected in the United States by federal and state statutes as well as common law. As discussed in Chapter 5.2, in order to be considered a trade secret, information must derive independent value from being kept secret and must be the subject of reasonable efforts to maintain confidentiality. These requirements are not always easy to establish, and information that is conveyed during the course of technical training, product demonstrations and support calls may not always qualify as trade secrets, despite its value to the recipient.
The concept of “know-how” has thus evolved to include both trade secrets as well as other information that is conveyed by the licensor to the licensee. So long as information is conveyed – whether orally, visually or in writing – it can be “know-how.” This lower bar is useful primarily to establish a basis for the payment of ongoing royalties with respect to the information conveyed. That is, if a license covered only trade secrets, and the information in question lost trade secret status for some reason, then it is unclear that the license would remain in effect or that the licensee would have a continuing obligation to pay royalties. But if the royalty were payable instead on know-how, then the loss of trade secret status with respect to some or all of that information would not affect the license or the obligation to pay royalties. Thus, know-how is a more flexible concept than trade secrets, supporting a stronger basis for the payment of royalties.
However it is defined, know-how is frequently featured in license agreements. As early as 1959, a study of more than 1,000 IP license agreements found that approximately 39 percent included grants of know-how rights.Footnote 9 More recently, Thomas Varner, in a study of over 1,400 publicly filed patent licenses and assignments, found that 56 percent included know-how.Footnote 10 So we know that know-how is being licensed, but what does this mean in practice?
There are two principal functions that know-how licensing plays in IP transactions. The first is straightforward. If the licensor provides training, support, consultation, expertise or some other technical services to the licensee, then the information and skills conveyed through those services are “licensed” as know-how. Even though most of this intangible knowledge is not protected (or even protectable) by formal IP rights, courts have long recognized that such information can be valuable and thus the subject of compensation.
In many cases, however, no such knowledge transfer is contemplated, yet the license agreement (usually a patent license agreement) still contains a license of know-how. In these cases, the know-how license is included simply as a clever way for the licensor to continue to collect royalties after the relevant patents expire. As we will see in Chapter 24, it is illegal under US law for a patent holder to continue to collect royalties for the use of a patented technology after the patent has expired. To get around this limitation, patent holders can license patents and know-how together, so that even after the patents expire, there is still a valuable asset to support the payment of royalties (albeit at a reduced level). The same logic applies to patents that are invalidated after being licensed, and to sales of products in countries where the patent holder does not seek patent protection. In all of these cases, royalties can be collected with respect to know-how, even though no enforceable patents are licensed.
“Licensed Rights” means the Licensed Patents and all associated Know-How conveyed by the Licensor to the Licensee hereunder.
“Know-How” means trade secrets, knowledge, techniques, methods and other information, whether or not patentable.
Notes and Questions
1. The risks of know-how licensing? In the 1950s and 1960s there was significant concern among scholars and policy makers that the licensing of vaguely defined know-how might run afoul of antitrust laws.Footnote 11 We will study antitrust issues arising in connection with licensing transactions in Chapter 25, but for now it is sufficient to understand that “licenses” of this amorphous set of rights were viewed as a potential cover-up for otherwise anticompetitive arrangements. What kind of anticompetitive behavior do you think a know-how license might conceal?
2. Know-how licensing and patent trolling. Professors Robin Feldman and Mark Lemley have observed that when a patent holder makes an unsolicited licensing proposal to a potential licensee (e.g., as a prelude to an express or tacit threat of litigation), the resulting licenses seldom include a transfer of know-how.Footnote 12 This result held whether the patent holder was a non-practicing entity, a university or a company. Professor Colleen Chien, in contrast, found in a study of publicly filed licenses of software patents that most patent licenses also included a license of know-how or software code.Footnote 13 She explains the difference between her results and those of Feldman and Lemley as follows:
Patent licenses that include knowledge, know-how, personnel, or joint venture relationships are more likely to represent direct transfers of technology, whereas the transfer of “naked” patent rights is more likely to primarily represent a transfer of liability between the parties.Footnote 14
What does Chien mean by a “transfer of liability between the parties”? Why would know-how transfers be more frequent in a broad sampling of licensing transactions than patent-owner-initiated demands for a license?
3. Taxing know-how. In addition to antitrust, early concerns over know-how licensing arose from tax law. Was know-how a taxable asset, or was the transfer of know-how a service? In each case, how was it valued? Together with “goodwill” (see Chapter 2.4), know-how presents one of the more interesting tax issues in the field of IP licensing.Footnote 15
6.1.4 Product Rights
So far, our consideration of licensed rights has focused on specific IP rights or groups of rights that the parties desire to license. This approach is natural when particular patents, copyrights, trademarks or trade secrets are known, or expected, to have value in themselves. However, it is often the case that a licensee is interested in exploiting a product that may be covered by a variety of IP rights held by the licensor, and neither the licensor nor the licensee knows, or particularly cares, which rights those may be.
Software programs often fall into this category. Software can be protected by large numbers of copyrights, patents, trade secrets, trade dress, trademarks and other forms of IP. But if a distributor wishes to resell a software program via an online store, or a consumer wishes to use that software on her laptop computer, it is unlikely that they are aware of, or have any desire to know about, the specific IP rights covering that software. In fact, in many cases, even the owner of the software, particularly if it is a large company, may not be aware of the many different IP rights that protect it. Thus, in software and other industries, the common practice is to license all rights pertaining to a particular product without any attempt to list or even categorize them.
The term “Licensed Software” means the executable object code of the SOFTMICRO application (version 1.0) and all of Licensor’s patent, copyright, trade secret, trade dress and other rights in and to such software application and its operation, but excluding trademarks.
In some cases, a licensor granting a license with respect to a full product, especially a software product, will not even recite the IP rights that are licensed at all, and will simply grant the license in the Grant clause of the agreement (see Section 6.3). Or, if it separately defines the licensed software, it will omit to mention any IP rights.
The term “Licensed Software” means the executable object code of the SOFTMICRO application (version 1.0).
There is a critical difference between licenses of rights and licenses of products. In a license of rights, for example a patent license, the licensee is permitted to create and exploit any product that it wishes within the bounds of the license grant (e.g., within the field of use and scope of license discussed in Section 6.2). Thus, if the licensed patent covers an amplifier, the licensee may make any amplifier that it wishes – large, small, low-power, portable, transistorized, heat-resistant, etc. In short, it may use the patented technology to create a product of its own. In contrast, a product license allows the licensee to make only the exact product that is licensed. Thus, if Microsoft licenses its Windows operating system to a PC manufacturer, the licensee is likely permitted to install Windows on its PCs, but not to create a new, improved version of Windows or any other operating system. This key difference is important to keep in mind when reviewing the many variants of license agreements that will be discussed in this book.
Notes and Questions
1. Code. The sample definition of “Licensed Software” relates to the “object code” version of the software. We will discuss the distinction between object code and source code in more detail in Section 18.2. For now, it is sufficient to understand that the object code version of software is the version that runs on a user’s computer or device, but does not allow the user to understand the internal functions of the software or how it is “written.” Why do you think most software distribution and use licenses are limited to object code?
2. No trademarks. Trademark rights are typically excluded from a product-based license or, if granted, are licensed separately. There are several reasons for this convention. First, a trademark license is not required to use a software program, even if the program displays the vendor’s trademarks (we will discuss trademark licensing in greater detail in Chapter 15). A distributor or reseller may require a license to advertise a software program, but that license will contain numerous qualifications and requirements and is thus best granted separately from the right to distribute the program. Finally, doctrines in trademark law such as “nominative fair use” permit parties to refer to a trademarked term in a factual manner (e.g., “We service BMW vehicles”), without the need for a formal license. As a result, a well-drafted definition of product rights should generally exclude trademarks.
6.1.5 Future Rights
It is a somewhat metaphysical question whether an IP right can actually be “licensed” before it is created. Is the license of a future IP right – a patent claiming an invention not yet made, the copyright in a book not yet written – a property interest that exists independent of the right itself, like a contingent remainder or other future interest in real property, or is it merely a promise to license the IP right once it exists? This is a question that deserves to be debated in the law reviews, but is not one that we will answer here. For all practical purposes, as we saw in Stanford v. Roche (Chapter 2.3), interests in IP that is not yet created can clearly be bought, sold and licensed. Yet, as that case also suggests, there is an important difference between a present license of future inventions and a promise to grant a license in the future (with the former clearly preferable to the latter).
In fact, we have already seen licenses of future rights above, in our example of a patent portfolio license. If, during the term of the license, the licensor comes into possession of a new patent that meets the other criteria for a licensed right, then that new patent is licensed along with the rest. But future rights may be licensed more explicitly, and they often are.
“Licensed Work” means the book that is written and delivered by Author hereunder, currently known under the working title THE GREAT AMERICAN NOVEL.
“Licensed Rights” means all patent, copyright, know-how, trade secret and other rights in all developments, inventions and discoveries in the Field made by Dr. Jekyll and the other members of the Jekyll Lab at Stevenson University during the Term.
Problem 6.1
For each of the following deals, draft a suitable definition of the “Licensed Rights”:
a. Transatlantic Corp. has agreed to sell its fleet of Atlantic fishing vessels to United Fishfry. After the sale, Transatlantic will continue to operate its remaining fleet of seven passenger cruise ships. Several years ago, Transatlantic developed a patented method of radar enhancement that greatly improves navigation at sea. The enhancement is now used on all of Transatlantic’s ships. The parties have agreed that, as part of the fleet sale, Transatlantic will grant an appropriate license to United.
b. Lobrow Corp. sells a popular line of children’s toys in the United States based on the popular YouTube character “Bo Weevil.” Assume that Lobrow owns all rights in and to this character and has protected it around the world. In an effort to go international, Lobrow has agreed to grant Downunder, Inc. the right to distribute Bo Weevil toys in Australia and New Zealand.
c. Don Juan has just published a bestselling memoir of his scandalous career in Hollywood. He was recently approached by RealTV, a producer, to develop the memoir into a Netflix television series.
d. Choco Corp. and PeaNot, Inc. are large snack food manufacturers. They have formed a joint venture (JV) to create and market a candy bar that combines the best features of each of their existing product lines (chocolate bars and synthetic peanuts). Each of them will receive 50 percent of the profits of the JV during its existence and has agreed to grant a license to the JV.
6.2 Scope of the License: Field of Use, Licensed Products, and Territory
Once the licensed IP rights are defined, we must define the markets and applications in which the licensee will be permitted to exploit those rights. In some rare cases, a licensor may wish to cede all potential markets and applications of its IP to the licensee throughout the world. If this is the case, then these concepts can simply be incorporated into the grant clause, discussed in Section 6.3. However, if the licensor wishes to grant the licensee only a subset of the total rights available, then careful attention must be paid to defining the scope of the licensee’s use. Three related definitions are often employed for this purpose: Field of Use, Licensed Products and Territory. While different agreements may combine some or all of these definitions, we will discuss each individually before considering how they can be combined.
6.2.1 Field of Use
The field of use (FOU) is the market segment or product category in which the licensee is authorized to exercise the licensed rights. There is a virtually unlimited range of fields that can be specified in an agreement, from extremely narrow to extremely broad. Following are examples of FOU for three different types of IP.
The limitation of a patent licensee’s FOU was validated by the Supreme Court in General Talking Pictures Corp. v. Western Electric, 304 U.S. 175 (1938). In that case, Western Electric, the holder of a patent on electronic amplifiers, licensed the patent to two different licensees: Transformer Co., in the field of amateur radio, and General Talking Pictures, in the field of movie projectors. When Transformer Co. began to sell amplifiers to General Talking Pictures for use in its projectors, Western Electric sued, alleging that Transformer Co. was not licensed to sell amplifiers for use in the theatrical projection market, and was thus infringing Western Electric’s patent. The Supreme Court agreed, holding that “patent owners may grant licenses extending to all uses or limited use in a defined field.”
Fields of use come in two flavors: those that limit the technical application of a licensed right (e.g., “treating emphysema”) and those that limit the customers to which products may be sold (e.g., manufacturers of amateur radio receivers versus movie projectors). In some respects, these two categories can appear to merge, as types of customers are easily defined by different technical applications (and the explicit allocation of customers is a violation of the antitrust laws – see Section 25.3). Nevertheless, analytically it is sometimes convenient to think of FOU as limiting either technical applications or customers.
Some agreements may define multiple fields of use: a licensee may have exclusive rights in some fields and nonexclusive rights in other fields; some fields may be prohibited to it; and it may have the option to acquire rights in still other fields, often upon the payment of a fee.
Biotech (e.g., a new molecule)
Treatment of hereditary breast cancer using a therapeutic agent targeted to variants in the BRCA1 or BRCA2 genes;
treatment of hereditary breast cancer using a therapeutic agent targeted to one or more genetic variants;
treatment of hereditary breast cancer;
treatment of breast cancer;
treatment of cancer;
human therapeutics;
all therapeutic applications, human and veterinary;
all applications, whether therapeutic, diagnostic, agricultural, industrial or military.
Electronics (e.g., part of a 5G telecommunications standard)
Implementation of wideband wireless communication functionality conforming to the 5G standard in a consumer handheld smartphone device;
implementation of wideband wireless communication functionality in a consumer handheld smartphone device;
implementation of wideband wireless communication functionality in a consumer device;
implementation of wideband wireless communication functionality in a communications device;
implementation of wireless communication functionality;
communications applications;
all applications.
Literary (e.g., a popular novel)
English-language print books for the US and Canadian market;
Spanish-language editions;
paperback editions;
ebooks;
magazine serializations;
audiobooks;
stage plays;
television and film adaptations;
action figures and other memorabilia;
T-shirts and other apparel;
theme park attractions.
Notes and Questions
1. Going broad. Generally, a licensee will desire an FOU that is as broad as possible, while the licensor will seek to limit the FOU so that it retains as many rights as possible to grant to others or exploit itself. Under what circumstances might a licensee be concerned about an FOU that is too broad?
2. Biotech FOU. In some industries, particularly biotechnology, there may be multiple potential uses for a licensed compound, such as a molecule, protein or gene. It is thus not uncommon in biotech licenses to see FOU that are limited to specific disease targets (e.g., cancer, cystic fibrosis, diabetes) or delivery mechanisms (e.g., intravenous, oral, topical, gene therapy). In many cases, license grants are exclusive with respect to these narrowly specified FOU. These licenses are typically granted at early stages of product research and development.
However, once a relatively complete drug or therapy is licensed (e.g., from a biotech company to a pharmaceutical manufacturer that will seek regulatory approval and then manufacture and market the drug), it is not typical to limit use by disease indication. The reason is that physicians are generally free to prescribe a medication for any use (i.e., the indicated use as well as “off label” uses), and the distributing company has little means of policing whether those uses fall within the scope of its license.
3. Anticompetitive fields? In General Talking Pictures, discussed above, Justice Black dissented, expressing concern that the allocation of different “fields” to different patent licensees, especially if numerous patents held by different owners were pooled together, could have the effect of creating a series of submonopolies that limited competition. We will discuss antitrust issues in greater detail in Chapter 25, but based on what you now know about FOU, do you agree with Justice Black’s concern?
4. FOU and the lawyer’s role. The FOU definition is one of the few parts of a license agreement that does not depend on legal terminology so much as a deep and accurate understanding of the licensed rights, the market and the potential business relationship between the parties. Clients will often provide their attorneys with a definition of the FOU that they feel is adequate, and that definition may even be embedded in a term sheet or letter of intent before the license drafting begins (see Section 5.3). But the diligent attorney should consider whether there are unanticipated pitfalls in the client’s FOU definition: Is it too broad or too narrow? Will it enable the licensee to carry out the business arrangement that is anticipated? How will it fare in the face of competition from others? Will the licensor have sufficient flexibility to license others in adjacent fields? Will the definition quickly become obsolete as technology advances? Asking questions like these, rather than cutting and pasting an FOU definition from a client’s email or term sheet, will serve the interests of both parties to the transaction.
Problem 6.2
For each of the following IP rights, describe the broadest FOU that you would realistically wish to obtain as the licensee, and the narrowest FOU that you would realistically wish to grant as the licensor:
a. a patented synthetic molecule that converts petroleum products into refined sugar;
b. the #1 R&B hit song “Bag of Fleas” by the megagroup Shag Shaggy Dog;
c. a little-known Bulgarian superhero comic character known as “Tarantula Man”;
d. a patented software encryption methodology that would reduce the effectiveness of cyberattacks by 90 percent;
e. the world-famous “squish” brand/logo that Squish Corp. has popularized through a line of high-end sports footwear;
f. The persona of the recently deceased pop superstar formerly known as Princess.
6.2.2 Licensed Product
The term “Licensed Product” means, essentially, a product made or sold by the licensee that uses or is covered by some or all of the licensed IP rights. The term Licensed Product is important because it often (but not always) defines the licensee’s payment obligation. That is, the licensee often must pay the licensor a royalty based on the licensee’s revenue earned from sales of Licensed Products. So, every Licensed Product triggers a payment. For this reason, the definition of Licensed Product must specify that the product in question is covered by the licensed IP. The licensor is typically not legally entitled to collect royalties on a licensee’s sale of products that are not covered by the licensor’s IP, a practice that is referred to as “misuse” (see Chapter 24).
A basic example of a Licensed Product definition is set forth below. The Cyrix case discussed in Section 6.3 introduces additional complexities to this definition, particularly clause (a).
“Licensed Product” means a product that is (a) manufactured or sold by or for the Licensee or its Affiliates and (b) which is covered by any claim of the Licensed Patents.
Licensed Product (Patent + Know-How)
“Licensed Product” means a product that is (a) manufactured or sold by or for the Licensee or its Affiliates and (b) which is covered by any claim of the Licensed Patents or which embodies, or is manufactured using, any of the Licensed Know-How.
6.2.3 Territory
Every IP license has a territorial scope, whether implicitly through the inherent national character of intellectual property rights or, more typically, as defined in the agreement.
Some licenses are worldwide. That is, they allow the licensee to exercise the licensed rights everywhere in the world. Of course, no license is needed in countries and regions where the licensor does not possess IP protection for the licensed rights. A few countries lack patent laws entirely (e.g., Eritrea, Myanmar, Somalia), and it is only the most determined patentee that seeks and obtains patent protection in every country that does. In terms of copyright, 179 countries are parties to the international Berne Convention for the Protection of Literary and Artistic Works, but Iran, Iraq, Cambodia, Ethiopia and a handful of others are not. Moreover, national IP laws are not recognized in international waters, or in space. Thus, while truly “worldwide” licenses may be overkill, there is little downside in granting worldwide rights when the licensor does not wish to impose any territorial restriction on the licensee’s activities.
Below the global level, parties may subdivide the world largely as they see fit. The territory of a license grant may be a city, state, country or larger region. Parties, however, often run into trouble when they try to define territories beyond national borders. Ill-defined regions such as “Asia Pacific”Footnote 16 the “Middle East” and the “US West Coast” (are Alaska and Hawaii included?) frequently appear in term sheets and letters of intent, but often lead to disagreements regarding the precise countries included within their scope. Even regions that may seem well-understood can harbor traps for the unwary. For example, when asked how many countries are in North America, many people will respond “three – Canada, the USA and Mexico.” But this is incorrect. There are around forty countries that make up the North American continent, including Caribbean nations such as Cuba, Jamaica, Haiti and the Dominican Republic, the Central American countries of Panama, Costa Rica, Nicaragua, Honduras, El Salvador, Guatemala and Belize, as well as Bermuda, off the Atlantic coast of the United States, and the massive territory Greenland (currently held by Denmark).
The territory of “Europe” presents even more complexities. When speaking of Europe, one might mean the European Union (EU) (27 countries), the European Economic Area (the EU plus Iceland, Liechtenstein and Norway), the Eurozone (19 of the EU countries), the European Patent Convention (16 countries), or the traditional “continent” of Europe, which includes Russia, Ukraine and other countries that are not a part of any of the major European trading coalitions. Moreover, even the EU is fluid, as the recent exit of the UK (via Brexit) demonstrates. License agreements that defined the licensee’s territory as spanning the European Union suddenly contracted on January 31, 2020, when the UK exited the EU.
The territory of “North America” consists of about forty different countries.
Perhaps the most precise manner of defining the territory of a license agreement is to list the specific countries included in the territory in a schedule or exhibit to the agreement, though this approach can have its hazards as well. Consider, for example, the patent and know-how licenses sponsored by the Medicines Patent Pool (MPP), an arm of the UN’s World Health Organization. The MPP obtains licenses from multinational pharmaceutical companies for the manufacture and distribution of lifesaving drugs in the developing world. A company granting such a license could specifically list the “developing” countries to which the license applied. But countries change status occasionally. India and China are, by some measures, still developing countries, yet many companies would hesitate to lump them together with far poorer countries for essentially philanthropic purposes. Instead of listing countries, a licensor could refer to an external list or index, such as the Organisation for Economic Co-operation and Development (OECD) list of “least developed countries,” a list that changes periodically.
A final note of caution with respect to territory definition is to ensure that the granting of licenses within defined territories is not a cover-up for the allocation of markets among competitors, a violation of the antitrust laws (see Section 25.3). Outside of the United States, competition laws and regional agreements may also limit the ability of parties to divide rights territorially. For example, the EU requires the free movement of goods, services, capital and persons among member states of the Union. Accordingly, agreements that prevent a party in one EU country from shipping goods to, or providing services in, another EU country may be invalid.
There is no foolproof method of correctly defining the territory of a license agreement, other than to draft carefully and thoughtfully with the intentions of the parties in mind and a good atlas at hand.
6.3 Grant Clause
With the nature of the licensed rights, and the markets in which the licensee may operate, established, the “grant” clause of a license agreement sets forth the precise legal rights that are granted to the licensee.
Grant Clause [Patent]
Licensor hereby grants [1] to Licensee a nonexclusive, [nonassignable] [2] license [3] under the Licensed Patent Rights, excluding the right to sublicense [4], to make, use, sell, offer for sale and import Licensed Products throughout the Territory.
Grant Clause [Copyright]
Licensor hereby grants [1] to Licensee a nonexclusive, [nonassignable] [2] license [3], excluding the right to sublicense [4], to reproduce, distribute, publicly perform and make derivative works of the Licensed Works throughout the Territory.
Grant Clause [Trademark]
Licensor hereby grants [1] to Licensee a nonexclusive, [nonassignable] [2] license [3], excluding the right to sublicense [4], to reproduce and display the Licensed Marks, without alteration, on Approved Products throughout the Territory and on advertising and promotional materials, tangible and electronic, promoting the Approved Products in the Territory.
[1] Present grant – although Stanford v. Roche (discussed in Section 2.3, Note 3) involved an assignment of rights rather than a license, its lessons about clear present grants of rights hold equally true in the realm of licensing. Avoid variants in the grant clause such as “shall grant,” “agrees to grant” and the like.
[2] Assignability – many license grants include the term “nonassignable.” Doing so could, however, conflict with the express assignment clause usually contained toward the back of the agreement (see Section 13.3). Rather than attempt to sort out any contradictory language when a merger or other corporate transaction is on the horizon, it is preferable to omit “nonassignable” in the grant clause.
[3] Right and license – the grant is of a “license.” Some agreements state that a “right and license” is granted, but this is unnecessary.
[4] Sublicensing – some licenses may be sublicensed (see Section 6.5), and if so, there will be a separate, often lengthy, section on sublicensing. However, if the intent is to prohibit sublicensing, it is efficient to do so in the grant clause.
Note that with respect to rights that are granted under statutory forms of IP (especially patents and copyrights), it is important to follow the statutory rights that are inherent in the licensed assets. Specifically:
The Patent Act establishes that the owner of a patent has the exclusive right to make, use, sell, offer for sale and import a patented article (35 U.S.C. § 271(a)).
The Copyright Act establishes that the owner of a copyright has the exclusive right to reproduce, prepare derivative works, distribute, perform and display various types of copyrighted works (17 U.S.C. § 106).
The Lanham Act establishes that the registrant of a federal trademark or service mark has the exclusive right to use in commerce, reproduce, copy, and imitate the mark (15 U.S.C. § 1114).
Keeping these distinctions in mind is critical when drafting the grant clause. Thus, if a patent is being licensed, it is nonsensical to grant a licensee the right to “display” the patented article or to “produce derivative works” of it, as these are not rights granted under the Patent Act. Likewise, granting the licensee under a copyright the right to “use” the copyrighted work can cause no end of confusion, as demonstrated by the decision in Kennedy v. NJDA, discussed in Section 9.1 (interpreting the word “use” in a copyright license to encompass the making of derivative works).
It is also important to note that these rights can often be granted separately, and not all rights need be granted to every licensee. For example, some patent licenses permit use of a patented apparatus, but do not grant the licensee the right to make or sell that apparatus. By the same token, some exclusive patent licenses may grant the licensee an exclusive right to sell a licensed product, but do not extend exclusivity to the use of that product. Copyright licenses can be limited to the right to reproduce a work, but not to create derivative works of it.
For IP assets that are not statutorily defined, such as know-how, unregistered trademarks, rights of publicity, database rights and the like, the drafter can be more creative regarding the authority granted to the licensee. Yet this additional flexibility can also lead to disputes, so the drafter must pay particular attention to defining the rights granted as precisely as possible to achieve the client’s objectives.
The Cyrix case excerpted below illustrates the importance of precisely defining the scope of the license granted.
77 F.3d 1381 (Fed. Cir. 1996)
LOURIE, CIRCUIT JUDGE
Intel Corporation appeals from the decision of the United States District Court for the Eastern District of Texas entering judgment in favor of Cyrix Corporation, SGS-Thomson Microelectronics, Inc. (ST), and International Business Machines Corporation (IBM), and holding that IBM and ST acted within the scope of their respective patent license agreements with Intel when IBM made, and ST had made, products for Cyrix. [We] affirm.
Background
Cyrix designed and sold microprocessors. Since it did not have its own facility for manufacturing the microprocessors it designed, it contracted with other companies to act as its foundries. Under such an arrangement, Cyrix provided the foundries with its microprocessor designs, and the foundries manufactured integrated circuit chips containing those microprocessors and sold them to Cyrix. Cyrix then sold the microprocessors in the marketplace under its own brand name.
It was Cyrix’s practice to use manufacturing facilities of companies that were licensed under Intel’s patents. IBM was such a company; it had obtained a license to Intel’s patents in a patent license agreement dated October 1, 1989. The granting clause of the IBM–Intel agreement provided as follows:
2.2 Subject to the provisions of Sections 2.7 and 3.3, INTEL, on behalf of itself and its Subsidiaries, hereby grants to IBM a worldwide, royalty-free, nonexclusive license under the INTEL Licensed Patents:
2.2.1 to make, use, lease, sell and otherwise transfer IBM Licensed Products and to practice any method or process involved in the manufacture or use thereof;
…
2.2.3 to have made IBM Licensed Products … by another manufacturer for the use, lease, sale or other transfer by IBM.
The agreement defined “IBM Licensed Products” as follows:
1.23 “IBM Licensed Products” shall mean IHS Products, … Supplies and any combination of any, some or all of the foregoing …
Cyrix also used ST as a foundry. Initially, ST manufactured the chips, but when ST was unable to meet Cyrix’s demands, ST requested its affiliate in Italy, SGS-Thomson Microelectronics S.r.L. (ST-Italy), to manufacture the needed chips, which ST then sold to Cyrix.
ST was operating under a license agreement between Mostek and Intel, which ST acquired by assignment. The agreement contains the following granting clause:
INTEL grants and agrees to grant to MOSTEK non-exclusive, non-transferrable, world-wide licenses under INTEL PATENTS and INTEL PATENT APPLICATIONS to make, to have made, to use, to sell (either directly or indirectly), to lease and to otherwise dispose of LICENSED PRODUCTS.
The agreement defined “LICENSED PRODUCTS” as follows:
“LICENSED PRODUCTS” shall mean any product manufactured, used or sold by either party covered by patents of the other party.
It is undisputed that ST-Italy is legally not a “subsidiary” of ST and is thus not licensed under the ST–Intel agreement. ST therefore relied upon its “have made” rights to obtain products from ST-Italy, which it then sold to Cyrix to fulfill its contractual obligation.
Cyrix filed a declaratory judgment action against Intel, alleging a “reasonable apprehension” that it would be sued for patent infringement.Footnote 17 Cyrix sought a declaration that it did not infringe the Intel patents, claiming immunity on the ground that IBM and ST were both licensed under the patents. Cyrix’s view was that because IBM and ST acted within the scope of their respective licenses from Intel, its sales of microprocessors were shielded from any holding of infringement, the microprocessors having been obtained from authorized licensees.
IBM and ST intervened, seeking an adjudication of their rights under their respective agreements with Intel. On motions for summary judgment by Intel, IBM, and ST, the district court granted summary judgment for IBM and ST, and denied summary judgment for Intel. The district court also entered judgment for Cyrix.
The district court held that IBM had a right to act as a foundry in supplying microprocessors to Cyrix. It found that the definition of “IBM Licensed Products” in the IBM–Intel agreement did not limit the products it was licensed to sell to those designed by IBM. The district court distinguished Intel Corp. v. U.S. Int’l Trade Comm’n, 946 F.2d 821, 828 (Fed.Cir.1991) (”Atmel”) (construing the term “Sanyo … products” in a license agreement as limiting the grant of rights to Sanyo-designed and Sanyo-manufactured products). The district court concluded that, unlike the situation in Atmel, an internal conflict in the IBM–Intel agreement was not created by construing the license grant to cover products other than IBM-designed products. The court considered the facts to be more analogous to those in ULSI, rather than to those in Atmel.
The district court also held that ST had the right to have microprocessors made for it by any third party, including ST-Italy, and the right to sell those microprocessors to Cyrix. The district court found that the microprocessors were made for ST, not Cyrix, and that the supply agreement between ST and ST-Italy was not a sublicense that exceeded ST’s rights under the ST–Intel agreement. The district court thus distinguished the case that Intel cited in support of its position, E.I. du Pont de Nemours and Co. v. Shell Oil Co., 498 A.2d 1108, 1114–15 (Del. 1985) (holding that a third-party’s manufacturing of a product for itself under a licensee’s “have made” rights was a prohibited sublicense). This appeal followed.
Discussion
IBM–Intel Agreement
Intel argues that the IBM–Intel agreement does not support a grant of foundry rights. Intel relies upon the word “IBM” as modifying the term “licensed products” in arguing that this modifier is a so-called “Sanyo limitation,” limiting the scope of the products licensed and indicating that the parties did not intend to provide foundry rights. Intel also asserts that the “have designed” provision in the license does not provide IBM with the right to act as a foundry in manufacturing products designed by Cyrix.
The complex flow of license rights in Cyrix v. Intel.
Cyrix and IBM argue that the plain language of the IBM–Intel agreement grants to IBM the right to make and sell to Cyrix microprocessors that Cyrix designed. They argue that the “IBM” modifier in section 2.2.1 of the agreement was intended to distinguish “IBM Licensed Products” from “Intel Licensed Products,” and that “IBM Licensed Products” as defined in the agreement are not limited to those products specifically designed by IBM and made for itself. They argue that the term “IBM” used in the term “IBM Licensed Products” is not a “Sanyo limitation.”
We agree with the district court. The agreement granted IBM the right to make and sell “IBM Licensed Products,” which are defined elsewhere in the agreement and are not limited to products designed by IBM. Sections 2.2.1, which grants a license to sell “IBM Licensed Products,” and 1.23, which defines “IBM Licensed Products,” must be read together. When this is done, the granting provision essentially reads as follows:
2.2.1 to make, use, lease, sell and otherwise transfer IHS Products, … Supplies and any combination of any, some or all of the foregoing … and to practice any method or process involved in the manufacture or use thereof;
The products so defined are not limited to IBM-designed products. They include categories of products defined without the IBM prefix. The agreement defined these items as follows:
1.1 “Information Handling System” shall mean any instrumentality or aggregate of instrumentalities primarily designed to compute, classify, process, transmit, receive, retrieve, originate, switch, store, display, manifest, measure, detect, record, reproduce, handle or utilize any form of information, intelligence or data for business, scientific, control or other purposes.
1.2 “IHS Product” shall mean an Information Handling System or any instrumentality or aggregate of instrumentalities (including, without limitation, any component or subassembly) designed for incorporation in an Information Handling System;
1.4 “Supply” shall mean, as to each party hereto, any article or matter designed for use in or by, and adapted to be effectively consumed in the course of operation of an IHS Product licensed herein to that party.
Accordingly, we conclude that the district court correctly held that “IBM Licensed Products” are not limited to products designed by IBM.
We also do not agree with Intel that the “IBM” modifier is analogous to the “Sanyo limitation” in Atmel. The agreement in Atmel contained the following provision:
Intel hereby grants and will grant to Sanyo an [sic] non-exclusive, world-wide royalty-free license without the right to sublicense except to its Subsidiaries, under Intel Patents which read on any Sanyo Semiconductor Material, Semiconductor Device, Magnetic Bubble Memory Device, Integrated Circuit and Electronic Circuit products, for the lives of such patents, to make, use and sell such products.
We construed the term “Sanyo” to limit the products listed after that term. Such a construction was required because it gave meaning to the term “Sanyo” which was consistent with other provisions of the contract. Otherwise, the term “Sanyo” would have lacked meaning, and a contract must be construed if possible to give meaning to all its provisions. In contrast, the term “IBM Licensed Products” is thoroughly defined in the IBM–Intel agreement to provide no Sanyo-type limitation. Moreover, as argued by IBM, the “IBM” modifier is readily explained by its being distinguished from “Intel Licensed Products.”
This case is more analogous to ULSI than Atmel. In ULSI, Hewlett-Packard Company (HP) acted as a foundry to make and sell math coprocessor chips to ULSI. HP obtained a license to Intel’s patents under an agreement in which “each granted to the other an ‘irrevocable, retroactive, nonexclusive, world-wide, royalty-free license.’” ULSI sought to be shielded from infringement of Intel’s patents by purchasing the math coprocessor chips from HP, which was acting as an authorized seller. In concluding that HP’s agreement with Intel provided HP with the right to act as a foundry for ULSI, we stated that, in contrast to the “Sanyo limitation” discussed in Atmel, “the licensing agreement between Intel and HP here contains no restriction on HP’s right to sell or serve as a foundry.” There was no “Sanyo limitation” in ULSI. The products that were licensed were defined broadly. Notwithstanding the presence of the modifier “IBM,” the same is true here.
Intel also argues that section 2.2.3, providing a right to “have made” products only when the designs are furnished by IBM, limits IBM’s right to have products designed by Cyrix. IBM did not have the products made for it, and thus this provision does not limit its rights to make and have designed the products it sold to Cyrix. In summary, IBM properly made and sold microprocessors under section 2.2.1; IBM properly had microprocessors designed under section 2.2.2; and IBM did not “have made” microprocessors under the more limited section 2.2.3. Thus, IBM did not act outside the terms of the Intel agreement.
Intel also makes a policy argument premised on a preamble clause in its agreement with IBM in which the parties stated that “each expects to continue a research and development effort which will produce further patents and each may require a nonexclusive license under such patents of the other.” Intel argues that interpreting the agreement in favor of IBM would discourage the research the agreement was intended to foster. That argument totally misses the mark. The meaning of that clause is simply that the parties were entering into the agreement to facilitate their future research, i.e., to provide themselves with patent freedom for the future. Even if Intel never intended IBM to act as a foundry, this vague preamble cannot be interpreted to give effect to that intention if doing so would override clear operative language in the agreement. This agreement clearly gave IBM the right to make and sell to Cyrix microprocessors designed by Cyrix.
ST–Intel Agreement
Intel argues that the arrangement between ST and ST-Italy is in effect a sublicense, which it is clear is not permissible under the ST–Intel agreement. In particular, it argues that under ST’s “have made” rights, ST is only permitted to have products made for itself. Intel posits that the arrangement among ST, ST-Italy, and Cyrix was a mere paper transaction, i.e., a “sham.” See E.I. du Pont, 498 A.2d at 1116 (holding that a third party made a product for itself, not for a licensee, when it made a product and sold it to the licensee, who simultaneously sold it back to the third party).
ST and Cyrix argue that ST was acting within the scope of its “have made” rights. ST denies that its arrangement with ST-Italy was a “sham” and claims that it was using ST-Italy to manufacture products for it in order to meet its obligation to supply microprocessors to Cyrix. They distinguish du Pont on its facts, noting that in du Pont the party manufacturing under the “have made” right was also using the product itself, whereas here the product made under the “have made” right was sent to and eventually sold by the licensee.
We start with the clear proposition that, under its agreement, ST had the right to have the product made for it and to sell that product to third parties. It relied upon that right to have the product made by ST-Italy and to sell it to Cyrix. The district court found that the arrangement was distinguishable from that in du Pont. In du Pont, Carbide sought a license under du Pont’s patent to manufacture a product known as methomyl, but du Pont refused to grant Carbide a license. Carbide then entered into an agreement with Shell, du Pont’s licensee, whereby Carbide would manufacture methomyl for Shell under Shell’s “have made” rights and Shell would sell it back to Carbide. Carbide would then use it (or sell it) as it wished. The Supreme Court of Delaware, whose law governed that agreement, concluded that the two agreements, one to enable Carbide to manufacture methomyl for Shell and the other whereby Shell sold it back to Carbide, were two halves of a single business transaction. The net result was that they enabled Carbide to make and use the patented product. The court held that that was in effect a sublicense, which was prohibited under the Shell–du Pont agreement.
The district court identified several important differences between the situation in du Pont and the arrangement among ST, ST-Italy, and Cyrix, and concluded in its Memorandum Opinion and Order as follows:
The substance of the arrangement between Cyrix and ST and ST and ST-Italy is that when Cyrix needs wafers, it issues a purchase order to ST. ST then either manufactures the wafers itself at its Carrollton, Texas, facility or arranges for ST-Italy to manufacture the wafers at its Italian facility. ST is selling wafers. It is not selling or receiving payment for the use of its license from Intel. It has not authorized ST-Italy to make the wafers for or sell them to anyone other than ST. The production of the wafers is for the use of ST, the original licensee, and not for the use of ST-Italy. This is a valid exercise of the have-made rights granted under the License Agreement and does not constitute a sublicense.
We agree with the district court that the facts here are thoroughly distinguishable from those in du Pont. In du Pont, the arrangement was a sham. The third-party (Carbide) acting under Shell’s “have made” rights was manufacturing and selling the product to Shell and then buying it back in what was only a set of paper transactions. Here, however, the third-party (ST-Italy) properly manufactured microprocessors under ST’s “have made” rights, and ST then properly sold the products to a different entity, Cyrix. The two agreements, one permitting ST-Italy to manufacture microprocessors for ST and the other providing for ST’s sale of microprocessors to Cyrix, were separate business transactions. As the district court found, ST was using both its own facility and ST-Italy’s to satisfy its obligation to provide microprocessors to Cyrix. The products manufactured by ST-Italy were made for ST. If the facts in this case had been that Cyrix made the product for ST under ST’s “have made” rights and then ST sold the product back to Cyrix, then they would have been analogous to those in du Pont, but those are not our facts. We accordingly conclude that the district court did not err in holding that the arrangements among ST, ST-Italy, and Cyrix were a valid exercise of ST’s “have made” rights under its agreement with Intel. The district court thus did not err in granting a declaratory judgment of noninfringement in favor of Cyrix and ST.
The “have-made” arrangements in du Pont v. Shell.
AFFIRMED.
Notes and Questions
1. Sublicensing. Sublicenses, which are discussed in greater detail in Section 6.5, play a major role in the court’s analysis in Cyrix. For now, suffice it to say that a sublicense is a grant by a licensee of a portion of the rights that it has received from the licensor. Sublicensing, like subleasing in the context of real property, may be prohibited by the “primary” license between the licensor and the licensee (i.e., the sublicensor). Did Intel’s licensing arrangements with ST and IBM permit them to sublicense rights to Cyrix?
2. Foundry use and the Sanyo limitation. In the microelectronics industry, a “foundry” is a manufacturing facility where integrated circuits are manufactured to the order of a customer, usually using the customer’s specifications. Because Cyrix lacked a license from Intel, Cyrix provided integrated circuit designs to ST and IBM, both Intel licensees, for manufacture. In effect, ST and IBM were acting as foundries for Cyrix. Why was Intel concerned about this arrangement? What is a “Sanyo limitation” and why did Intel argue that the license included one?
3. Generality as permissiveness? The Cyrix court interpreted Intel’s license grant to IBM as including the right to make products designed by other entities. This determination was based in large part on the rather broad and general definitions given to the term “IBM Licensed Products.” The license grant clause reads:
[INTEL] hereby grants to IBM a worldwide, royalty-free, nonexclusive license under the INTEL Licensed Patents … to make, use, lease, sell and otherwise transfer IBM Licensed Products and to practice any method or process involved in the manufacture or use thereof …
If you were to redraft the grant in a manner more favorable to Intel, how would you do so to prevent IBM from acting as a foundry for Cyrix?
4. Have-made and foundry rights. Unlike the right to “make,” “use,” “offer to sell” and “sell” a patented article, the right to have the article made by a third party is not one of the exclusive rights granted to a patent holder under 35 U.S.C. § 271(a). Courts thus have some flexibility in interpreting the have-made right, but have often interpreted it, at least in the electronics industry, as specifically permitting a customer to have products manufactured by a foundry.Footnote 18 The have-made right then immunizes the foundry manufacturer from claims of infringement. Do you think that Intel granted IBM a “have-made right”? If not, would your interpretation change if Intel knew, at the time the license was negotiated, that IBM had all of its products fabricated by third parties? What is the difference between a “have-made” right and a sublicense?
5. More on have-made and foundry rights. If a have-made right has been granted, courts must often determine the limits of permitted foundry activity, and whether it includes the manufacture of products that are not made to the specifications of a particular customer, but are stock or off-the-shelf products. For example, in Thorne EMI N. Am. v. Hyundai Elec. Indus., 1996 U.S. LEXIS 21170 (Dist. Del. 1996), the court held that “a foundry commissioned by IBM to manufacture [Hyundai] products would have the protection of the license agreement, [but] a manufacturer of ‘off the shelf’ products is not a foundry … [and] therefore, whether or not it sold the products to IBM, would not be protected by the agreement.” Assuming that Intel granted a have-made right to IBM, was IBM operating within its scope as a foundry for Cyrix?Footnote 19
6. Have-made under copyright. Have-made rights are usually discussed in the context of patent licensing, but they can arise under copyright law as well. In Great Minds v. Fedex Office & Print Servs., Inc., 886 F.3d 91, 94 (2d Cir. 2018), the court dismissed a copyright infringement action against a commercial printer that copied materials at request of an authorized licensee, noting the “mundane ubiquity of lawful agency relationships.” Where the text of a license “provides no basis for distinguishing between” a licensee that directs its own employees to make copies versus one that “achieves an identical result by enlisting a temporary independent contractor—or a commercial duplication service,” the contractor is not liable for infringement. How does this reasoning work in terms of the exclusive rights granted under the Copyright Act, which do not include a right to “make,” but do include the right to “reproduce”? Does this decision effectively create a right to sublicense under copyright law, or should it be interpreted more narrowly, like the “have-made” right under patent law?
7. Branding as a restriction. Another way that licensors sometimes try to prevent their licensees from acting as third-party foundries is to limit the scope of their licenses to products bearing the licensee’s brands. Thus, Intel could have limited IBM’s license to the manufacture of products “marketed and sold under IBM’s brands.” This certainly would have prevented IBM from manufacturing Cyrix-branded chips. But how might such a restriction be circumvented by a determined licensee? Would such circumvention result in as effective a situation for the third-party customer?
8. Granting what you have the right to grant: the legal authority limitation. Recall the Spindelfabrik case from Section 6.1. Murata licensed the ’011 patent to Schubert, then assigned the patent and the Murata–Schubert license to Suessen. Murata did not own the ’946 patent. Schubert argued that Suessen, which did hold the ’946 patent, should be deemed to have licensed it to Schubert when Suessen acquired the ’011 patent and associated license. But the court disagreed, holding that the license could only convey to Schubert what the original licensor, Murata, could legally convey. Because Murata never held the ’946 patent, Murata could not license it to Schubert, and Suessen, which acquired the Murata–Schubert license, had no obligation to grant Schubert more than Murata did.
We discussed Spindelfabrik in the context of defining licensed rights (via the definition of “Control”). But the idea that a licensor cannot grant more than it holds also finds its way into license grant clauses. Consider the highlighted language in the following license grant.
Licensor hereby grants to Licensee, during the Term of this Agreement, and solely to the extent that Licensor has the authority to do so, a nonexclusive, nonassignable worldwide right and license under the Licensed Rights, excluding the right to sublicense, to make, use, sell, offer for sale and import Licensed Products.
The above clause limits the license grant to rights that the licensor has the legal authority to grant. At first blush, this limitation might seem tautological: of course the licensor can’t grant more rights than it has, as the court in Spindelfabrik emphasized. So is such a clause mere legal surplusage? Not exactly.
Suppose, for example, that “Licensed Rights” encompasses all of the licensor’s worldwide patent rights with respect to a particular technology. Also suppose that the licensor previously granted to Company A an exclusive right to use such technology in France. When the licensor grants further rights to Company B, it cannot grant Company B the right to use the technology in France. So rather than modify the grant to exclude France (and every other country and subfield in which it has granted rights to others), the licensor can simply limit the license to the rights that the licensor has the authority to grant to Company B.
Should Company B, the licensee, be concerned about such a limitation? Absolutely. But it can protect itself by insisting that the licensor list any previous license grants with respect to the licensed rights in a schedule (see Section 10.2.2, Note 6). How might such a disclosure protect the licensee?
9. Use not sell. As noted in the introduction to this part, some nonexclusive patent licenses grant the right to use but not to sell a licensed product, and some exclusive licenses grant exclusivity with respect to the right to sell, but not the right to use. What is the reasoning behind splitting the use and sale rights in this manner? How might the right to make a licensed product be addressed in these scenarios?
6.4 Changes to License Scope
Some IP rights – copyrights, trademarks, trade secrets – can last a very long time, sometimes in excess of a century and sometimes indefinitely. It is not surprising, therefore, that technologies and business practices that were contemplated when license agreements were drafted may change radically during the term of those agreements. How should unanticipated future uses be treated? The following cases explore this important issue.
145 F.3d 481 (2d Cir. 1998)
LEVAL, CIRCUIT JUDGE
Boosey & Hawkes Music Publishers Ltd., an English corporation and the assignee of Igor Stravinsky’s copyrights for “The Rite of Spring,” brought this action alleging that the Walt Disney Company’s foreign distribution in video cassette and laser disc format (”video format”) of the film “Fantasia,” featuring Stravinsky’s work, infringed Boosey’s rights. In 1939 Stravinsky licensed Disney’s distribution of The Rite of Spring in the motion picture. Boosey, which acquired Stravinsky’s copyright in 1947, contends that the license does not authorize distribution in video format … We hold that summary judgment was properly granted to Disney with respect to Boosey’s Lanham Act claims, but that material issues of fact barred the other grants of summary judgment. [We] remand all but the Lanham Act claim for trial.
Background
During 1938, Disney sought Stravinsky’s authorization to use The Rite of Spring (sometimes referred to as the “work” or the “composition”) throughout the world in a motion picture. Because under United States law the work was in the public domain, Disney needed no authorization to record or distribute it in this country, but permission was required for distribution in countries where Stravinsky enjoyed copyright protection. In January 1939 the parties executed an agreement (the “1939 Agreement”) giving Disney rights to use the work in a motion picture in consideration of a fee to Stravinsky of $6000.
The 1939 Agreement provided that:
In consideration of the sum of Six Thousand ($6,000.) Dollars, receipt of which is hereby acknowledged, [Stravinsky] does hereby give and grant unto Walt Disney Enterprises, a California corporation … the nonexclusive, irrevocable right, license, privilege and authority to record in any manner, medium or form, and to license the performance of, the musical composition hereinbelow set out
Under “type of use” in ¶ 3, the Agreement specified that
The music of said musical composition may be used in one motion picture throughout the length thereof or through such portion or portions thereof as the Purchaser shall desire. The said music may be used in whole or in part and may be adapted, changed, added to or subtracted from, all as shall appear desirable to the Purchaser in its uncontrolled discretion.
Igor Stravinsky and Walt Disney.
The Agreement went on to specify in ¶ 4 that Disney’s license to the work “is limited to the use of the musical composition in synchronism or timed-relation with the motion picture.”
Finally, ¶ 7 of the Agreement provided that “the licensor reserves to himself all rights and uses in and to the said musical composition not herein specifically granted” (the “reservation clause”).
Disney released Fantasia, starring Mickey Mouse, in 1940. The film contains no dialogue. It matches a pantomime of animated beasts and fantastic creatures to passages of great classical music, creating what critics celebrated as a “partnership between fine music and animated film.” The soundtrack uses compositions of Bach, Beethoven, Dukas, Schubert, Tchaikovsky, and Stravinsky, all performed by the Philadelphia Orchestra under the direction of Leopold Stokowski. As it appears in the film soundtrack, The Rite of Spring was shortened from its original 34 minutes to about 22.5; sections of the score were cut, while other sections were reordered. For more than five decades Disney exhibited The Rite of Spring in Fantasia under the 1939 license. The film has been re-released for theatrical distribution at least seven times since 1940, and although Fantasia has never appeared on television in its entirety, excerpts including portions of The Rite of Spring have been televised occasionally over the years. Neither Stravinsky nor Boosey has ever previously objected to any of the distributions.
In 1991 Disney first released Fantasia in video format. The video has been sold in foreign countries, as well as in the United States. To date, the Fantasia video release has generated more than $360 million in gross revenue for Disney.
Discussion
Boosey’s request for declaratory judgment raises … whether the general grant of permission under the 1939 Agreement licensed Disney to use The Rite of Spring in the video format version of Fantasia (on which the district court found in Disney’s favor) …
Boosey contends that the license to use Stravinsky’s work in a “motion picture” did not authorize distribution of the motion picture in video format, especially in view of the absence of an express provision for “future technologies” and Stravinsky’s reservation of all rights not granted in the Agreement. Disputes about whether licensees may exploit licensed works through new marketing channels made possible by technologies developed after the licensing contract – often called “new-use” problems – have vexed courts since at least the advent of the motion picture.
In Bartsch v. Metro-Goldwyn-Mayer, Inc., [391 F.2d 150 (2d Cir.1968)] we held that “licensees may properly pursue any uses which may reasonably be said to fall within the medium as described in the license.” 391 F.2d at 155. We held in Bartsch that a license of motion picture rights to a play included the right to telecast the motion picture. We observed that “if the words are broad enough to cover the new use, it seems fairer that the burden of framing and negotiating an exception should fall on the grantor,” at least when the new medium is not completely unknown at the time of contracting.
The 1991 VHS videotape version of Disney’s Fantasia sparked a dispute with Igor Stravinsky’s estate.
The 1939 Agreement conveys the right “to record [the composition] in any manner, medium or form” for use “in [a] motion picture.” We believe this language is broad enough to include distribution of the motion picture in video format. At a minimum, Bartsch holds that when a license includes a grant of rights that is reasonably read to cover a new use (at least where the new use was foreseeable at the time of contracting), the burden of excluding the right to the new use will rest on the grantor. The license “to record in any manner, medium or form” doubtless extends to videocassette recording and we can see no reason why the grant of “motion picture” reproduction rights should not include the video format, absent any indication in the Agreement to the contrary. If a new-use license hinges on the foreseeability of the new channels of distribution at the time of contracting – a question left open in Bartsch – Disney has proffered unrefuted evidence that a nascent market for home viewing of feature films existed by 1939. The Bartsch analysis thus compels the conclusion that the license for motion picture rights extends to video format distribution.
We recognize that courts and scholars are not in complete accord on the capacity of a broad license to cover future developed markets resulting from new technologies. The Nimmer treatise describes two principal approaches to the problem. According to the first view, advocated here by Boosey, “a license of rights in a given medium (e.g., ‘motion picture rights’) includes only such uses as fall within the unambiguous core meaning of the term (e.g., exhibition of motion picture film in motion picture theaters) and exclude any uses that lie within the ambiguous penumbra (e.g., exhibition of motion picture on television).” Under this approach, a license given in 1939 to “motion picture” rights would include only the core uses of “motion picture” as understood in 1939 – presumably theatrical distribution – and would not include subsequently developed methods of distribution of a motion picture such as television videocassettes or laser discs.
The second position described by Nimmer is “that the licensee may properly pursue any uses that may reasonably be said to fall within the medium as described in the license.” Nimmer expresses clear preferences for the latter approach on the ground that it is “less likely to prove unjust.” As Judge Friendly noted in Bartsch, “So do we.”
We acknowledge that a result which deprives the author-licensor of participation in the profits of new unforeseen channels of distribution is not an altogether happy solution. Nonetheless, we think it more fair and sensible than a result that would deprive a contracting party of the rights reasonably found in the terms of the contract it negotiates. This issue is too often, and improperly, framed as one of favoritism as between licensors and licensees. Because licensors are often authors – whose creativity the copyright laws intend to nurture – and are often impecunious, while licensees are often large business organizations, there is sometimes a tendency in copyright scholarship and adjudication to seek solutions that favor licensors over licensees. Thus in [Cohen v. Paramount Pictures, Inc., 845 F.2d 851 at 854 [(9th Cir. 1988)], the Ninth Circuit wrote that a “license must be construed in accordance with the purpose underlying federal copyright law,” which the court construed as the granting of valuable, enforceable rights to authors and the encouragement of the production of literary works. Asserting that copyright law “is enacted for the benefit of the composer,” the court concluded that it would “frustrate the purposes of the [copyright] Act” to construe the license as encompassing video technology, which did not exist when the license was granted.
In our view, new-use analysis should rely on neutral principles of contract interpretation rather than solicitude for either party. Although Bartsch speaks of placing the “burden of framing and negotiating an exception … on the grantor,” it should not be understood to adopt a default rule in favor of copyright licensees or any default rule whatsoever. What governs under Bartsch is the language of the contract. If the contract is more reasonably read to convey one meaning, the party benefited by that reading should be able to rely on it; the party seeking exception or deviation from the meaning reasonably conveyed by the words of the contract should bear the burden of negotiating for language that would express the limitation or deviation. This principle favors neither licensors nor licensees. It follows simply from the words of the contract.
The words of Disney’s license are more reasonably read to include than to exclude a motion picture distributed in video format. Thus, we conclude that the burden fell on Stravinsky, if he wished to exclude new markets arising from subsequently developed motion picture technology, to insert such language of limitation in the license, rather than on Disney to add language that reiterated what the license already stated.
Other significant jurisprudential and policy considerations confirm our approach to new-use problems. We think that our view is more consistent with the law of contract than the view that would exclude new technologies even when they reasonably fall within the description of what is licensed. Although contract interpretation normally requires inquiry into the intent of the contracting parties, intent is not likely to be helpful when the subject of the inquiry is something the parties were not thinking about. Nor is extrinsic evidence such as past dealings or industry custom likely to illuminate the intent of the parties, because the use in question was, by hypothesis, new, and could not have been the subject of prior negotiations or established practice. Moreover, many years after formation of the contract, it may well be impossible to consult the principals or retrieve documentary evidence to ascertain the parties’ intent, if any, with respect to new uses. On the other hand, the parties or assignees of the contract should be entitled to rely on the words of the contract. Especially where, as here, evidence probative of intent is likely to be both scant and unreliable, the burden of justifying a departure from the most reasonable reading of the contract should fall on the party advocating the departure.
Nor do we believe that our approach disadvantages licensors. By holding contracting parties accountable to the reasonable interpretation of their agreements, we encourage licensors and licensees to anticipate and bargain for the full value of potential future uses. Licensors reluctant to anticipate future developments remain free to negotiate language that clearly reserves the rights to future uses. But the creation of exceptional principles of contract construction that places doubt on the capacity of a license to transfer new technologies is likely to harm licensors together with licensees, by placing a significant percentage of the profits they might have shared in the hands of lawyers instead.
Neither the absence of a future technologies clause in the Agreement nor the presence of the reservation clause alters that analysis. The reservation clause stands for no more than the truism that Stravinsky retained whatever he had not granted. It contributes nothing to the definition of the boundaries of the license. And irrespective of the presence or absence of a clause expressly confirming a license over future technologies, the burden still falls on the party advancing a deviation from the most reasonable reading of the license to insure that the desired deviation is reflected in the final terms of the contract. As we have already stated, if the broad terms of the license are more reasonably read to include the particular future technology in question, then the licensee may rely on that language.
Bartsch therefore continues to articulate our “preferred” approach to new-use questions, and we hold that the district court properly applied it to find that the basic terms of Disney’s license included the right to record and distribute Fantasia in video format.
Notes and Questions
1. Other new uses. In Random House, Inc. v. Rosetta Books, LLC, 150 F. Supp. 2d 613 (S.D.N.Y. 2002), aff’d, 283 F.3d 490 (2d Cir. 2002), the court held that an agreement granting Random House the exclusive right to “print, publish and sell” certain works by William Styron, Kurt Vonnegut and other prominent authors “in book form” did not convey a right to release the works in electronic form as “ebooks.” The court explained,
Manifestly, paragraph #1 of each contract – entitled either “grant of rights” or “exclusive publication right” – conveys certain rights from the author to the publisher. In that paragraph, separate grant language is used to convey the rights to publish book club editions, reprint editions, abridged forms, and editions in Braille. This language would not be necessary if the phrase “in book form” encompassed all types of books. That paragraph specifies exactly which rights were being granted by the author to the publisher. Indeed, many of the rights set forth in the publisher’s form contracts were in fact not granted to the publisher, but rather were reserved by the authors to themselves. For example, each of the authors specifically reserved certain rights for themselves by striking out phrases, sentences, and paragraphs of the publisher’s form contract. This evidences an intent by these authors not to grant the publisher the broadest rights in their works.
The court distinguished Boosey & Hawkes and other early cases by characterizing them as encompassing within the licensed rights “new uses” within the “same medium as the original grant” (i.e., the display of a motion picture, whether on television or a videocassette). Ebooks, on the other hand, are “a separate medium from the original use – printed words on paper.” Do you agree with this distinction? Is the difference between ebooks and printed books very different than the difference between videotapes and cinematic films? Or is it, as the court claims, merely “a determination, relying on neutral principles of contract interpretation”?
William Styron’s Sophie’s Choice, one of the titles at issue in Random House v. Rosetta Books.
2. Tasini and database rights. The advent of digital formats such as databases and the Internet complicated the licensing of traditional print works such as newspaper articles. The Supreme Court in New York Times Co., Inc. v. Tasini, 533 U.S. 483 (2001) held that a newspaper that obtained the right to publish stories written by freelance journalists did not automatically obtain the right to place those stories in an online searchable database. After Tasini, a publisher specifically must obtain the right to publish the work both in the original newspaper or other compilation as well as on the Internet, in a database or in other digital formats. How do you think publishers reacted to the Tasini decision? Do you think that they made any changes to their standard agreements with freelance journalists?
Problem 6.3
Scent-o-Matic is a new technology for ebooks that gives users an olfactory overlay, such as providing the fragrance of baking bread in a recipe book or the aroma of a city back alley in a 1930s detective story. Scent-o-Matic is patented and works using software that causes certain scents to be produced when certain keywords are on the page. Scent-o-Matic is owned by Nile Books, a popular ebook publisher with existing licenses to deliver the numerous titles in its library to its ebook readers. The standard ebook publication license grants Nile Books the right to “reproduce and distribute the Work in English as an electronic book of the full-length verbatim text of the Work, including any illustrations, in a digital format. Such digital format may include necessary modifications to allow an end user to access, read, and interact with the Work in digital format.”
Analyze whether the Scent-o-Matic technology would be allowed under the license in the following circumstances.
a. The original ebooks are not altered – the Scent-o-Matic ebook reader contains a program that analyzes words on the screen and produces scents when it recognizes certain keywords.
b. Niles Books edits the original ebook by adding a non-visible notation to certain words such that when those words appear on the screen, Scent-o-Matic produces certain scents.
c. Your client is an established author with a new series of books soon to be published. She does not want Niles Books to deploy its Scent-o-Matic technology with her books. Redraft the language to make it clear that neither Scent-o-Matic nor any other technologies that add sensory inputs can be used with her books without her permission.
6.5 Sublicensing
A sublicense is a grant of rights by a licensee to a third party (the sublicensee) which encompasses some or all of the rights that have been granted to the licensee under a primary license agreement. Unlike an assignment of a license, the licensee that grants a sublicense generally remains bound by the terms of the original license. By the same token, the sublicense only exists so long as the underlying license remains in force.
Generally speaking, nonexclusive licensees may not grant sublicenses unless expressly permitted to do so in the primary license agreement. In some cases, however, exclusive licensees are permitted, under the law, to grant sublicenses without express permission from the licensor.Footnote 20 As a result, it is prudent, whether drafting an exclusive or a nonexclusive license, to specify whether, and to what degree, the licensee may grant sublicenses.Footnote 21
If sublicensing will be permitted under a licensing agreement, the licensor will often seek to impose some degree of control over the nature and identity of sublicensees.
Graphical representation of license and sublicense rights.
a. The licenses granted under this Agreement shall include the right to grant sublicenses without the consent of Licensor to any of Licensee’s Affiliates for so long as it remains an Affiliate of Licensee. Except as provided above, Licensee has no right to sublicense any licenses granted under this Agreement without the prior written consent of Licensor [1], which shall not be unreasonably withheld or delayed [2].
b. [Any sublicense granted by Licensee for any in-kind or nonmonetary consideration (including, but not limited to, services, equipment, supplies, usage of facilities, advertising, barter, bandwidth, data, intellectual property of any kind, releases from liability, options, interests in litigation, security interests, loans, debt forgiveness, covenants not to sue or to assert rights, software, technology, know-how, marketing rights, improvements, capital stock, units, partnership interests or other ownership interests in entities of any kind, or rights to receive dividends, revenue, royalties or other monies in the future) may be granted only with Licensor’s express prior written consent.] [3]
c. No sublicense shall relieve Licensee of its obligations under this Agreement, including the obligation to pay Licensor any and all fees, royalties and other amounts due. [4] Any breach of a sublicense agreement by the Sublicensee shall be deemed to constitute a breach of this Agreement by Licensee, and Licensee shall be liable for any action by a Sublicensee that would constitute a breach of this Agreement had it been committed by Licensee. [5]
d. Licensee shall provide a fully executed copy of each agreement pursuant to which it grants sublicense hereunder to Licensor immediately following its execution [6]. Without limiting the generality of the foregoing, each sublicense agreement shall provide that:
(i) Licensor shall have no responsibility, obligation or liability of any kind or manner to any Sublicensee;
(ii) Licensor shall be an express third party beneficiary of such sublicense, entitled to enforce it in accordance with its terms; [7]
(iii) Sublicensees shall have no further right to grant sublicenses of the rights granted under this Agreement; [8]
(iv) in the event of any inconsistency between the terms of a sublicense agreement and this Agreement, this Agreement shall control; [9]
(v) in the event that any Sublicensee (or any entity or person acting on its behalf) initiates any proceeding or otherwise asserts any claim challenging the validity or enforceability of any Licensed Right in any court, administrative agency or other forum, Licensee shall, upon written request by Licensor, terminate forthwith the sublicense agreement with such Sublicensee, and the sublicense agreement shall provide for such right of termination by Licensee; [10]
(vi) such sublicense shall terminate automatically upon the termination of this Agreement. [11]
(vii) the sublicensee shall be bound by provisions equivalent to those found in Sections xxx of this Agreement [e.g., audit, reporting, indemnification, non-competition, confidentiality, etc.]. [12]
[1] Approval rights – the licensor often retains the right to approve sublicensees, though in some cases a licensee’s Affiliates are automatically approved (this is especially the case when it is anticipated that the licensee will distribute a product through an international network of affiliated companies). In approving or rejecting sublicensees, the licensor must be careful to avoid potential antitrust issues that can arise from customer allocation and group boycotts (see Sections 25.3 and 25.7).
[2] Reasonableness – not all licensors will want to prove that their refusal to approve a sublicensee is “reasonable.” Accordingly, “shall not be unreasonably withheld … ” can be replaced by “, which approval licensor may extend or withhold in its sole discretion.” As a compromise, the agreement can specify types of sublicensees that are either prohibited outright, or require approval of the licensor (e.g., competitors in the licensor’s markets). Be careful, though. Naming specific companies to which sublicenses cannot be granted can run afoul of antitrust laws as concerted refusals to deal or group boycotts (see Section 25.7).
[3] Nonmonetary compensation – this clause is necessary to protect the licensor only if the licensee will pay the licensor a running royalty based on net sales or share sublicensing income with the licensor (see Section 8.4).
[4] No release from obligations – it is important that the licensee remain obligated to the licensor for all of its obligations under the prime licensing agreement. A sublicense is not intended to release the licensee from liability. If that were the case, the licensor could enter into a license agreement directly with the proposed sublicensee.
[5] Cross-breach – because the licensor lacks privity of contract with sublicensees, it is useful to attribute breaches by sublicensees to the prime licensee. Without such attribution, the licensor may have limited recourse against breaches by sublicensees (which may, in fact, be preferable to the licensee/sublicensor, who may argue that so long as it complies with its obligations under the prime license, the sublicensor–sublicensee relationship is not the concern of the prime licensor). See Section 12.3, Note 10, further discussing breach and termination by sublicensees.
[6] Copies – it is advisable for the licensor to obtain copies of all sublicenses granted when sublicensees will have substantial rights to exploit the licensed IP. It is unnecessary, for example, in the case of consumer end user sublicense agreements (see Section 17.1, Note 2). Licensees are sometimes reluctant to disclose sublicense agreements, but may agree to disclosure of redacted versions that contain at least the terms necessary to verify compliance with the sublicensing conditions, including financial terms when relevant to the licensor.
[7] Third-party beneficiary – under § 302(1) of the Restatement (Second) of Contracts, a third party’s capacity to sue under a contract depends on whether that party is an intended beneficiary of the contract. If the contracting parties intended that a third party benefit from performance of the contract, then that third party is an intended beneficiary and is entitled to enforce the contract. As a result, a licensor may seek to declare itself a third-party beneficiary of each sublicense agreement.
[8] No further sublicenses – this restriction, like the limitation on the number of sublicensees, seeks to contain the dissemination of the licensed IP and the prime licensor’s control over it.
[9] Precedence – see the discussion of order of precedence in Section 13.10.
[10] No challenge – see the discussion of no-challenge clauses in Section 22.4.
[11] Termination – it is often the case that sublicenses will terminate upon termination of the prime license, though there are exceptions – for example, in the case of software end user sublicenses. See Section 12.5.6 for a discussion.
[12] Pass-down obligations – if the licensor itself has licensed IP rights from a third party, then it may be required to pass down additional obligations to sublicensees.
The Cyrix case discussed in Section 6.1 turns to a great extent on whether or not ST was permitted to grant a sublicense to its affiliate in Italy. ST’s primary license with Intel appears to have prohibited sublicensing. If sublicensing is permitted, however, the sublicensor (the primary licensee) can clearly not grant the sublicensee more rights than the sublicensor obtained from the primary licensor. The following case explores what happens when a sublicensee acquires a sublicense from a licensee/sublicensor that itself may not be in good standing with the primary licensor.
284 F.3d 1323 (Fed. Cir. 2002)
DYK, CIRCUIT JUDGE
Rhône-Poulenc Agro, S.A. (“RPA”) appeals from the decision of the United States District Court for the Middle District of North Carolina granting summary judgment of non-infringement on the ground that Monsanto Co. (“Monsanto”) has a valid license to U.S. Patent No. 5,510,471 (“the ’471 patent”). The issue here is whether a sublicensee (Monsanto) that acquired the sublicense from a licensee (DeKalb Genetics Corp. (“DeKalb”)), that acquired the original license by fraud, may retain the sublicense by establishing that the sublicensee was a bona fide purchaser for value …
We hold that the bona fide purchaser defense is governed by federal law and is not available to non-exclusive licensees in the circumstances of this case. Accordingly, we vacate the decision of the district court and remand for further proceedings consistent with this opinion.
Background
From 1991 through 1994, RPA and DeKalb collaborated on the development of biotechnology related to specific genetic materials. During this time, a scientist at RPA, Dr. DeRose, developed an optimized transit peptide (“OTP”) with a particular maize gene, which proved useful in growing herbicide resistant corn plants. The OTP is covered by the claims of the ’471 patent and is the subject of RPA’s patent infringement claim against Monsanto.
In 1994, RPA, DeKalb, and non-party Calgene, Inc. (“Calgene”) entered into an agreement (the “1994 Agreement”) that provided:
RPA and CALGENE hereby grant to DEKALB the world-wide, paid-up right to use the RPA/CALGENE Technology and RPA/CALGENE Genetic Material in the field of use of corn. DEKALB shall have the right to grant sublicenses to the aforementioned right to use without further payment being made to RPA or CALGENE.
The RPA/CALGENE Technology and RPA/CALGENE Genetic Material included the invention claimed in the ’471 patent. In 1996, DeKalb sublicensed its rights to the RPA/Calgene Technology and Genetic Material to Monsanto. At the same time Monsanto granted to DeKalb licenses to use certain intellectual property related to genetically-engineered corn …
On October 30, 1997, RPA filed suit against DeKalb and Monsanto, seeking, inter alia, to rescind the 1994 Agreement on the ground that DeKalb had procured the license (the “right to use”) by fraud. RPA also alleged that DeKalb and Monsanto were infringing the ’471 patent and had misappropriated RPA’s trade secrets. Monsanto defended, inter alia, on the ground that it had a valid license to practice the invention of the patent and use the trade secrets, based on the rights owned under the 1994 Agreement that were transferred by DeKalb to Monsanto in 1996. At trial, a jury found, inter alia, that DeKalb had fraudulently induced RPA to enter into the 1994 Agreement. The district court ordered rescission of the 1994 Agreement. Nonetheless, Monsanto moved the district court for summary judgment that it had a valid license to the ’471 patent and the right to use RPA’s trade secrets because under the 1996 Agreement Monsanto was a bona fide purchaser for value of the sublicense to the patent and the trade secrets. The district court … granted this motion and dismissed the infringement and misappropriation claims against Monsanto.
The district court found that, as a sublicensee of the ’471 patent and the trade secrets, Monsanto was “entitled to be considered a bona fide purchaser, because it paid value for the right to use the technology without knowledge of any wrongdoing by DeKalb.” Because “Monsanto [was] a bona fide purchaser of the … technology, [it] therefore [could not] be liable as a patent infringer or a trade secret misappropriater.” The district court explicitly did not reach the issues of whether Monsanto’s bona fide purchaser defense would apply to any future licenses of RPA’s technology or whether, in light of the 1994 RPA–DeKalb–Monsanto Agreement granting DeKalb the right to sublicense, the bona fide purchaser defense would benefit sublicensees of Monsanto.
RPA filed this timely appeal, which concerns only the validity of Monsanto’s license to practice the ’471 patent.
Discussion
In Rhône-Poulenc I, we affirmed the judgment of the district court, rescinding the 1994 licensing agreement based on a jury verdict finding that DeKalb acquired its patent license by fraud. RPA asserts that it necessarily follows that the Monsanto sublicense to the ’471 patent is void, and that Monsanto can be sued for patent infringement. We agree …
35 U.S.C. § 261 … provides that a later bona fide purchaser for value without notice (a later assignee) prevails if the earlier assignment was not timely recorded in the patent office.Footnote 22 This case, however, involves a different situation – the circumstance in which the interest in the patent held by the grantor is voidable and the question is whether a grantee may retain its interest even if the grantor’s interest is voided. Section 261 does not directly govern the resolution of this question.
Since section 261 does not apply directly, we must turn to other provisions of the Patent Act. Section 271 of the Act provides: “whoever without authority makes, uses, offers to sell, or sells any patented invention … infringes the patent.” 35 U.S.C. § 271(a). We are charged with the task of determining the meaning of the term “without authority.” Under this provision, as under other provisions of the Patent Act, the courts have developed a federal rule, where appropriate, and have deferred to state law, where that is appropriate. This issue of whether to apply state or federal law has particular importance in this case because North Carolina state law, the law of the forum state, does not recognize a bona fide purchaser defense unless there has been a title transfer.
In general, the Supreme Court and this court have turned to state law to determine whether there is contractual “authority” to practice the invention of a patent. Thus, the interpretation of contracts for rights under patents is generally governed by state law. Aronson v. Quick Point Pencil Co., 440 U.S. 257, 262 (1979); Lear, Inc. v. Adkins, 395 U.S. 653, 661–62 (1969). Just as the interpretation of patent license contracts is generally governed by state law, so too the consequences of fraud in the negotiation of such contracts is a matter generally governed by state law. It may be argued that the impact of fraud upon the validity of a license as against a bona purchaser defense should also be governed by state law. However, we confront here a unique situation in which a federal patent statute explicitly governs the bona fide purchaser rule in some situations but not in all situations. It would be anomalous for federal law to govern that defense in part and for state law to govern in part. There is quite plainly a need for a uniform body of federal law on the bona fide purchaser defense.
On the related question of the transferability of patent licenses, many courts have concluded that federal law must be applied. In so holding, courts generally have acknowledged the need for a uniform national rule that patent licenses are personal and non-transferable in the absence of an agreement authorizing assignment, contrary to the state common law rule that contractual rights are assignable unless forbidden by an agreement.
In short, because of the importance of having a uniform national rule, we hold that the bona fide purchaser defense to patent infringement is a matter of federal law. Because such a federal rule implicates an issue of patent law, the law of this circuit governs the rule. Of course, the creation of a federal rule concerning the bona fide purchaser defense is informed by the various state common law bona fide purchaser rules as they are generally understood.
Congress has specifically provided that patents are to be treated as personal property. 35 U.S.C. § 261. At common law, a bona fide purchaser (also known as a “good faith buyer”) who acquired title to personal property was entitled to retain the property against the real owner who had lost title to the property, for example, by fraud. Generally, a bona fide purchaser is one who purchases legal title to property in good faith for valuable consideration, without notice of any other claim of interest in the property. The bona fide purchaser rule exists to protect innocent purchasers of property from competing equitable interests in the property because “[s]trong as a plaintiff’s equity may be, it can in no case be stronger than that of a purchaser, who has put himself in peril by purchasing a title, and paying a valuable consideration, without notice of any defect in it, or adverse claim to it …. ” Boone v. Chiles, 35 U.S. 177, 210 (1836).
At common law, however, it was quite clear that one who did not acquire title to the property could not assert the protection of the bona fide purchaser rule. Many courts have held that a party to an executory contract to purchase title, the owner of a lease, or a purchaser from a vendor who did not have title cannot benefit from the bona fide purchaser rule. It is clear under the law of North Carolina (the state in which RPA filed suit) that “[i]n the absence of an estoppel, one is not entitled to protection as a bona fide purchaser unless he holds the legal title to the property in dispute.”
Monsanto urges that the cases requiring that one obtain title to benefit from the bona fide purchaser defense are “antiquated,” and the Uniform Commercial Code’s (“U.C.C.”) modern approach has rejected the requirement of title. In fact, the title rule is recognized in modern property law, and has been confirmed by the U.C.C. Under U.C.C. Article 2-403, even “[a] person with voidable title has power to transfer a good title to a good faith purchaser for value.”
Monsanto also relies on statements from various treatises on patent licensing for the proposition that a sublicense continues, even when the principal license is terminated. But the statements address the situation where the original licensee is terminated as a matter of contract law, e.g., for breach of contract. These treatises do not address the operation of the bona fide purchaser rule with respect to sublicenses and do not state or suggest that a sublicense continues even when the principal license is rescinded because it has been obtained by fraud.
Even if the general common law extended the protection of the bona fide purchaser rule to holders of non-exclusive licenses, it would not be appropriate for us to extend such protection to non-exclusive licenses as a matter of federal common law. Section 261 of title 35 reflects a determination by Congress that only those who have obtained an “assignment, grant or conveyance” may benefit from the protection of the statute. This provision thus reflects a congressional judgment that the protections of the bona fide purchaser rule extend only to those who have received an “assignment, grant or conveyance.” Under such circumstances, the Supreme Court has made clear that we must consider the purposes of federal statutes in framing a rule of federal common law, even if the statutes are not directly applicable.
Although our precedent has recognized that in some circumstances an exclusive patent license may be tantamount to an assignment of title to the patent, this is so only when “the licensee holds ‘all substantial rights’ under the patent.” Textile Prods., Inc. v. Mead Corp., 134 F.3d 1481, 1484 (Fed. Cir.), cert. denied, 525 U.S. 826 (1998). Here the license is non-exclusive, and there is no contention that the license agreement transferred “all substantial rights.” Thus, an assignment did not occur, and in the absence of an “assignment, grant or conveyance,” Congress contemplated that there would be no bona fide purchaser defense.
Conclusion
In sum, the bona fide purchaser defense does not apply to non-exclusive licensees. We accordingly vacate the decision of the district court and remand for further proceedings consistent with this opinion.
Notes and Questions
1. Exclusive versus nonexclusive licensee. The Federal Circuit in Rhone-Poulenc v. DeKalb holds that the bona fide purchaser defense does not apply to nonexclusive licensees. Why? Does the court imply that a different result might apply to exclusive licensees? Do you agree?
2. Section 261 and bona fide purchasers. Much like state recording statutes for real property, Section 261 of the Patent Act provides that a purchaser of a patent without notice of a prior sale will prevail over a previous purchaser of the same patent if the prior sale was not recorded at the Patent and Trademark Office within three months of the purchase. Why do you think that Congress enacted this rule? How does this rule differ from a traditional state “race-notice” or “notice” recording statutes for real property?
3. Termination of sublicenses. In arguing that its sublicense should continue notwithstanding DeKalb’s original fraudulent license acquisition, Monsanto relies on treatise authors who suggest that a valid sublicense should continue notwithstanding the termination of the primary license. The Federal Circuit in Rhone-Poulenc sidestepped this issue, noting that the question was not whether Rhône-Poulenc’s sublicense was terminated, but whether it was ever valid in the first place, considering DeKalb’s original fraudulent license. For some time after this decision, it was unclear whether a sublicense would survive the termination of its primary license. In 2018, however, the Federal Circuit clarified its position, holding in Fraunhofer-Gesellschaft v. Sirius XM Radio, 940 F.3d 1372, 1380 (Fed. Cir. 2018) that “our law does not provide for automatic survival of a sublicense” and expressly rejecting any implication to the contrary in Rhone-Poulenc. Which default rule do you find more persuasive: that sublicenses do or do not automatically survive the termination of the primary license?
Of course, the parties themselves provide for the survival of sublicenses by clearly stating in the primary license agreement that all sublicenses will, or will not, terminate upon termination of the primary license. See Section 12.3, Note 9, and Section 12.E.6, which discuss in greater detail issues surrounding the breach and termination of sublicenses.
4. Licensor’s approval of sublicenses. If sublicensing is permitted under a primary license, the parties will sometimes agree to include the template for the sublicensing agreement that the licensee/sublicensor must use as an exhibit to the primary license. This gives the licensor comfort that the terms that its licensee will grant to sublicensees are understood and agreed up-front. In other cases, when a template sublicense agreement is not attached, the licensor may reserve the right to review and approve any sublicense agreements or individual sublicensees. If the primary licensor reserves the right to approve sublicensees, the reasons for rejection should usually be spelled out in the primary agreement so as to avoid allegations of anticompetitive behavior (see Chapter 25 relating to antitrust considerations in license agreements).
5. Sublicensing in the biotech industry. One of the industries in which sublicensing is standard practice is biotechnology. In many cases, a university will grant a license to a biotechnology company, which is sometimes a university spinout or start-up founded by university researchers (see Section 14.3). The scope of this license is often broad and exclusive, covering the entire output of a particular university laboratory. The biotech company will then continue the research begun by the university, often working alongside university researchers. The biotech company’s goal is to develop or discover promising drug or diagnostic candidates that it can then sublicense on an individual basis to a larger pharmaceutical company, which will have the resources to conduct the large-scale clinical trials necessary to secure FDA approval for the product. Sometimes, the pharmaceutical company will license several compounds or drug candidates (each a different FOU) from the biotech company, often on an exclusive basis. The pharmaceutical company may also obtain an option to acquire licenses in additional FOUs, typically upon payment of an option fee. When a biotech company announces that it has signed a large deal with a pharmaceutical company, it is usually counting on the exercise of all such options, the payment of all milestone payments and an estimate of the royalty revenue that it will receive. As a licensee/sublicensor, the biotech company itself will be obligated to pay a portion of its earnings from the pharmaceutical company back to the university holding the patents and the primary license. This business pattern has been used for the last thirty years and has, to a large degree, defined the modern biotechnology industry. Nevertheless, as discussed in Section 14.3, Note 1, universities have been criticized for granting sublicenses of such breadth to for-profit companies that are not obliged to abide by the universities’ public missions. What alternative licensing and sublicensing structures might exist to address these concerns?
Open access