Hachette Book Grp. Inc. v. Internet Archive

Defendant-Appellant Internet Archive appeals from a judgment of the United States District Court for the Southern District of New York (Koeltl, J.) denying its motion for summary judgment and granting Plaintiffs-Appellees’ motion for summary judgment. Internet Archive creates digital copies of print books and posts those copies on its website where users may access them in full, for free, in a service it calls the “Free Digital Library.” Other than a period in 2020, Internet Archive has maintained a one-to-one owned-to-loaned ratio for its digital books: Initially, it allowed only as many concurrent “checkouts” of a digital book as it has physical copies in its possession. Subsequently, Internet Archive expanded its Free Digital Library to include other libraries, thereby counting the number of physical copies of a book possessed by those libraries toward the total number of digital copies it makes available at any given time. Plaintiffs-Appellees — four book publishers — sued Internet Archive in 2020, alleging that its Free Digital Library infringes their copyrights in 127 books and seeking damages and declaratory and injunctive relief. Internet Archive asserted a defense of fair use under Section 107 of the Copyright Act. The district court rejected that defense and entered summary judgment for Plaintiffs. This appeal presents the following question: Is it “fair use” for a nonprofit organization to scan copyright-protected print books in their entirety, and distribute those digital copies online, in full, for free, subject to a one-to-one owned-to-loaned ratio between its print copies and the digital copies it makes available at any given time, all without authorization from the copyright-holding publishers or authors? Applying the relevant provisions of the Copyright Act as well as binding Supreme Court and Second Circuit precedent, we conclude the answer is no. We therefore AFFIRM. BETH ROBINSON, C.J. Defendant-Appellant Internet Archive appeals from a judgment of the United States District Court for the Southern District of New York (Koeltl, J.) denying its motion for summary judgment and granting Plaintiffs-Appellees’ motion for summary judgment. Internet Archive creates digital copies of print books and posts those copies on its website where users may access them in full, for free, in a service it calls the “Free Digital Library.” Other than a period in 2020, Internet Archive has maintained a one-to-one owned-to-loaned ratio for its digital books: Initially, it allowed only as many concurrent “checkouts” of a digital book as it has physical copies in its possession. Through its Open Libraries Project, Internet Archive subsequently expanded its Free Digital Library to include partner libraries’ collections, thereby counting physical copies of a book possessed by partner libraries towards the total number of digital copies it makes available at any given time. Plaintiffs-Appellees — four book publishers — sued Internet Archive in 2020, alleging that its Free Digital Library infringed their copyrights in 127 books and seeking damages and declaratory and injunctive relief. Internet Archive asserted a defense of fair use under Section 107 of the Copyright Act. The district court rejected that defense and entered summary judgment for Plaintiffs. This appeal presents the following question: is it “fair use” for a nonprofit organization to scan copyright-protected print books in their entirety and distribute those digital copies online, in full, for free, subject to a one-to-one owned-to-loaned ratio between its print copies and the digital copies it makes available at any given time, all without authorization from the copyright-holding publishers or authors? Applying the relevant provisions of the Copyright Act as well as binding Supreme Court and Second Circuit precedent, we conclude the answer is no. We therefore AFFIRM. BACKGROUND¹ Plaintiffs-Appellees Hachette Book Group, Inc., HarperCollins Publishers LLC, John Wiley & Sons, Inc., and Penguin Random House LLC (“Publishers”) are four of the leading book publishers in the United States. They obtain from authors the exclusive right to publish their works in multiple formats, including hardcover, paperback, and eBook. Publishers and their authors profit from this arrangement via sales in each format. In addition to selling traditional print books, Publishers collectively invest millions of dollars in developing new formats and markets suited for the digital age, including the eBook market. Publishers distribute eBooks in two principal ways. First, they sell eBooks direct to consumer through electronic retail platforms such as the Amazon Kindle Store. Second, Publishers — via commercial distributors such as OverDrive — license eBooks to libraries under a one-copy, one-user model that allows each eBook to be “checked out” by one patron at a time, in a similar fashion to traditional library lending of print books. Library patrons access eBooks through distributors’ online platforms, like OverDrive’s Libby app. Publishers require that libraries lend each eBook only to their own verified library members, who are residents of the geographical areas served by the library. The library eBook lending market is thriving. Checkouts of eBooks on OverDrive by library patrons increased dramatically between 2010 and 2020. This surge in lending translates to greater profits for Publishers, some of whom find library eBook licenses occupying an increasing percentage of their overall eBook revenues. With more than 93 percent of public libraries participating in eBook lending, Publishers and their authors have tapped in to a profitable, growing market. The story is a bit more complicated for libraries. When libraries acquire print books, they may circulate those books for as much time and as many borrows as they desire, until the books wear down beyond usefulness. At first, Publishers offered eBook licenses to libraries on a perpetual basis, allowing them to lend eBooks in a similar manner to print books, though without the same concerns for wear and tear. Now, Hachette offers two-year licenses, Penguin offers two-year and pay-per-use licenses, HarperCollins offers 26-borrows and pay-per-use licenses, and Wiley offers perpetual licenses and licenses for a set number of days or uses. Although Hachette, Penguin, and HarperCollins still offer perpetual licenses to academic libraries, they stopped offering perpetual licenses to public libraries in 2019, 2018, and 2011, respectively. While some Publishers price their eBook licenses the same as print books, others price their licenses higher than the suggested retail price of print books or even consumer eBooks.² For libraries, the result is regular renegotiation of eBook licenses that often come at a steeper price and for a shorter term than print copies of the same books. Enter Internet Archive: a nonprofit whose stated mission “is to provide universal access to all knowledge.” App’x 5778. Internet Archive (“IA”) “works with libraries, museums, universities, and the public to preserve and offer free online access to texts, audio, moving images, software, and other cultural artifacts.” Id. at 5779. One of IA’s longest running projects is the “Wayback Machine,” which archives public webpages that might otherwise be lost to history. In 2011, IA launched the project at issue in this appeal: the “Free Digital Library.” Appellant’s Br. at 5. Starting in 2011, IA partnered with the Open Library of Richmond, a nonprofit, and Better World Books (“BWB”), a for-profit used bookstore, to acquire print books for digitizing and posting on IA’s website. The Open Library of Richmond buys or accepts donations of print books — primarily from BWB — and sends those books to IA scanning centers, where operators scan each page using a book-digitization device. The Open Library of Richmond holds legal title to and maintains physical possession of the print books, which are placed in shipping containers and kept out of circulation after scanning. Once IA scans a book, it posts the digital copy³ on its website, where IA account holders may access it in full, for free. IA does not charge a fee to open an account or for any subsequent service, including digital book borrowing. Account holders may check out up to ten digital books at a time for up to fourteen days each, and may read those copies on IA’s BookReader web browser platform or download an encrypted PDF or EPUB version of the digital book. IA represents that it secures these downloadable versions with software that “allows only the authorized patron to download and read the borrowed book…and prevents the patron from copying or further distributing the book or from accessing it after their loan has ended.” App’x 5788. Other than a period in 2020, IA has maintained a one-to-one owned-to-loaned ratio for its digital books: Prior to 2018, it allowed only as many concurrent “checkouts” of a digital book as it has physical copies in storage. IA refers to this practice as “Controlled Digital Lending” or “CDL,” which it likens to traditional library lending of print books. In 2018, IA launched the “Open Libraries Project,” allowing libraries to “contribute” their noncirculating print books to the number of concurrent checkouts available on IA’s website. Participating libraries send a catalog of their noncirculating books to IA, which then runs an “overlap analysis.” App’x 6095. If the library’s catalog includes a book for which IA already has a digital copy, IA increases the number of available concurrent checkouts by one.⁴ Participating libraries may, in turn, integrate links to IA’s digital books in their own catalogs. IA asserts that the Open Libraries Project allows it to expand its lending capacity while remaining within the confines of CDL. In other words, IA still maintains a one-to-one owned-to-loaned ratio, but the category of books “owned” now includes those books owned by participating libraries. In 2020, in response to the COVID-19 pandemic, IA launched the “National Emergency Library” (the “NEL”). Out of concern for a potential lack of access to books due to the closure of schools and libraries, IA lifted its one-to-one owned-to-loaned ratio, allowing its digital books to be checked out by up to 10,000 users at a time, without regard to the corresponding number of physical books in storage or in partner libraries’ possession — a practice IA acknowledges was a “deviat[ion] from controlled digital lending.” Appellant’s Br. at 9. The NEL ran from March 24, 2020, to June 16, 2020, when IA reinstated its lending controls after this lawsuit was filed. IA markets its lending services to libraries as a free alternative to Publishers’ print books and eBook licenses. For example, in a presentation to libraries, IA’s Director of Open Libraries noted that the Open Libraries Project “ensures [libraries] will not have to buy the same content over and over, simply because of a change in format.” App’x 6099. Another presentation advertising the Open Libraries Project bore the title: “Maximizing institutional investments in print resources through controlled digital lending,” with the subtitle, “Or, You Don’t Have to Buy it Again!” Id. at 6099. In another instance, a vendor acting on behalf of IA emailed libraries asking whether they “would like to participate [in the Project] and get free ebooks.” Id. at 6100; see also Appellant’s Br. at 5 (referring to its lending services as the “Free Digital Library”). As of December 2021, 62 partner libraries participated in the Open Libraries Project. IA hosts over 3.2 million digital copies of copyrighted books on its website. Its 5.9 million users effectuate about 70,000 book “borrows” a day — approximately 25 million per year. Critically, IA and its users lack permission from copyright holders to engage in any of these activities. They do not license these materials from publishers, nor do they otherwise compensate authors in connection with the digitization and distribution of their works. On June 1, 2020, Publishers sued IA and five Doe defendants in the Southern District of New York, alleging that IA infringed their copyrights in 127 books (the “Works” or the “Works in Suit”). The Works in Suit are published fiction and nonfiction books ranging from Sandra Cisneros’ The House on Mango Street (1984), Sylvia Plath’s The Bell Jar (1963), and J.D. Salinger’s The Catcher in the Rye (1951), to Les Carter’s Enough About You, Let’s Talk About Me: How to Recognize and Manage the Narcissists in Your Life (2005), and Ian Woofenden’s Wind Power for Dummies (2009). The authors of the Works in Suit each assigned to one of the Publishers the exclusive right to publish their works in print and eBook formats. All 127 Works are available as authorized eBooks that may be purchased by consumers or licensed to libraries; and, at the time of the complaint, all 127 Works were available as free digital copies on IA’s website. IA answered the complaint, denying that its practices violate Publishers’ copyrights and asserting an affirmative defense of fair use under the Copyright Act.⁵ See 17 U.S.C. §107. Following extensive discovery, the parties cross-moved for summary judgment. On March 24, 2023, the district court granted summary judgment to Publishers, concluding that Publishers established the elements of copyright infringement and that IA’s infringement was not excused by the defense of fair use. Hachette Book Group, Inc. v. Internet Archive, 664 F. Supp. 3d 370, 378, 391 (S.D.N.Y. 2023). In evaluating the four statutory fair use factors set forth in Section 107, the district court found that: (1) IA’s use of the Works in Suit is (a) non-transformative because IA reproduces the Works in full and its digital copies serve the same purpose as the originals; and (b) commercial because, despite IA’s nonprofit status, it exploits the Works by soliciting donations on its website and taking a cut of the proceeds when users buy a physical book from BWB using a link embedded on IA’s website; (2) the Works are original fiction and nonfiction books “close to the core of intended copyright protection”; (3) IA copies the books wholesale; and (4) IA “brings to the marketplace a competing substitute” for library eBook licenses, “usurping a market that properly belongs to the copyright-holder.” Id. at 380-91 (alterations adopted). As each factor favored Publishers, the district court concluded that IA’s fair use defense failed as a matter of law. Id. at 391. It further determined that the above analysis applied “even more forcefully to the NEL.” Id. The district court entered judgment on August 11, 2023, including a permanent injunction barring IA from, among other things, distributing or reproducing Publishers’ copyrighted works, including but not limited to the Works in Suit. The district court specifically limited its judgment to print books that, like the 127 Works at issue in the suit, are also “available for electronic licensing.” No. 1:20-cv-04160-JGK, Dkt. 216, at 3. IA timely appealed. DISCUSSION “We review without deference the district court’s grant of summary judgment when, as here, the parties filed cross-motions for summary judgment and the district court granted one motion but denied the other.” Loomis v. ACE American Insurance Company, 91 F.4th 565, 572 (2d Cir. 2024) (internal quotation marks omitted). Summary judgment should be granted if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a). Although fair use “presents a mixed question of law and fact, it may be resolved on summary judgment where, as here, the material facts are not in dispute.” Andy Warhol Foundation for Visual Arts, Inc. v. Goldsmith, 11 F.4th 26, 36 (2d Cir. 2021) (“Warhol I”), aff’d, 598 U.S. 508 (2023). The Constitution empowers Congress to enact copyright laws “[t]o promote the Progress of Science and useful Arts.” U.S. Const. art. I, §8, cl. 8. To that end, the Copyright Act of 1976 grants the author of an original work “a bundle of exclusive rights,” Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539, 546 (1985), including the rights “to reproduce the copyrighted work,” “to prepare derivative works,” to distribute copies of the work via sale, rental, lease, or lending, and, in the case of literary works, “to display the copyrighted work publicly,” 17 U.S.C. §106. The Act defines derivative works “largely by example, rather than explanation.” Authors Guild v. Google, Inc., 804 F.3d 202, 215 (2d Cir. 2015) (hereinafter “Google Books”). Those examples include “translation, musical arrangement, dramatization, fictionalization, motion picture version, sound recording, art reproduction, abridgment, condensation, or any other form in which a work may be recast, transformed, or adapted.” 17 U.S.C. §101. Copyright infringement occurs when a person or entity “violates any of the exclusive rights of the copyright owner.” 17 U.S.C. §501(a). The monopoly created by the Copyright Act “rewards the individual author in order to benefit the public”; the idea being that authors and inventors will be more motivated to produce new works if they know those works will be protected, and the public will benefit from both restricted access to those works in the short term and unfettered access in the long term, once the period of exclusive control expires. Harper & Row, 471 U.S. at 546. The Act therefore “reflects a balance of competing claims upon the public interest: Creative work is to be encouraged and rewarded, but private motivation must ultimately serve the cause of promoting broad public availability of literature, music, and the other arts.” Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156 (1975). This “balancing act” is reflected throughout the Copyright Act. Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, 598 U.S. 508, 526 (2023) (“Warhol II”). For instance, Section 109 codifies the first sale doctrine, which allows the owner of a physical copy of a work to “sell or otherwise dispose of” it without the copyright owner’s permission. 17 U.S.C. §109(a). Section 108 permits libraries and archives to make a small number of copies of a protected work for certain limited purposes, including preservation and replacement. 17 U.S.C. §108. The Act also allows for certain “fair” uses of copyrighted works. Section 107 provides: “[n]otwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work…for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright.” 17 U.S.C. §107 (emphasis added). The statute sets out four non-exclusive factors for courts to consider in determining whether a particular use is “fair”: (1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and (4) the effect of the use upon the potential market for or value of the copyrighted work. Id. The fair use defense thus excuses what might otherwise be considered infringing behavior, allowing courts to “avoid rigid application” of the Copyright Act when it would “stifle the very creativity” the Act is meant to promote. Stewart v. Abend, 495 U.S. 207, 236 (1990). Given the diverse array of copyrightable material, fair use is a “flexible” concept, whose application varies depending on the context. Google LLC v. Oracle America, Inc., 593 U.S. 1, 20 (2021); Harper & Row, 471 U.S. at 560 (“The factors enumerated in [Section 107] are not meant to be exclusive: Since the doctrine is an equitable rule of reason, no generally applicable definition is possible, and each case raising the question must be decided on its own facts.” (internal quotation marks omitted and alteration adopted)). On appeal, IA does not challenge the district court’s conclusion that Publishers met their affirmative burden of establishing the elements of copyright infringement. Rather, it argues as a defense to the infringement claim that its lending practices constitute fair use under Section 107 of the Copyright Act. Applying the fair use factors as set forth below, we conclude that the challenged practices — IA’s lending of its “own” digital books that “are commercially available for sale or license in any electronic text format,” the Open Libraries Project, and the NEL — are not fair use. No. 1:20-cv-04160-JGK, Dkt. 216 at 2 (internal quotation marks omitted).⁶ I. The purpose and character of the use The first fair use factor is “the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes.” 17 U.S.C. §107(1). In assessing the first factor, courts consider two sub-factors: (i) the extent to which the secondary use is transformative and (ii) whether the secondary use is commercial in nature. Warhol I, 11 F.4th at 37. Though we consider both transformativeness and commerciality, transformativeness is the “central” question. Warhol II, 598 U.S. at 528; Warhol I, 11 F.4th at 37 (“[O]ur assessment of this first factor has focused chiefly on the degree to which the use is ‘transformative.’”); Fox News Network, LLC v. TVEyes, Inc., 883 F.3d 169, 176 (2d Cir. 2018) (describing transformativeness as the “primary inquiry”). A. Transformativeness The first fair use factor focuses primarily on the extent to which the secondary use is transformative; that is, whether the new work merely supplants the original, “or instead adds something new, with a further purpose or different character, altering the [original] with new expression, meaning, or message.” Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579 (1994). The use of a work to achieve “a purpose that is the same as, or highly similar to, that of the original” is more likely to substitute or supplant the original work, and less likely to be considered transformative. Warhol II, 598 U.S. at 528. Paradigmatic examples of transformative uses are those listed in the preamble to Section 107: criticism, comment, news reporting, teaching, scholarship, and research. 17 U.S.C. §107. But transformative uses are not limited to those examples. Parodies of an original work are “routinely held transformative.” Warhol I, 11 F.4th at 37; Campbell, 510 U.S. at 583. Likewise, a secondary use may be transformative if it “improve[s] the efficiency of delivering [original] content,” see TVEyes, 883 F.3d at 177-78 (service that allowed users to view segments of TV programs responsive to their interests “at least somewhat transformative” because it improved the efficiency of delivering content), or expands the utility of the original, see Google Books, 804 F.3d at 214 (scanning books to create a full-text searchable database was transformative because it expanded the utility of the originals). Whether the use finds its approval in the preamble to Section 107 or in our decisions, each of these examples contemplate the use of a work to “communicate[] something new and different from the original or expand[] its utility, thus serving copyright’s overall objective of contributing to public knowledge.” Google Books, 804 F.3d at 214. However, “[n]ot every instance will be clear cut.” Warhol II, 598 U.S. at 528. Whether a secondary use has a different purpose or character from the original “is a matter of degree.” Id. Though “[m]ost copying has some further purpose” and “[m]any secondary works add something new,” that alone “does not render such uses fair.” Id. The word “transformative,” then, “cannot be taken too literally.” Google Books, 804 F.3d at 214. While the right of copyright owners to prepare derivative works is “subject to” fair use, an “overbroad concept of transformative use…that includes any further purpose, or any different character, would [improperly] narrow the copyright owner’s exclusive right to create derivative works.” Warhol II, 598 U.S. at 529 (emphases added). To preserve that right, the degree of transformation must “go beyond that required to qualify as a derivative.” Id. IA argues that its Free Digital Library is transformative because it uses technology “to make lending more convenient and efficient” and “deliver[s] the work only to one already entitled to view it — the one person borrowing the book at a time.” Appellant’s Br. at 30-32 (citing Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417, 488-50 (1984)); TVEyes, 883 F.3d at 178; Capitol Records, LLC v. ReDigi Inc., 910 F.3d 649, 661 (2d Cir. 2018)). Additionally, IA asserts that its Free Digital Library “enables uses not possible with print books and physical borrowing,” such as allowing “authors writing online articles [to] link directly to” a digital book in IA’s library. Id. at 32. Thus, IA concludes, each digital book “serves a new and different function from the original work and is not a substitute for it.” Id. (quoting Authors Guild, Inc. v. HathiTrust, 755 F.3d 87, 96 (2d Cir. 2014)). Publishers, meanwhile, argue that IA’s Free Digital library “does nothing ‘more than repackage or republish’ the Works.” Appellees’ Br. at 26 (quoting HathiTrust, 755 F.3d at 96). They assert that IA does not copy the Works to provide criticism, commentary, or information, but to serve the same purpose as the originals. Id. In Publishers’ view, to hold IA’s use transformative would “destroy the value of [their] exclusive right to prepare derivative works,” including the right to publish their authors’ works as eBooks. Id. at 27. Nor does IA’s adherence to CDL render its use transformative: “[W]ithout a different purpose from the original,” Publishers argue, “IA’s ebooks are not transformative.” Id. We conclude that IA’s use of the Works is not transformative. IA creates digital copies of the Works and distributes those copies to its users in full, for free. Its digital copies do not provide criticism, commentary, or information about the originals. Nor do they “add[] something new, with a further purpose or different character, altering the [originals] with new expression, meaning or message.” Campbell, 510 U.S. at 579. Instead, IA’s digital books serve the same exact purpose as the originals: making authors’ works available to read. IA’s Free Digital Library is meant to — and does — substitute for the original Works. Warhol II, 598 U.S. at 528 (“The use of an original work to achieve a purpose that is the same as, or highly similar to, that of the original work is more likely to substitute for, or supplant, the work.” (cleaned up)). As we have said, “[w]hen all or a substantial portion of text that contains protectable expression is included in another work, solely to convey the original text to the reader without adding any comment or criticism, the second work may be said to have supplanted the original because a reader of the second work has little reason to buy a copy of the original.” Ringgold v. Black Entm’t TV, Inc., 126 F.3d 70, 79 (2d Cir. 1997). Though “[n]ot every instance will be clear cut,” this one is. Warhol II, 598 U.S. at 528. True, there is some “change” involved in the conversion of print books to digital copies. See Infinity Broadcast Corp. v. Kirkwood, 150 F.3d 104, 108 n.2 (2d Cir. 1998) (“[A] change in format…is not technically a transformation.”). But the degree of change does not “go beyond that required to qualify as derivative.” Warhol II, 598 U.S. at 529. Unlike transformative works, derivative works “ordinarily are those that re-present the protected aspects of the original work, i.e., its expressive content, converted into an altered form.” Google Books, 804 F.3d at 225. To be transformative, a use must do “something more than repackage or republish the original copyrighted work.” Authors Guild, Inc. v. HathiTrust, 755 F.3d 87, 96 (2d Cir. 2014); see also TVEyes, 883 F.3d at 177 (“[A] use of copyrighted material that merely repackages or republishes the original is unlikely to be deemed a fair use.” (internal quotation marks omitted)). Changing the medium of a work is a derivative use rather than a transformative one. Warhol II, 598 U.S. at 529. In fact, we have characterized this exact use — “the recasting of a novel as an e-book” — as a “paradigmatic” example of a derivative work. Google Books, 804 F.3d at 215. Digitizing physical copies of written work is not transformative, because the act “merely transforms the material object embodying the intangible article that is the copyrighted original work.” American Geophysical Union v. Texaco Inc., 60 F.3d 913, 923 (2d Cir. 1994). To construe IA’s use of the Works as transformative would significantly narrow — if not entirely eviscerate — copyright owners’ exclusive right to prepare (or not prepare) derivative works. See Warhol II, 598 U.S. at 528-29; 17 U.S.C. §106(2). The Supreme Court has cautioned against such an “overbroad concept of transformative use.” Warhol II, 598 U.S. at 529. In Google Books, we identified a fair use in Google’s “snippet view” search function, which allowed users to search the text and view snippets of it online, 804 F.3d at 217-218, but we noted that had the Plaintiffs’ claim been “based on Google’s converting their books into a digitized form and making that digitized version accessible to the public, their claim would [have been] strong,” id. at 225. As the district court noted here, that is “precisely what the Publishers allege in this case.” Hachette, 664 F. Supp. 3d at 381. IA’s arguments to the contrary are unpersuasive. IA argues that the format change alone is not what makes its use transformative. Rather, its use of the Works is transformative because of “the sort of lending the scanning enables,” namely, (1) making lending more efficient, and (2) enabling other uses not possible with print books and physical borrowing. Appellant’s Br. at 33. Neither of these justifications render IA’s use of the Works transformative because the underlying purpose of making the Works available in a derivative format is still the same. IA argues that its use of the Works is transformative because it “make[s] lending more convenient and efficient” and “uses technology to deliver the work only to one already entitled to view it — the one person borrowing the book at a time.” Id. at 31-32. IA derives this argument from three cases: Sony, TVEyes, and ReDigi. In Sony, the defendant manufactured and sold home video tape recorders that allowed users to record TV programs broadcast at set times over the airwaves so they could watch them later. 464 U.S. at 422-23. The Supreme Court, in a decision that predated our use of the word “transformative” as a term of art, concluded that “time-shifting for private home use…[is] a noncommercial, nonprofit activity” that “merely enables a viewer to see such a work which [they] had [already] been invited to witness in its entirety free of charge.” Id. at 449. For those reasons, the Court concluded that the first fair use factor favored the defendant. Id. We contextualized Sony within the modern fair use framework in TVEyes. There, the defendant recorded vast quantities of live TV programs, compiled those recordings into a text-searchable database in which they would be held for thirty-two days before being deleted, and allowed its paying subscribers to search for and watch up to ten-minute video clips of recordings that featured their terms of interest. 883 F.3d at 173-74. In assessing whether the use was transformative, we reframed the Supreme Court’s discussion from Sony: While Sony was decided before “transformative” became a term of art, the apparent reasoning was that a secondary use may be a fair use if it utilizes technology to achieve the transformative purpose of improving the efficiency of delivering content without unreasonably encroaching on the commercial entitlements of the rights holder. Id. at 177. Expanding on that reasoning, although we ultimately rejected TVEyes’ fair use claim, we concluded that the use at issue was at least “somewhat transformative” because it “enhanc[ed] efficiency” by enabling users to view programming that discussed topics of interest to them “without having to monitor thirty-two days of programming in order to catch each relevant discussion.” Id. at 177. Later that same year, we addressed the fair use defense of a defendant-website that hosted an online platform for reselling digital music files. ReDigi, 910 F.3d at 652-54. We repeated our characterization of Sony from TVEyes, adding: In Sony, the “apparent reasoning was that a secondary use may be a fair use if it utilizes technology to achieve the transformative purpose of improving the efficiency of delivering content without unreasonably encroaching on the commercial entitlements of the rights holder” because the improved delivery was to one entitled to receive the content. Id. at 661 (quoting TVEyes, 883 F.3d at 177) (emphasis added). Still, we determined that the use in ReDigi was not transformative. Id. The website provided “neither criticism, commentary, nor information about” the original music files, nor did it “deliver the content in more convenient and usable form to one who has acquired an entitlement to receive the content.” Id. Instead, it effectively “provide[d] a market for the resale of digital music files, which…compete[d] with sales of the same recorded music by the rights holder.” Id. For that reason, the first fair use factor favored the plaintiffs. Id. The “efficiencies” identified in Sony and TVEyes are distinct from the purported efficiencies offered by IA’s Free Digital Library. Sony was decided long before modern technology made it possible for one to view virtually any content at any time. Put in context, the “time-shifting” permitted by the defendant’s tape recorders in Sony was a unique efficiency not widely available at the time, and certainly not offered by the plaintiff-television producer. 464 U.S. at 449. The defendant in TVEyes likewise offered a uniquely efficient service: allowing users to watch short clips of television programs featuring their selected search terms. 883 F.3d at 173-74. Here, by contrast, IA’s Free Digital Library offers few efficiencies beyond those already offered by Publishers’ own eBooks. IA argues that its use is more efficient because it “replace[s] the burdens of physical transportation with the benefits of digital technology,” but this ignores the fact that IA’s digital books compete directly with Publishers’ eBooks — works derivative of the original print books. Appellant’s Br. at 32. Understood this way, IA’s Free Digital Library does not “improv[e] the efficiency of delivering content” without unreasonably encroaching on the rights of the copyright holder; it offers the same efficiencies as Publishers’ derivative works while greatly impinging on their exclusive right to prepare those works. TVEyes, 883 F.3d at 177; ReDigi, 910 F.3d at 661. In Sony, “timeshifting merely enable[d] a viewer to see such a work which [the viewer] had been invited to witness in its entirety free of charge” by the broadcasters, Sony, 464 U.S. at 449, and therefore did not “unreasonably encroach[] on the commercial entitlements of the rights holder,” TVEyes, 883 F.3d at 177 (characterizing Sony). The Publishers in this case never “invited” readers to read their books for free from an unlicensed digital library. Sony, 464 U.S. at 449. Nor does IA’s asserted adherence to CDL render its use transformative. IA maintains that it delivers each Work “only to one already entitled to view [it]” — i.e., the one person who would be entitled to check out the physical copy of each Work. Appellant’s Br. at 32. But this characterization confuses IA’s practices with traditional library lending of print books. IA does not perform the traditional functions of a library; it prepares derivatives of Publishers’ Works and delivers those derivatives to its users in full. That Section 108 allows libraries to make a small number of copies for preservation and replacement purposes does not mean that IA can prepare and distribute derivative works en masse and assert that it is simply performing the traditional functions of a library. 17 U.S.C. §108; see also, e.g., ReDigi, 910 F.3d at 658 (“ We are not free to disregard the terms of the statute merely because the entity performing an unauthorized reproduction makes efforts to nullify its consequences by the counterbalancing destruction of the preexisting phonorecords.”). Whether it delivers the copies on a one-to-one owned-to-loaned basis or not, IA’s recasting of the Works as digital books is not transformative. Google Books, 804 F.3d at 215. IA also argues that its use is transformative “because it enables uses not possible with print books and physical borrowing,” such as allowing “authors writing online articles [to] link directly to [IA's digital books].” Appellant’s Br. at 32. We rejected a similar argument in TVEyes. There, the defendant argued that its provision of reproduced content identified through targeted search tools was transformative “because it allow[ed] clients to conduct research and analysis of television content by enabling them to view [video] clips responsive to their research needs[,] [and] [r]esearch…[was] a purpose not shared by users of the original content.” TVEyes, 883 F.3d at 178 n.4. Although we concluded that TVEyes’ use was at least “somewhat transformative,” id. at 178, we rejected the research-based argument, explaining that simply because “a secondary use can facilitate research does not itself support a finding that the secondary use is transformative,” id. at 178 n.4. To deem a use transformative on that basis would extend the concept of transformativeness “beyond recognition.” Id. So too here. That authors of online articles may embed links to IA’s Free Digital Library does not render the Library a significantly transformative secondary use of the Works. In sum, because IA’s Free Digital Library primarily supplants the original Works without adding meaningfully new or different features that avoid unduly impinging on Publishers’ rights to prepare derivative works, its use of the Works is not transformative. TVEyes, 883 F.3d at 177 (“Although transformative use is not absolutely necessary for a finding of fair use, transformative works lie at the heart of the fair use doctrine, and a use of copyrighted material that merely repackages or republishes the original is unlikely to be deemed a fair use.” (cleaned up)). B. Commerciality We next consider “whether [the] use is of a commercial nature or is for nonprofit educational purposes.” 17 U.S.C. §107(1). The commercial nature of a secondary use generally weighs against a finding of fair use, especially where the secondary use is only “modest[ly] transformative.” TVEyes, 883 F.3d at 178. However, we assess commerciality “with caution” since “nearly all of the illustrative uses listed in the preamble to Section 107 are generally conducted for profit in this country.” Warhol I, 11 F.4th at 44 (cleaned up). “The crux of the profit/nonprofit distinction,” then, “is not whether the sole motive of the use is monetary gain[,] but whether the user stands to profit from exploitation of the copyrighted material without paying the customary price.” Harper & Row, 471 U.S. at 562. Importantly, commerciality is only one part of the first factor inquiry and the broader fair use analysis. See, e.g., Sony, 464 U.S. at 448-49 (noting that the commercial or nonprofit character of a work is “not conclusive” but rather a factor to be “weighed [alongside the others] in any fair use decision”). Accordingly, “the mere fact that a use is educational and not for profit does not insulate it from a finding of infringement, any more than the commercial character of a use bars a finding of fairness.” Campbell, 510 U.S. at 584; see also Google Books, 804 F.3d at 219 n.20 (“[T]here is…no reason to presume categorically that a nonprofit educational purpose should qualify as a fair use.”). Here, the district court concluded that IA’s use of the Works is commercial for two principal reasons. Hachette, 664 F. Supp. 3d at 383-84. Though it acknowledged IA’s nonprofit status, it determined that IA still “exploits the Works in Suit without paying the customary price” by (1) soliciting donations on its website and (2) taking a cut of the proceeds when users buy a physical book from BWB using a link embedded on IA’s website. Id. at 383-84. The district court found “largely irrelevant” the fact that IA distributes its digital books for free, instead emphasizing that “[w]hat matters is whether IA profited from copying the Works.” Id. at 384. IA objects to the district court’s “expansive” understanding of commerciality, asserting that it “receives no profit or other private benefit” from its use of the Works. Appellant’s Br. at 21, 23. Publishers argue that IA’s use is commercial in nature because IA reaps some economic advantages from its partnership with BWB and obtains “other non-monetary calculable benefits” from its use of the Works, including advancing its mission of ensuring universal access to all knowledge. Appellees’ Br. at 39. However, Publishers “do not contend that a non-profit…is commercial merely because it solicits donations.” Id. at 38 n.12. We conclude, contrary to the district court, that IA’s use of the Works is not commercial in nature. It is undisputed that IA is a nonprofit entity and that it distributes its digital books for free. Of course, IA must solicit some funds to keep the lights on: its website includes a link to “Donate” to IA, and it has previously received grant funding to support its various activities. App’x 6091. But unlike the defendant in TVEyes, who charged users a fee to use its text-searchable database, IA does not profit directly from its Free Digital Library. 883 F.3d at 175. It offers this service free of charge. See American Society for Testing and Materials v. Public.Resource.Org, Inc., 896 F.3d 437, 449 (D.C. Cir. 2018) (“ASTM I”) (concluding that use was not commercial where there was no evidence that nonprofit organization profited from its dissemination of reproduced industry standards that had been incorporated into law). We therefore disagree with the district court that IA’s partnership with BWB renders its use commercial. When users read a digital book on IA’s web browser platform, a “Purchase at Better World Books” button appears at the top of their screens. This button links to BWB’s website, where users may buy a used print copy of the book they were just reading, or any other book in BWB’s library. BWB compensates IA whenever someone buys a book via the “Purchase at Better World Books” link; although the exact profits IA earns from this arrangement is disputed. Any link between the funds IA receives from its partnership with BWB and its use of the Works is too attenuated for us to characterize the use as commercial on that basis. See Swatch Group Management Services Ltd. v. Bloomberg L.P., 756 F.3d 73, 83 (2d Cir. 2014). In Swatch, Bloomberg posted a recording of a Swatch conference call on its website for access by its paying subscribers. Id. at 78-79. Although the recording — like other Bloomberg content — could be accessed only by paying subscribers, we concluded that “the link between the defendant’s commercial gain and its copying [was] attenuated such that it would be misleading to characterize the use as commercial exploitation.” Id. at 83 (cleaned up). Unlike in TVEyes, where the defendant charged a fee in direct connection with its use of the works, Bloomberg’s subscription fee covered a broad range of services, including access to the recording. Compare id., with TVEyes, 883 F.3d at 175. Because “it would strain credulity to suggest that providing access to [the recording] more than trivially affected the [overall] value of [Bloomberg's] service[s],” we assigned the commercial nature of Bloomberg’s use “little weight.” Swatch, 756 F.3d at 83. IA does not charge a fee for any of its services, let alone the Free Digital Library. While IA may incidentally profit from its use of the Works when account holders check out a digital book via the BookReader platform, click on the “Purchase at Better World Books” link, and purchase a book from BWB, the link between this commercial gain and IA’s use of the Works to create the Free Digital Library is “attenuated[,] such that it would be misleading to characterize the use as commercial exploitation.” Id. (internal quotation marks omitted). Although IA exploits the Works in Suit without paying the customary price, it does not profit directly from that exploitation. See Harper & Row, 471 U.S. at 562. We likewise reject the proposition that IA’s solicitation of donations renders its use of the Works commercial. IA does not solicit donations specifically in connection with its digital book lending services — nearly every page on IA’s website contains a link to “Donate” to IA. App’x 6091. Thus, as with its partnership with BWB, any link between the funds IA receives from donations and its use of the Works is too attenuated to render the use commercial. Swatch, 756 F.3d at 83. To hold otherwise would greatly restrain the ability of nonprofits to seek donations while making fair use of copyrighted works. See ASTM I, 896 F.3d at 449 (rejecting the argument that because free distribution of copyrighted industry standards enhanced a nonprofit organization’s fundraising appeal, the use was commercial). Finally, to the extent Publishers argue that IA’s use is commercial because it derives “other non-monetary calculable benefits” from the Free Digital Library, we disagree. Appellees’ Br. at 39. It is true that the “profit/non-profit distinction is context specific, not dollar dominated,” such that the “absence of a dollars and cents profit” will not necessarily exclude a finding of commerciality. Weissmann v. Freeman, 868 F.2d 1313, 1324 (2d Cir. 1989). Indeed, in Weissmann we refused to weigh the first fair use factor against the plaintiff where the defendant appropriated her unpublished research article and presented it as his own. Id. at 1316. Despite the apparent lack of economic profit, the defendant “stood to gain recognition among his peers in the profession…without paying the usual price that accompanies scientific research and writing, that is to say, by the sweat of his brow.” Id. at 1324. In this context, however, IA does not stand to gain the direct reputational benefits that accompany presenting another’s work as one’s own. Rather, it obtains only those nonmonetary benefits that attend most other legitimate, secondary uses, including advancing its mission and bolstering its reputation. Characterizing these general benefits as commercial profits would render commercial the activities of virtually any nonprofit organization that bolsters its reputation through its own nonprofit activities. IA does not profit directly from its exploitation of the Works in Suit. For that reason, its use of the Works is non-commercial in nature. But the nonprofit character of IA’s use is not conclusive, Sony, 464 U.S. at 448, and transformativeness remains the “central” focus of the first factor, Warhol II, 598 U.S. at 542. Thus, because IA’s use of the Works is not transformative, the first fair use factor favors Publishers. See TVEyes, 883 F.3d at 177-78. II. The nature of the copyrighted works The second fair use factor is “the nature of the copyrighted work.” 17 U.S.C. §107(2). In assessing the second factor, courts consider (i) whether the work is expressive or creative, with “greater leeway being allowed to a claim of fair use where the work is factual or informational”; and (ii) “whether the work is published or unpublished, with the scope of fair use involving unpublished works being considerably narrower.” Blanch v. Koons, 467 F.3d 244, 256 (2d Cir. 2006). This factor recognizes that “some works are closer to the core of intended copyright protection than others.” Campbell, 510 U.S. at 586. However, it “rarely play[s] a significant role in the determination of a fair use dispute.” Google Books, 804 F.3d at 220; see also HathiTrust, 755 F.3d at 98 (“The second fair-use factor…is not dispositive.”). IA argues that this factor weighs neutrally because the Works, though published, include both fiction and nonfiction books. Although we have previously stated that the second factor tends to favor fair use when the copied works are factual, rather than fictional, that general statement is an oversimplification of the Works in this case. While the Works in Suit include both fiction and nonfiction books, even the nonfiction books contain original expression “close[] to the core of intended copyright protection.” Campbell, 510 U.S. at 586. The Copyright Act “does not protect facts or ideas set forth in a work, [but] it does protect that author’s manner of expressing those facts and ideas.” Google Books, 804 F.3d at 220 (emphasis added). Wind Power for Dummies may contain factual information regarding the mechanics of wind power, but the author’s compilation of those facts into a book suitable for the average “dummy” is an original expression that the Copyright Act protects. The nonfiction Works here are therefore unlike the copyrighted materials in Swatch, one of the few cases in which the second fair use factor bore some significance. There, as noted above, Bloomberg obtained a sound recording and transcript of a private conference call between Swatch executives and a group of securities analysts and posted both online for access by its paying subscribers. Swatch, 756 F.3d at 78-79. Swatch sued Bloomberg, obtaining a copyright to cover the statements made by Swatch executives during the meeting. Id. at 79. In concluding that the second factor favored fair use, we remarked on the “thinness” of Swatch’s copyright: the copyrighted statements were of a “manifestly factual character,” intended to “convey financial information about the company,” not to showcase any original expression. Id. at 89. Here, while the nonfiction Works undoubtedly convey factual information and ideas, they also represent the authors’ original expressions of those facts and ideas — and those “subjective descriptions and portraits” reflect “the author’s individualized expression.” Harper & Rowe, 471 U.S. at 563. Thus, because the Works in Suit are “of the type that the copyright laws value and seek to protect,” the second fair use factor favors Publishers. HathiTrust, 755 F.3d at 98. III. The amount and substantiality of the use The third fair use factor is “the amount and substantiality of the portion used in relation to the copyrighted work as a whole.” 17 U.S.C. §107(3). Generally, a finding of fair use is more likely when “small amounts, or less important passages [of the work] are copied than when the copying is extensive, or encompasses the most important parts of the original.” Google Books, 804 F.3d at 221. The third and fourth factors are interrelated, as the greater the amount and substantiality of the use, “the greater the likelihood that the secondary work might serve as an effectively competing substitute for the original.” Id. Still, there is no bright line rule separating permissible from impermissible copying. Warhol I, 11 F.4th at 46. Sometimes it is necessary to copy the entirety of a work in order to achieve a legitimate, transformative secondary purpose. See, e.g., Google Books, 804 F.3d at 221-22 (copying entirety of books necessary to create transformative search function and snippet view); HathiTrust, 755 F.3d at 98-99 (copying entirety of books necessary to create transformative search function). Other times, “taking even a small percentage of the original work [is] unfair use.” Rogers v. Koons, 960 F.2d 301, 311 (2d Cir. 1992). The relevant consideration, then, is not the amount of copyrighted material used by the copier, but “the amount of copyrighted material made available to the public.” TVEyes, 883 F.3d at 179. Here, it is undisputed that IA copies the Works in their entirety and distributes those copies to the public in full. Still, IA argues that the third factor weighs neutrally because “copying the entire work is necessary for Controlled Digital Lending.” Appellant’s Br. at 43 (cleaned up). As IA itself recognizes, this argument hinges entirely on its assumption that its use of the Works is transformative.⁷ But IA does not scan the Works in their entirety to achieve a transformative secondary purpose; it scans the Works in their entirety to substitute IA’s digital books for Publishers’ print books and eBooks. IA’s use is therefore unlike the copying that took place in HathiTrust and Google Books. In those cases, the defendants scanned copyrighted books to create, among other things, searchable databases that allowed users to view snippets of text pertaining to their search terms, or to learn on which pages and with what frequency their search terms appeared in any given book. HathiTrust, 755 F.3d at 91; Google Books, 804 F.3d at 208-10. Though the defendants copied the books in their entirety, doing so was necessary to achieve a transformative, secondary purpose — the searchable databases. HathiTrust, 755 F.3d at 98-99; Google Books, 804 F.3d at 221-23. And critically, while the defendants “ma[de] an unauthorized digital copy of the entire book, [they did] not reveal that digital copy to the public.” Google Books, 804 F.3d at 221 (second emphasis added). Here, IA makes unauthorized digital copies of the Works and reveals those copies to the public in their entirety. The “amount and substantiality” of the copying is not necessary to achieve a transformative secondary purpose, but serves to substitute Publishers’ books. 17 U.S.C. §107(3). For that reason, the third fair use factor favors Publishers. See ReDigi, 910 F.3d at 662 (third factor favored plaintiffs where the defendant made available “identical copies of the whole of Plaintiffs’ copyrighted sound recordings.”); TVEyes, 883 F.3d at 179 (third factor favored plaintiff where the defendant made available “virtually the entirety of the [plaintiff's] programming that TVEyes users want[ed] to see and hear.”). IV. The effect of the use on the potential market for or value of the works The fourth and final fair use factor is “the effect of the use upon the potential market for or value of the copyrighted work.” 17 U.S.C. §107(4). This factor “focuses on whether the copy brings to the marketplace a competing substitute for the original, or its derivative, so as to deprive the rights holder of significant revenues because of the likelihood that potential purchasers may opt to acquire the copy [rather than] the original.” Google Books, 804 F.3d at 223. Analysis of this factor “requires us to balance the benefit the public will derive if the use is permitted” against “the personal gain the copyright owner will receive if the use is denied.” Warhol I, 11 F.4th at 48. The fourth factor is “undoubtedly the single most important element of fair use.” Harper & Row, 471 U.S. at 566. In assessing the fourth factor, we consider “not only the market harm caused by the particular actions of the alleged infringer, but also the market harm that would result from unrestricted and widespread conduct of the same sort.” TVEyes, 883 F.3d at 179 (cleaned up); Campbell, 510 U.S. at 590. We ask not “whether the second work would damage the market for the first (by, for example, devaluing it through parody or criticism), but whether it usurps the market for the first by offering a competing substitute.” Warhol I, 11 F.4th at 48. This inquiry encompasses both the market for the original work and the market for any derivative works the rightsholder might develop. Campbell, 510 U.S. at 590. The first and fourth factors are closely related, as the more copying is done to achieve a purpose that is the same as or substantially similar to the original, the more likely it is that the copy will serve as a “satisfactory substitute for the original.” Google Books, 804 F.3d at 223. Ultimately, the “burden of proving that the secondary use does not compete in the relevant market is…borne by the party asserting the defense.” Warhol I, 11 F.4th at 49. Before discussing the parties’ arguments, we must identify the relevant market in which IA’s Free Digital Library allegedly competes. Publishers focus on harm to their markets for eBooks, including consumer eBooks and licensed eBooks. IA, meanwhile, argues that we should consider harm (or lack of harm) to Publishers’ eBook and print book markets. In this case, the relevant market for purposes of analyzing the fourth fair use factor is the market for the Works in general, without regard to format. The Copyright Act protects authors’ works in whatever format they are produced. See generally 17 U.S.C. §106 (granting copyright owners the exclusive right to reproduce their works in numerous formats). Even if we consider the print editions to be the “original” forms of the Works in this case — and we recognize that the concept of a singular “original” form may be difficult to grasp in the digital age — the Act recognizes that presentation of the same content in a different medium is a derivative work subject to the same legal protections as the purported original. 17 U.S.C. §106(2). Here, Publishers obtained from authors the exclusive right to publish their Works in numerous formats, including print and eBooks, and it is this exclusive right that IA is alleged to have violated via its Free Digital Library. For that reason, the relevant harm — or lack thereof — is to Publishers’ markets for the Works in any format. A. IA does not disprove market harm IA argues that its Free Digital Library causes no harm to Publishers’ markets for the Works because IA’s lending “offers a distinct service” from Publishers’ eBooks, not a substitute; and “all available evidence shows that IA’s lending caused no harm at all” to Publishers’ markets for the Works. Appellant’s Br. at 45. We address each point in turn. IA’s first argument is a simple one: because its Free Digital Library is a transformative use of the Works, it cannot serve as a substitute for the originals. This argument, like numerous other arguments raised by IA in this appeal, is premised on its assumption that its use of the Works is transformative. But there is nothing transformative about IA’s use of the Works or its adherence to CDL — its digital copies serve the same purpose as the originals. And the more copying is done to achieve a purpose that is the same as the original, the more likely it is that the copy will compete in the market as a “satisfactory substitute for the original.” Google Books, 804 F.3d at 223. Google Books and HathiTrust underscore this point. In HathiTrust, the fourth factor supported a finding of fair use because the ability to search the text of a book to determine whether it includes select words did not “serve as a substitute for the books [themselves].” 755 F.3d at 100. Likewise, in Google Books, the ability to view small fragments of books pertaining to particular search terms did not “provide a significant substitute for the purchase of the [actual] book[s].” 804 F.3d at 224-25. Critical to our conclusions in those cases was the interrelatedness of the first and fourth factors: because the copying was “done to achieve a purpose that differ[ed] from the purpose of the original[s],” it was unlikely that the secondary uses would “serve as a satisfactory substitute for the original[s].” Id. at 223. Here, not only is IA’s Free Digital Library likely to serve as a substitute for the originals, the undisputed evidence suggests it is intended to achieve that exact result. IA copies the Works in full and makes those copies available to the public in their entirety. It does not do this to achieve a transformative secondary purpose, but to supplant the originals. IA itself advertises its digital books as a free alternative to Publishers’ print and eBooks. See, e.g., App’x 6099 (“[T]he Open Libraries Project ensures [libraries] will not have to buy the same content over and over, simply because of a change in format.” (internal quotation marks omitted)); id. at 6100 (marketing the Free Digital Library as a way for libraries to “get free ebooks”); id. at 6099 (“You Don’t Have to Buy it Again!”). IA offers effectively the same product as Publishers — full copies of the Works — but at no cost to consumers or libraries. At least in this context, it is difficult to compete with free. IA’s non-transformative use of the Works is a strong indicator that its digital books “usurp[]the market for the [originals] by offering a competing substitute.” Warhol I, 11 F.4th at 48 (emphasis omitted). Still, even where a secondary use serves the same purpose as the originals, the defendant may present evidence of a lack of market harm to support their defense of fair use. IA presents data that purportedly show a lack of harm to Publishers’ print and eBook markets, and faults Publishers for failing to provide empirical data of their own. IA’s expert, Dr. Rasmus Jørgensen, examined OverDrive checkouts of the Works before, during, and after the National Emergency Library (IA’s COVID-era program pursuant to which it lifted its one-to-one owned-to-loaned ratio, allowing each digital book to be checked out by up to 10,000 users at a time without regard to the corresponding number of physical books in storage or in partner libraries’ possession) to assess potential harm to Publishers’ eBook licensing market. If IA’s lending were indeed a substitute for Publishers’ library eBook licenses, he theorized, then the shutdown of the NEL and reinstitution of IA’s lending controls should correspond to an increase in demand for the Works on OverDrive (the commercial service used by many libraries who license eBooks). But Dr. Jørgensen found the opposite: OverDrive checkouts of the Works decreased following the shutdown of the NEL in June 2020. From this, IA concludes that its lending “has no effect on demand for borrowing on OverDrive” and, therefore, there is “no reason to imagine, much less assume, that digital lending affects Publishers’ ebook license revenue at all.” Appellant’s Br. at 51. Dr. Jørgensen used the same methodology to analyze the effect of IA’s lending on Hachette’s⁸ monthly print and eBook sales of the Works. Once again, he theorized that if IA’s lending were a substitute for Hachette’s consumer print and eBooks, then the closure of the NEL should correspond to an increase in Hachette’s sales of the Works in those mediums. But it did not: at the same time IA reinstituted its lending controls in June 2020, Hachette’s print and eBook sales decreased by 21 percent and 29 percent, respectively. IA also submits the expert report of Dr. Imke Reimers, who examined the effect of IA’s lending on Amazon sales rankings for print copies of the Works. Dr. Reimers analyzed whether Amazon sales rankings changed when IA (1) first added the Work to its Free Digital Library, (2) launched the NEL, or (3) removed the Work from its Free Digital Library in response to this lawsuit. She found “no statistically significant evidence” that either inclusion in IA’s library or increased lending through the NEL harmed print sales rankings on Amazon, and that removal of the Works from IA’s Free Digital Library actually correlated with a decrease in sales rankings of the Works on Amazon. App’x 4934. From this and Dr. Jørgensen’s report, IA concludes that its Free Digital Library has no effect on Publishers’ markets for print and eBooks. This evidence does not satisfy IA’s “burden of proving that the secondary use does not compete in the relevant market[s].” Warhol I, 11 F.4th at 49. Dr. Jørgensen’s report shows, at best, a weak correlation between the closure of the NEL and a decline in Publishers’ checkouts and sales. As Publishers’ expert Dr. Jeffrey Prince points out, Dr. Jørgensen “does not even have a[n] [empirical] model,” let alone a model that accounts for the effect of various other causal factors on Publishers’ sales. App’x 5476. Instead, Dr. Jørgensen simply analyzes “two points in time during a period where major macroeconomic events were occurring,” id. at 5382, and “say[s] any difference that you see is because of the takedown of the NEL,” id. at 5530. But “[j]ust because one event follows another doesn’t mean it was caused by the prior event.” Id. at 5383. There were a broad range of factors at play in June 2020 for which Dr. Jørgensen’s report fails to account, including the “COVID-related surge in ebook sales” which “slowed down by summer 2020″ as public spaces began to reopen. Id. at 5469. Dr. Jørgensen’s own graph reveals that checkouts spiked for OverDrive and IA between March 2020 and June 2020, followed by a steady decline for both as checkouts reverted to pre-pandemic levels. Id. at 4850. In short, Dr. Jørgensen’s report does not indicate, let alone prove, the effect of IA’s Free Digital Library on Publishers’ markets for the Works. Dr. Reimers’ report suffers from deeper infirmities. First, her report is of limited relevance as it analyzes Amazon sales rankings for print editions of the Works in Suit, not print revenues, eBook Amazon sales rankings, or eBook revenues. As Dr. Prince explains, Amazon sales rankings are not necessarily a proxy for revenues, and the sales rankings of print books tell us nothing about the impact of IA’s lending practices on sales and checkouts of eBooks. Id. at 5397-98. Moreover, the Amazon algorithm that determines the rankings is “vague” and reportedly “susceptible to manipulation.” Id. at 5474. This makes it difficult to discern whether and to what extent those rankings correspond to sales revenues. For that reason, the conclusions drawn by IA from Dr. Reimers’ report regarding the impact of their lending services on Publishers’ sales are ill supported. Although they do not provide empirical data of their own, Publishers assert that they (1) have suffered market harm due to lost eBook licensing fees and (2) will suffer market harm in the future if IA’s practices were to become widespread. IA argues that Publishers cannot rely on the “common-sense inference” of market harm without data to back that up, citing American Society for Testing & Materials v. Public.Resource.Org, Inc., 82 F.4th 1262 (D.C. Cir. 2023) (“ASTM II”). Appellant’s Br. at 55-56. We agree with Publishers’ assessment of market harm. “It is indisputable that, as a general matter, a copyright holder is entitled to demand a royalty for licensing others to use its copyrighted work, and that the impact on potential licensing revenues is a proper subject for consideration in assessing the fourth factor.” Bill Graham Archives v. Dorling Kindersley Ltd., 448 F.3d 605, 614 (2d Cir. 2006). While “only an impact on potential licensing revenues for traditional, reasonable, or likely to be developed markets should be legally cognizable” when evaluating market harm, here, Publishers’ library eBook licensing market is both reasonable and developed. TVEyes, 883 F.3d at 180 (cleaned up). And it is undisputed that IA appropriated the Works “without payment of [the] customary licensing fee.”⁹ Ringgold, 126 F.3d at 81. IA has thus “usurped a market that properly belongs to the copyright holder.”¹⁰ TVEyes, 883 F.3d at 180 (cleaned up). We are likewise convinced that “unrestricted and widespread conduct of the sort engaged in by [IA] would result in a substantially adverse impact on the potential market for [the Works in Suit].” Campbell, 510 U.S. at 590 (cleaned up). IA’s Free Digital Library serves as a satisfactory substitute for the original Works. Were we to approve IA’s use of the Works, there would be little reason for consumers or libraries to pay Publishers for content they could access for free on IA’s website. See Warhol I, 11 F.4th at 50. Though Publishers have not provided empirical data to support this observation, we routinely rely on such logical inferences where appropriate in assessing the fourth fair use factor. See, e.g., Google Books, 804 F.3d at 223-25 (relying on logical inferences rather than empirical data to conclude that the secondary use did not serve as a competing substitute for the originals); ReDigi, 910 F.3d at 662-63 (relying on logical inferences rather than empirical data to conclude that the secondary use served as a competing substitute for the originals); Warhol I, 11 F.4th at 50 (relying on “self-evident” harm that would result if the defendant’s use were to become widespread). Thus, we conclude it is “self-evident” that if IA’s use were to become widespread, it would adversely affect Publishers’ markets for the Works in Suit. Warhol I, 11 F.4th at 50. ASTM II does not convince us otherwise. That D.C. Circuit case involved various private organizations that promulgate, copyright, and sell copies of standards establishing best practices for their respective industries, which are often incorporated by reference into federal agency rules and state and local regulations. ASTM II, 82 F.4th at 1265. Three of those standard-setting organizations sued nonprofit Public.Resource.Org (“PR”), alleging that it infringed their copyrights in numerous standards that had been incorporated into government regulations by posting those standards on its website for users to access for free. Id. at 1266. The district court determined that PR’s use of the standards was a fair use, and the organizations appealed. Id. at 1265. The D.C. Circuit affirmed, concluding that the first three fair use factors strongly favored PR. Id. at 1267. With regard to market harm, the organizations pressed the “common-sense” argument that, if users can download an identical copy of a standard for free, few will pay to buy the standard. Id. at 1271. The court rejected this argument for two reasons. First, PR published only versions of the regulations that had been incorporated into government regulations. Id. Because the organizations regularly update their standards, and government agencies are slow to incorporate those new standards, if builders, engineers, or other consumers wanted the most up-to-date version of a regulation, they would have to purchase the organization’s copy. Id. Thus, there was reason to doubt the “common-sense inference” that PR’s activities would usurp the market for the organizations’ standards. Id. Second, the organizations did not provide any empirical evidence of market harm, while PR introduced data indicating a lack of such harm. Id. The Court explained: Public Resource has been posting incorporated standards for fifteen years. Yet the plaintiffs have been unable to produce any economic analysis showing that Public Resource’s activity has harmed any relevant market for their standards. To the contrary, ASTM’s sales have increased over that time; NFPA’s sales have decreased in recent years but are cyclical with publications; and ASHRAE has not pointed to any evidence of its harm. The plaintiffs’ primary evidence of harm is an expert report opining that Public Resource’s activities could put the plaintiffs’ revenues at risk. Yet although the report qualitatively describes harms the plaintiffs could suffer, it makes no serious attempt to quantify past or future harms. Like the district court, we find it telling that the plaintiffs do not provide any quantifiable evidence, and instead rely on conclusory assertions and speculation long after Public Resource first began posting the standards. Id. (cleaned up). The Court thereafter concluded that the “fourth fair-use factor does not significantly tip the balance one way or the other.” Id. at 1272. IA urges us to adopt the same analysis of market harm. It argues that Publishers’ lack of data is similarly “telling,” and that we, like the D.C. Circuit, should reject their common-sense claim that if users can download an identical copy of a Work for free, few will pay to buy or lease Publishers’ editions. Appellant’s Br. at 56. To the extent IA faults Publishers’ lack of empirical data, it forgets the burden of proof. Recall the broader context: Publishers have already established a prima facie case of copyright infringement. The only issue in this appeal is whether IA’s Free Digital Library constitutes a fair use of the Works. “Fair use is an affirmative defense; as such, the ultimate burden of proving that the secondary use does not compete in the relevant market is appropriately borne by the party asserting the defense: the secondary user.” Warhol I, 11 F.4th at 49. While the rightsholder may bear some initial burden of identifying relevant markets, “we have never held that the rightsholder bears the burden of showing actual market harm.” Id. Publishers need not present empirical data of their own in connection with IA’s asserted affirmative defense. Additionally, the nature of the works at issue in ASTM II was of a manifestly different character than the Works at issue here. The organizations in ASTM II updated their standards frequently, such that their production of new copyrighted standards typically outpaced governments’ incorporation of those standards into law. ASTM II, 82 F.4th at 1271. Given that PR posted only standards that had been incorporated into law, its use did not usurp the organizations’ market for the most up-to-date version of the standards. Id. The interrelatedness of the first and fourth factors was apparently critical to the D.C. Circuit’s conclusion: because PR’s copies did not replace the originals — i.e., the most up-to-date versions of the standards — it was unlikely that those copies would bring to the marketplace a competing substitute for the originals. Id. at 1271, 1268 (“[PR] publishes only what the law is, not what industry groups may regard as current best practices.”). This lack of usurpation made the organizations’ failure to quantify past or future harms “telling.”¹¹ Id. Here, by contrast, IA’s use of the Works, which remain unchanged through time, is not transformative. Because IA’s Free Digital Library functions as a replacement for the originals, it is reasonable and logical to conclude not only that IA’s digital books currently function as a competing substitute for Publishers’ licensed editions of the Works, but also that, if IA’s practices were to become “unrestricted and widespread,” TVEyes, 883 F.3d at 179, it would decimate Publishers’ markets for the Works in Suit across formats. B. Any public benefits of IA’s Free Digital Library do not outweigh the harm to Publishers’ markets Although IA cannot disprove market harm, we still “balance the benefit the public will derive if the use is permitted [against] the personal gain the copyright owner will receive if the use is denied.” Warhol I, 11 F.4th at 48. IA argues that its Free Digital Library provides “significant public benefits,” including expanding access to free literary materials. Appellant’s Br. at 59. It asserts that prohibiting its CDL practices would cause harm “to those who have difficulty accessing physical libraries and to researchers and authors who use controlled digital lending in creating new works.” Id. We conclude that both Publishers and the public will benefit if IA’s use is denied. To be sure, expanding access to knowledge would, in a general sense, benefit the public. But “[a]ny copyright infringer may claim to benefit the public by increasing public access to the copyrighted work.” Harper & Row, 471 U.S. at 569. That does not alone render the infringement lawful. Indeed, the Copyright Act and its empowering constitutional authority reflect a considered judgment that “the Progress of Science and useful Arts” is best promoted by laws that protect authors’ original works and permit authors to set the terms of engagement, at least for a limited time. See Sony, 464 U.S. at 429. Doing so benefits the public “by providing rewards for authorship.” Google Books, 804 F.3d at 212. This monopolistic power is a feature, not a bug, of the Copyright Act. Within the framework of the Copyright Act, IA’s argument regarding the public interest is shortsighted. True, libraries and consumers may reap some short-term benefits from access to free digital books, but what are the long-term consequences? If authors and creators knew that their original works could be copied and disseminated for free, there would be little motivation to produce new works. And a dearth of creative activity would undoubtedly negatively impact the public. It is this reality that the Copyright Act seeks to avoid. While IA claims that prohibiting its practices would harm consumers and researchers, allowing its practices would — and does — harm authors. With each digital book IA disseminates, it deprives Publishers and authors of the revenues due to them as compensation for their unique creations. Sandra Cisneros, an author who submitted a declaration in support of Publishers’ summary judgment motion, captured the effect of IA’s infringement: I worked hard to earn the financial security that I now have and which enables me to earn a living from my pen without fear of poverty. And, as my agent reminds me, the royalty revenues I receive from the sales of books I have written are precious and must be closely guarded because this is ultimately going to generate the money that supports me in old age…. When I went on the Internet Archive’s website and saw that scans of my books were being distributed to anybody who wanted them for free — without my permission or any payment — I was appalled. I found the experience so viscerally upsetting that I could not stay on the website for long. It was like I had gone to a pawn shop and seen my stolen possessions on sale. App’x 249-50. Though IA and its amici may lament the consolidation of editorial power and criticize Publishers for being motivated by profits, behind Publishers stand authors who are entitled to compensation for the reproduction of their works and whose “private motivation…ultimately serve[s] the cause of promoting broad public availability of literature, music, and the other arts.” Warhol II, 598 U.S. at 526 (quoting Twentieth Century Music Corp., 422 U.S. at 156). IA’s Free Digital Library undermines that motivation. In sum, IA has not met its “burden of proving that the secondary use does not compete in the relevant market[s].” Warhol I, 11 F.4th at 49. Its empirical evidence does not disprove market harm, and Publishers convincingly claim both present and future market harm. Any short-term public benefits of IA’s Free Digital Library are outweighed not only by harm to Publishers and authors but also by the long-term detriments society may suffer if IA’s infringing use were allowed to continue. For these reasons, the fourth fair use factor favors Publishers. V. Weighing the factors together While fair use “presents a mixed question of law and fact, it may be resolved on summary judgment where, as here, the material facts are not in dispute.” Warhol I, 11 F.4th at 36. Having considered the four factors, we find that each favors Publishers. Even if certain tools associated with IA’s reproductions offer some minor functional benefits not associated with the original works (such as linkability), because IA offers these functionalities by reproducing and making publicly available the copyrighted works in their entirety, those benefits are overwhelmingly outweighed by the other fair use considerations. See TVEyes, 883 F.3d at 174 (concluding that even though TVEyes’ re-distribution of Fox’s audiovisual content had some transformative character, “because that redistribution makes available virtually all of Fox’s copyrighted audiovisual content — including all of the Fox content that TVEyes’s clients wish to see and hear — and because it deprives Fox of revenue that properly belongs to the copyright holder, TVEyes has failed to show that the product it offers to its clients can be justified as a fair use.”). Accordingly, IA’s defense of fair use fails as a matter of law.¹² CONCLUSION The parties in this case represent potentially serious interests. On the one hand, eBook licensing fees may impose a burden on libraries and reduce access to creative work. On the other hand, authors have a right to be compensated in connection with the copying and distribution of their original creations. Congress balanced these “competing claims upon the public interest” in the Copyright Act. Twentieth Century Music Corp., 422 U.S. at 156. We must uphold that balance here. IA asks this Court to bless the large scale copying and distribution of copyrighted books without permission from or payment to the Publishers or authors. Such a holding would allow for widescale copying that deprives creators of compensation and diminishes the incentive to produce new works. This may be what IA and its amici prefer, but it is not an approach that the Copyright Act permits. For these reasons, we AFFIRM.