Full opinion text
OPINION PATEL, Chief Judge. The matter before the court concerns the boundary between sharing and theft, personal use and the unauthorized worldwide distribution of copyrighted music and sound recordings. On December 6, 1999, A & M Records and seventeen other record companies (“record company plaintiffs”) filed a complaint for contributory and vicarious copyright infringement, violations of the California Civil Code section 980(a)(2), and unfair competition against Napster, Inc., an Internet start-up that enables users to download MP3 music files without payment. On January 7, 2000, plaintiffs Jerry Leiber, Mike Stoller, and Frank Music Corporation filed a complaint for vicarious and contributory copyright infringement on behalf of a putative class of similarly-situated music publishers (“music publisher plaintiffs”) against Napster, Inc. and former CEO Eileen Richardson. The music publisher plaintiffs filed a first amended complaint on April 6, 2000, and on May 24, 2000, the court entered a stipulation of dismissal of all claims against Richardson. Now before this court is the record company and music publisher plaintiffs’ joint motion to preliminarily enjoin Napster, Inc. from engaging in or assisting others in copying, downloading, uploading, transmitting, or distributing copyrighted music without the express permission of the rights owner. In opposition to this motion, defendant seeks to expand the “fair use” doctrine articulated in Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417, 104 S.Ct. 774, 78 L.Ed.2d 574 (1984), to encompass the massive downloading of MP3 files by Napster users. Alternatively, defendant contends that, even if this third-party activity constitutes direct copyright infringement, plaintiffs have not shown probable success on the merits of their contributory and vicarious infringement claims. Defendant also asks the court to find that copyright holders are not injured by a service created and promoted to facilitate the free downloading of music files, the vast majority of which are copyrighted. Having considered the parties’ arguments, the court grants plaintiffs’ motion for a preliminary injunction against Napster, Inc. The court makes the following Findings of Fact and Conclusions of Law to support the preliminary injunction under Federal Rules of Civil Procedure 65(d). I. FINDINGS OF FACT A. MP3 Technology I. Digital compression technology makes it possible to store audio recordings in a digital format that uses less memory and may be uploaded and downloaded over the Internet. See David M. Lisi Dec. (Ty-gar Rep.) at 11. MP3 is a popular, standard format used to store such compressed audio files. See Edward Kessler Dec. ¶ 3; Lisi Dec. (Tygar Rep.) at 11. Compressing data into MP3 format results in some loss of sound quality. See List Dec. (Tygar Rep.) at 12. However, because MP3 files are smaller, they require less time to transfer and are therefore better suited to transmission over the Internet. See id. at 11. 2. Consumers typically acquire MP3 files in two ways. First, users may download audio recordings that have already been converted into MP3 format by using an Internet service such as Napster. See Lisi Dec. (Tygar Rep.) at 11. Second, “ripping” software makes it possible to copy an audio compact disc (“CD”) directly onto a computer hard-drive; ripping software compresses the millions of bytes of information on a typical CD into a smaller MP3 file that requires a fraction of the storage space. See id.; Kessler Dec. ¶ 32; 1 Laurence F. Pulgram Dec., Exh. A (Conroy Dep.) at 13:19-24. B. Defendant’s Business 1. Napster, Inc. is a start-up company based in San Mateo, California. It distributes its proprietary file-sharing software free of charge via its Internet website. People who have downloaded this software can log-on to the Napster system and share MP3 music files with other users who are also logged-on to the system. See Kessler Dec. ¶ 6. It is uncontradicted that Napster users currently upload or download MP3 files without payment to each other, defendant, or copyright owners. According to a Napster, Inc. executive summary, the Napster service gives its users the unprecedented ability to “locate music by their favorite artists in MP3 format.” 1 Frackman Dec., Exh. A (Richardson Dep.), Exh. 127 at ER000131. Defendant boasts that it “takes the frustration out of locating servers with MP3 files” by-providing a peer-to-peer file-sharing system that allows Napster account holders to conduct relatively sophisticated searches for music files on the hard drives of millions of other anonymous users. See A & M Records, Inc. v. Napster, Inc., 2000 WL 573136, at *1 (N.D.Cal. May 12, 2000) (citing Def.’s Mot. for Summ. Adjud.) at 4. 2. Although Napster was the brainchild of a college student who wanted to facilitate music-swapping by his roommate, see 1 Frackman Dec., Exh. B (Fanning Dep.) at 31:10-35:1, it is far from a simple tool of distribution among friends and family. According to defendant’s internal documents, there will be 75 million Napster users by the end of 2000. See 1 Frackman Dec., Exh. A (Richardson Dep.) at 318:19-319:1, Exh. 166 at 002725. At one point, defendant estimated that even without marketing, its “viral service” was growing by more than 200 percent per month. Id., Exh. 127 at ER00130. Approximately 10,-000 music files are shared per second using Napster, and every second more than 100 users attempt to connect to the system. See Kessler Dec. ¶ 29. 3. Napster, Inc. currently collects no revenues and charges its clientele no fees; it is a free service. See, e.g., 1 Frackman Dec., Exh. A (Richardson Dep.) at 179:15. However, it has never been a non-profit organization. See id. at 116:10. It plans to delay the maximization of revenues while it attracts a large user base. See id., Exh. 127 at ER00130; 1 Frackman Dec., Exh. C (Parker Dep.) at 160:1-162:14, Exh. 254 at SF00099. The value of the system grows as the quantity and quality of available music increases. See id. at 112:18-113:2, Exh. 127 at ER00130; David J. Teece Rep. at 4. Defendant’s internal documents reveal a strategy of attaining a “critical mass” of music in an “ever-expanding library” as new members bring their MP3 collections online. See 1 Frackman Dec. (Richardson Dep.), Exh. 127 at ER00130; Exh. C (Parker Dep.) at 160:1-162:14, Exh. 254 at SF00099. Defendant eventually plans to “monetize” its user base. See id. at 115:24-116:13; Teece Rep. at 4, 7-11. Potential revenue sources include targeted email; advertising; commissions from links to commercial websites; and direct marketing of CDs, Napster products, and CD burners and rippers. See 1 Frackman Dec., Exh. C (Parker Dep.) at 160:1-162:14, Exh. 254 at SF00099-100; Teece Rep. at 2-3. Defendant also may begin to charge fees for a premium or commercial version of its software. See Teece Rep. at 8; cf. 1 Frackman Dec., Exh. A (Richardson Dep.) at 179:6-25. The existence of a large user base that increases daily and can be “monetized” makes Napster, Inc. a potentially attractive acquisition for larger, more established firms. See Teece Rep. at 7. 4. Napster Inc.’s value — which is measured, at least in part, by the size of its user base — lies between 60 and 80 million dollars. See Teece Rep. at 11-12; Def.’s Opp. at 35. Defendant obtained substantial capital infusions after the onset of this litigation. For example, in May 2000, the venture firm Hummer Winblad purchased a twenty-percent ownership interest in the company for 13 million dollars; other investors simultaneously invested 1.5 million dollars. See Hank Barry Dec. ¶ 7. 5. The evidence shows that virtually all Napster users download or upload copyrighted files and that the vast majority of the music available on Napster is copyrighted. Eighty-seven percent of the files sampled by plaintiffs’ expert, Dr. Ingram Olkin, “belong to or are administered by plaintiffs or other copyright holders.” Olkin Rep. at 7. After analyzing Olkin’s data, Charles J. Hausman, anti-piracy counsel for the RIAA, determined that 834 out of 1,150 files in Olkin’s download database belong to or are administered by plaintiffs; plaintiffs alone own the copyrights to more than seventy percent of the 1,150 files. See Charles J. Hausman Dec. ¶ 8. Napster users shared these files without authorization. See id. 6. Napster, Inc. has never obtained licenses to distribute or download, or to facilitate others in distributing or downloading, the music that plaintiffs own. See Kevin Conroy Dec. ¶ 4; Richard Cottrell Dec. ¶ 5; Mark R. Eisenberg Dec. ¶ 21; Lawrence Kenswil Dec. ¶ 15; Paul Vidich Dec. ¶ 8; Mike Stoller Dec. ¶ 11. 7. Defendant’s internal documents indicate that it seeks to take over, or at least threaten, plaintiffs’ role in the promotion and distribution of music. See, e.g., 1 Frackman Dec., Exh. C (Parker Dep.), 160:1-162:14, Exh. 254, at SF00099 (declaring that “[ujltimately Napster could evolve into a full-fledged music distribution platform, usurping the record industry as we know it today and allowing us to digitally promote and distribute emerging artists at a fraction of the cost” but noting that “we should focus on our realistic short-term goals while wooing the industry before we try to undermine it”). 8.Defendant’s internal documents also demonstrate that its executives knew Napster users were engaging in unauthorized downloading and uploading of copyrighted music. See, e.g. 1 Frackman Dec., Exh. C (Parker Dep.) at 160:1-162:14, Exh. 254 at SF00100 (stating that Napster users “are exchanging pirated music.”); id. at SF00102 (“[W]e are not just making pirated music available but also pushing demand”). Several Napster executives admitted in their depositions that they believed many of the millions of MP3 music flies available on Napster were copyrighted. See, e.g., 1 Frackman Dee., Exh. B (Fanning Dep.) at 105:10-108:2. 9. At least on paper, the promotion of new artists constituted an aspect of defendant’s plan as early as October 1999. See Sean F. Parker Dec. ¶ 5 & Exh. B ; Scott Krause Dec. ¶ 6. New or unsigned artists now may promote their works and distribute them in MP3 format via the Napster service. See Krause Dec. ¶¶ 8-15. Napster, Inc. has sought business alliances and developed both Internet- and software-based technologies to support its New Artist Program. See Parker Dec. ¶ 6. However, the court finds that the New Artist Program accounts for a small portion of Napster use and did not become central to defendant’s business strategy until this action made it convenient to give the program top billing. An early version of the Napster website advertised the ease with which users could find their favorite popular music without “wading through page after page of unknown artists.” 1 Frackman Dec., Exh. C (Parker Dep.) at 104:16-105:10, Exh. 235. Defendant did not even create the New Artist Program that runs on its Internet website until April 2000 — well after plaintiffs filed this action. See Krause Dec. ¶ 9, Exh. A. Moreover, in Olkin’s sample of 1,150 files (which were randomly selected from over 550,000), only 232 files matched any of the 19,440 names that were listed in defendant’s new artist database as of July 2000. See Olkin Reply Dec. ¶¶ 3-5; Hausman Reply Dec. ¶¶ 3-6. An RIAA representative who analyzed the data also noted that the list of so-called new artists actually contained many popular stars represented by major record labels — among them teen sensation Britney Spears and the legendary alternative rock band Nirvana. See Hausman Reply Dec. ¶ 5. Once established artists were eliminated from the results, only eleven new artists and fourteen of their music files remained in Ol-kin’s sample of 1,150 files. See id. ¶ 6. 10. Defendant employs the term “space-shifting” to refer to the process of converting a CD the consumer already owns into MP3 format and using Napster to transfer the music to a different computer — from home to office, for example. See Def. Opp. at 12. The court finds that space-shifting accounts for a de minimis portion of Napster use and is not a significant aspect of defendant’s business. According to the court’s understanding of the Napster technology, a user who wanted to space-shift files from her home to her office would have to log-on to the system from her home computer, leave that computer online, commute to work, and log-on to Napster from her office computer to access the desired file. In the meantime, many users might download it before she reached the office. Common sense dictates that this use does not draw users lo the system. Defendant fails to cite a single Napster, Inc. document indicating that the company saw space-shifting as an attraction for its user base, and survey evidence shows that almost half of college-student survey respondents previously owned less than ten percent of the songs they have downloaded. See E. Deborah Jay Rep. at 4, 21 & Tbl. 7. C. The Napster Technology 1. Internet users may download defendant’s proprietary MusicShare software free of charge from the Napster website. This free software enables users to access the Napster computer network. See Kes-sler Dec. ¶ 6. 2. The software becomes fully functional after users register with Napster by selecting an account name, or “user name,” and a password. See Kessler Dec. ¶¶ 6, 23. Persons who register may include biographical data, but registration does not require a real name or address. See 2 Frackman Dec., Exh. E (Kessler Dep.) at 255:20-257:22. Napster does not associate user names with the biographical information that individuals provide at registration. See id. Indeed, after a user logs-on, her physical address information is no longer available to the Napster server. See id. 3. The software features a browser interface, search engine, and chat functions that operate in conjunction with defendant’s online network of servers. See id. ¶¶ 6, 13. The software also contains a “hotlist” tool that allows users to compile and store lists of other account holders’ user names. See id. ¶ 8. In addition, the Napster software may be used to play and categorize audio files, which users can store in specific file directories on their hard drives. See id. ¶¶ 6-7. Those directories, which allow account holders to share files on Napster, constitute the “user library.” Id. Some users store their MP3 files in such directories; others do not. See id. 4. Defendant maintains clusters of servers that compose its network or system. See Kessler Dec. ¶ 13. Account holders who access the Napster network may communicate, share files, and learn of designated hotlist names only within the cluster to which they are assigned. See id. Users can access the network of servers free of charge. 5. Once an account holder signs on to the Napster network, the Napster browser interacts with its proprietary server-side software. See id. ¶¶ 7, 8; 2 Frackman Dec., Exh. E (Kessler Dep.) at 54:16-56:10. If a user sets the “allowable uploads” function of the MusicShare software above zero, all of the MP3 file names she stores in her user library automatically become available to other online Napster users. See Kessler Dec. ¶ 7. However, before the client software uploads MP3 file names to defendant’s master servers, it “validates” the files stored in the user library directories. See 2 Frackman Dec., Exh. E (Kessler Dep.) at 145:2-18. The client software reads those files to ensure they are indeed MP3 files, checking to see whether they contain the proper syntax specification and content. See id. If the files are not properly formatted, their file names will not be not uploaded to the Napster servers. See id. Once the file names are successfully uploaded to the servers, each user library, identified by a user name, becomes a “location” on the servers. Kessler Dec. ¶ 8. Napster locations are short-lived; they are respectively added or purged every time a user signs on or off of the network. See id. Thus, a user’s MP3 files are only accessible to other users while she is online. 6. A user who is logged-on to the Napster servers via the client software may access the content of other users’ uploaded “locations” in one of two ways: (a) by utilizing defendant’s proprietary search engine, or (b) by employing the hotlist tool featured in the client software. See id. ¶ 12. 7.An account holder may use the search tools included in the Napster client software to find MP3 files. See id. ¶ 10. The server-side application software maintains a search index that is updated in real time as users log-on and -off of the system. See id.; 2 Frackman Dec., Exh. E (Kessler Dep.) at 56:3-10. The file-name index contains the names of MP3 files that on-line users save in their designated user library directories. See Kessler Dec. ¶¶ 7, 14; 2 Frackman Dec., Exh. E (Kessler Dep.) at 55:14-56:10; Exh. 2. Users who wish to search for a song or artist may do so by entering the name of the song or artist in the search fields of the client software and then clicking the “Find It” button. When the search form is transmitted to the Napster network, the Napster servers send the requesting user a list of files that include the same term(s) she entered on the search form. See Kessler Dec. ¶ 5; 2 Frackman Dec., Exh. E (Kessler Dep.) at 56:3-10. After the application software returns a list of specific MP3 file names to the requesting user, the user then must peruse the list to determine whether she desires any of those files. See id. ¶ 10. She must read through the list because the Napster application software does not search for a particular song or recording artist per se. Napster does not organize MP3 files based on content because, currently, they are not designed for such indexing. See id. ¶ 11. Instead, Napster performs a text search of the file names indexed in a particular cluster. Those file names may contain typographical errors or otherwise inaccurate descriptions of the content of the files since they are designated by other users. See id. ¶¶ 13, 10, 27; 2 Pulgram Dec. Exh. B (Fanning Dep.) at 116:8-19. In addition to listed text results from an executed search, Napster’s servers provide other information about particular MP3 files. For instance, the client software can sort the results of “echo packets” or “ping requests” that it sends out to host users; these requests help gauge the “responsiveness value” of a transmission between two users by. calculating the amount of time it takes for ping responses to be returned to the client software. See 2 Frackman Dec., Exh. E (Kessler Dep.) at 56:3-10, Exh. 5 at 3; Shawn Fanning Dec. ¶ 8. Users can also search for files that meet certain technical criteria, such as the host user’s bandwidth. See id. Finally, the file name or “data object description” includes the size and bytes stored and “attributes of quality,” such as bit rate. See 2 Frackman Dec., Exh. E (Kessler Dep.) at 153:16-154:24; Fanning Dec. ¶ 8. These Napster options contribute to the ease with which the user can locate and obtain the music she wants. 8. Alternatively, users may access MP3 files via the hotlist function. This function enables a Napster user to archive other user names and learn whether account holders who access the network under those names are online. See Kessler Dec. ¶¶ 8-9. A requesting user can access or browse all files listed in the user libraries of hotlisted users. See id. ¶ 9. Then she can request a particular file in a host user’s user library by selecting, or clicking on, that file name. See id. The hotlist function is a feature that helps make Napster users a virtual community — they are not only able to download the music they desire, but also to obtain files from particular individuals whom they know by user name. 9. The Napster network facilitates the same mode of file-transfer, whether a requesting user accesses a specific MP3 file with the search engine or the hotlist. See id. ¶ 12. Once a requesting user locates and selects the file she wishes to download, the server-side software engages in a dialogue with her browser and that of the “host user” (that is, the user who makes the MP3 available for downloading). See Kessler Dec. ¶ 12; 2 Frackman Dec., Exh. E (Kessler Dep.) at 80:19-22; 56:3-10. Napster servers obtain the necessary IP address information from the host user. See Daniel Farmer Dec. ¶ 17; Frackman Dec., Exh. 1 (Kessler Dep.) at 103-05. The servers then communicate the host user’s address or routing information to the requesting user; the requesting user’s computer employs this information to establish a connection with the host user’s browser software and download the MP3 file from the host user’s library. See Kessler Dec. ¶¶ 10-13; 2 Frackman Dec., Exh. E (Kessler Dep.) at 56:3-10. The content of the actual MP3 file is transferred over the Internet between users, not through the Napster servers. See Kessler Dec. ¶ 12; A & M Records, Inc. v. Napster, Inc., 2000 WL 573136, at *7 (N.D.Cal. May 12, 2000). However, users would not be able to access the uploaded file names and corresponding routing data without signing on to the Napster system. See Kessler Dec. ¶ 23. 10. In some instances, a requested file is not immediately ready for download. Those files are “queued” or deferred until the host user is able to transmit the file. See 2 Frackman Dec., Exh. E (Kessler Dep. at 80:2-22). The request may be deferred, for example, because the host user has limited the number of downloads she can provide simultaneously, or because the host user has signed off the Napster network. See id. Defendant employs technology that permits users to resume queued downloads at a later time. See 2 Frackman Dec., Exh. E (Kessler Dep.) at 112:3-13. Every MP3 file has a mathematically-generated and unique fingerprint or “checksum.” See Kessler Dec. ¶ 32; 2 Frackman Dec., Exh. E (Kessler Dep.) at 112:3-13. Any requesting user who is unable to download a particular MP3 file may use the client software to send the file’s checksum and full intended size to the Napster servers and attempt to locate a match for download. See 2 Frackman Dec., Exh. E (Kes-sler Dep.) at 112:3-13. 11. Defendant also provides Napster users with a chat service. Its central servers permit users who are logged-on to communicate with other online users, including those whose user names comprise the hotlist. See Kessler Dec. ¶ 13. Aside from communicating with specific online users logged-on to the same cluster of servers, the chat service allows users to communicate in groups. Defendant organizes these groups within “channels” or “chat rooms” named after particular musical genres. See id.; 2 Pulgram Dec., Exh. B (2 Fanning Dep.) at 219: 4-14. Alternatively, users can create their own channels in which to communicate. See Kessler Dec. ¶ 13. 12.Defendant’s New Artist Program technology functions in two interrelated environments: (a) on its Internet website and (b) through its network-client browser and search technology. See Krause Dec. ¶¶ 9-15. The website version performs several functions. It allows new or unsigned artists to create a “profile” that consists of certain biographical and descriptive data including artist and band names, similar artists or influences, and news about the band. See id. ¶¶ 9-10, Exh. C. Defendant only accepts completed forms if the submitting artist authorizes Napster users to share his music. See id. ¶ 10, Exh. D. Once defendant accepts the profile, it stores all of the relevant information in a database linked to its Internet website. See id. ¶ 11. Defendant has accepted several thousand such profiles. See id. ¶¶ 9,16. Members of the public can then search the new artist database in several ways: (1) by artist name, (2) by artist influence, or (3) by browsing the different genres of music and then scrolling down lists of new artists categorized in those genres. See id. ¶ 12. The Napster site does not store any of the new artists’ music, however. See id. ¶ 9. Instead, those who access the website-based service acquire information about an artist, such as his name. See id., Exh. E. Once an individual obtains this data, she is directed to switch to Napster’s software- and network-based service to search for and download the new artist’s music. See id. ¶ 15. Napster account holders who use Mu-sicShare software and log-on to the Napster system can locate and download new artists’ songs in the same manner they would find and download any other files: by utilizing the search engine, or by browsing user libraries. See id. ¶ 9. While on the Napster network, both new artists and other users may use the chat function to market music directly or learn about new artists. See id. ¶¶ 8,15. D. Plaintiffs’ Business 1. The music publisher plaintiffs compose music and write songs. See, e.g., Stoller Dec. ¶ 2. They depend financially upon the sale of sound recordings because they earn royalties from such sales. See id. at ¶¶ 2,11,13. However, they do not get a royalty when a Napster user uploads or downloads an MP3 file of their compositions without payment or authorization. See id. ¶ 11. The record company plaintiffs’ sound recordings also result from a substantial investment of money, time, manpower, and creativity. See Conroy Dec. ¶ 5; Cottrell Dec. 5; Eisenberg Dec. ¶¶ 5, 21; Kenswil Dec. ¶ 5. In contrast, defendant invests nothing in the content of the music which means that, compared with plaintiffs, it incurs virtually no costs in providing a wide array of music to satisfy consumer demand. See Teece Rep. at 14. 2. To make a profit, the record company plaintiffs largely rely on the success of “hit” or popular recordings, which may constitute as little as ten or fifteen percent of albums released. See, e.g., Eisenberg Dec. ¶ 7. Many, or all, of their top recordings have been available for free on Napster. See Frank Creighton Dec. ¶ 5. 3. The record company plaintiffs have invested substantial time, effort, and funds in actual or planned entry into the digital downloading market. BMG Music (“BMG”) began to explore digital downloading in early 1996 and has made more than twenty tracks commercially available for downloading through the digital service providers (“DSPs”) Amplified.com and the Liquid Music Network. See Conroy Dec. ¶ 9. BMG has entered several business partnerships, strategic marketing agreements, and clearinghouse relationships to develop a plan for secure, commercial digital downloading; July 2000 was the target date for BMG’s launch. See id. at ¶¶ 10-17. Plaintiffs Capitol Record, Inc. and Virgin Records America are affiliated with EMI Recorded Music, North America (“EMI”). See Cottrell Dec. ¶ 1. EMI has developed business plans to distribute its music through several DSPs which represent more than 800 retail websites. See id. ¶ 7. All digital downloads that EMI offers will be encrypted and watermarked. See id. ¶ 12. Sony Music Entertainment (“Sony”) has already begun to make selected singles available through its websites and those of its artists; to- obtain a permanent copy of this music, consumers must pay for the download. See Eisenberg Dec. ¶ 13. As of May 31, 2000, Sony also began selling downloadable music through a distribution network of about thirty-five retail sites. See id. at ¶ 17. Plaintiffs A & M Records, Geffen Records, Interscope Records, Island Records, MCA Records, Motown Records, UMG Records, and Universal Records (collectively, “Universal”) have spent millions of dollars preparing a secure digital distribution system scheduled for launch in midsummer 2000. See Kenswil Dec. ¶ 9-16. Warner Music Group and its associated labels — plaintiffs Atlantic Recording Corp., London-Sire Records Inc. (f/k/a Sire Records Group Inc.), Elektra Entertainment Group Inc., and Warner Bros. Records (collectively, ‘Warner”) — have done due diligence and dedicated a substantial budget to digital distribution. See Vidich Dec. ¶ 7(a)-(e). Warner expects to launch its commercial digital distribution of hundreds of recordings by the fourth quarter of 2000. See id. at ¶ 7(e) 4. Promotional samples offered by plaintiffs and other retail sites differ significantly from using Napster to decide whether to buy a CD. The record company plaintiffs have made some free downloads available but have limited them in amount and duration. They have not provided entire albums, and the downloads typically have been “timed-out” so that users can only play them for a finite period of time— often less than a month. See Conroy Dec. ¶ 9-17; Cottrell Dec. ¶ 15; Eisenberg Dec. ¶ 13; Kenswil Dec. ¶ 12; Vidich Dec. ¶¶ 7(d), 8. Although plaintiffs have not been completely successful in managing the rights to promotional downloads, record company executives accord importance to the security of music distributed in this manner. See Conroy Dec. ¶ 9-17; Cott-rell Dec. ¶ 15; Eisenberg Dec. ■ ¶ 13; Ken-swil Dec. ¶ 12; Vidich Dec. ¶¶ 7(d), 8. Retail sites, such as Amazon.com, offer thirty-to-sixty-second song samples in streaming audio format, rather than as downloads. See David Lambert Reply Dec. ¶ 2. Unlike downloading, streaming does not copy the music onto the listener’s computer hard drive; it merely allows her to hear it. See id. Because companies like DiscoverMusic that provide song samples to these Internet retailers enter licensing agreements, rights holders earn royalties from this form of sampling. See id. ¶ 3. In contrast, persons who obtain MP3 files for free using Napster can retain and play them indefinitely- — and, even if they download a song to make a purchasing decision, they may decide not to buy the music. While Napster users can burn CDs comprised of unauthorized downloads they obtained to “sample” new songs, sampling on sites affiliated with plaintiffs does not substitute for purchasing the entire disc. See Teece Rep. at 17. E. Effect of Napster on the Market for Plaintiffs’ Copyrighted Works 1. The court finds that Napster use is likely to reduce CD purchases by college students, whom defendant admits constitute a key demographic. See Jay Rep. at 4, 18; Michael Fine Rep. at 1; Julia Greer Reply Dec. (Brooks Dep.) at 145:10-12 (“We believes [sic] ourselves to have a high college demographic, and beyond that to be primarily [ages] 12 to 24.”). Plaintiffs’ expert, Dr. E. Deborah Jay, opined that forty-one percent of her college-student survey respondents “gave a reason for using Napster or described the nature of its impact on their music purchases in a way which either explicitly indicated or suggested that Napster displaces CD sales.” Jay Rep. at 4, 18. She also found that twenty-one percent of the college students surveyed revealed that Napster helped them make a better selection or decide what to buy. See id., Tbl. 4. However, Jay’s overall conclusion was that “[t]he more songs Napster users have downloaded,” the more likely they are to admit or imply that such use has reduced their music purchases. See id. at 4, 18. The report of Soundscan CEO Michael Fine lends support to Jay’s findings. After examining data culled from three types of retail stores near college or university campuses, Fine concluded that “on-line file sharing has resulted in a loss of album sales within college markets.” Fine Rep. at 1. For the reasons discussed in the court’s separate order, the report by defendant’s expert, Dr. Peter S. Fader, does not provide credible evidence that music file-sharing on Napster stimulates more CD sales than it displaces. Nor do the recording industry documents that defendant cites reliably show increased music sales due to Napster use. One such memorandum deals with the effect of Warner’s promotional downloads, which are “timed-out” and thus differ from MP3 files obtained using Napster. See 1 Pulgram Dec., Exh. N (Vidich Dep.), Exh. 279 at T3122-23; Vidich Dec. ¶¶7(d), 8 (stating that free, promotional downloads are “timed-out.”). Another purported “smoking gun” is a Universal survey on music-purchasing by people who download MP3 files. See 1 Pulgram Dec., Exh. F (Kenswil Dep.) at 110:22-111:15. However, the court has too little information about this survey to rely on it, and the deponent, Universal representative Lawrence Kenswil, declined to vouch for the survey’s accuracy. See id. at 111:8,14-15. 2. Because plaintiffs entered the digital download market very recently, or plan to enter it in the next few months, they are especially vulnerable to direct competition from Napster, Inc. See Teece Rep. at 15-16. The court finds that, in choosing between the free Napster service and pay-per-download sites, consumers are likely to choose Napster. See id. at 14; Jay Rep. at 4, 18 (reaching this conclusion with regard to a survey sample of college students). Defendant’s economic expert, Dr. Robert E. Hall, opines that plaintiffs’ music could still command a high price after a period when the price has been zero due to Napster use; thus, he concludes, plaintiffs will not suffer irreparable harm between now and a trial verdict against defendant. See Lisi Dec. (Hall Rep.) ¶¶ 39, 54. This argument does not square with Hall’s assertion that preliminarily enjoining defendant will put it out of business because users will switch to services offered by “kindred spirits.” Hall Rep. ¶ 15-19, 73; see also Barry Dec. ¶ 13. If this is true, consumers will not necessarily resume buying music if Napster is enjoined; rather, they will go to other sites offering free MP3 files. Indeed, as Dr. David J. Teece avers, defendant has contributed to a new attitude that digitally-downloaded songs ought to be free — an attitude that creates formidable hurdles for the establishment of a commercial downloading market. See Teece Rep. at 14-18. Hall also maintains that Napster, Inc. will increase the volume of plaintiffs’ online sales by stimulating consumer investment in the hardware and software needed to obtain and play MP3 files. See Hall Rep. ¶¶ 45-49. However, he ignores evidence of reduced CD-buying among college students due to Napster use, see Jay Rep. at 18, and the data upon which he relies to argue that Napster has enhanced sales is either weak (in the case of the Fader Report) or unavailable for the court’s review. See, e.g., id. ¶ 17 (relying on IDC and Forrester Research studies), ¶¶ 27-28 (discussing reports by Student Monitor and Andersen Consulting), ¶ 34 (citing a study by the University of Southern California). The court therefore finds that the barriers to commercial distribution posed by an emerging sense of entitlement to free music probably outweigh the benefits that defendant purports to confer. 3. Downloading on Napster also has the potential to disrupt plaintiffs’ promotional efforts because it does not involve any of the restrictions on timing, amount, or selection that plaintiffs impose when they offer free music files. See Conroy Dec. ¶ 19-17; Cottrell Dec. ¶ 15; Eisen-berg Dec. ¶ 13; Kenswil Dec. ¶ 12; Vidich Dec. ¶¶ 7(d), 8; see also Teece Rep. at 18. Even if Napster users sometimes download files to determine whether they want to purchase a CD, sampling on Napster is vastly different than that offered by plaintiffs. On Napster, the user — not the copyright owner — determines how much music to sample and how long to keep it. II. CONCLUSIONS OF LAW A. Legal Standard 1. The Ninth Circuit authorizes preliminary injunctive relief for “a party who demonstrates either (1) a combination of probable success on the merits and the possibility of irreparable harm, or (2) that serious questions are raised and the balance of hardships tips in its favor.” Prudential Real Estate Affiliates, Inc. v. PPR Realty, Inc., 204 F.3d 867, 874 (9th Cir.2000). 2. The standard is a sliding scale which requires a greater degree of harm the lesser the probability of success. See id. In a copyright infringement case, demonstration of a reasonable likelihood of success on the merits creates a presumption of irreparable harm. See Micro Star v. Formgen, Inc., 154 F.3d 1107, 1109 (9th Cir.1998). B. Proof of Direct Infringement 1. To prevail on a contributory or vicarious copyright infringement claim, a plaintiff must show direct infringement by a third party. See Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 434, 104 S.Ct. 774, 78 L.Ed.2d 574 (1984). As a threshold matter, plaintiffs in this action must demonstrate that Napster users are engaged in direct infringement. 2. Plaintiffs have established a prima facie case of direct copyright infringement. As discussed above, virtually all Napster users engage in the unauthorized downloading or uploading of copyrighted music; as much as eighty-seven percent of the files available on Napster may be copyrighted, and more than seventy percent may be owned or administered by plaintiffs. See Olkin Rep. at 7; Hausman Dec. ¶ 8. C. Affirmative Defense of Fair Use and Substantial Non-Infringing Use 1. Defendant asserts the affirmative defenses of fair use and substantial non-infringing use. The latter defense is also known as the staple article of commerce doctrine. See Sony, 464 U.S. at 442, 104 S.Ct. 774. Sony stands for the rule that a manufacturer is not liable for selling a “staple article of commerce” that is “capable of commercially significant non-infringing uses.” Id. The Supreme Court also declared in Sony, “Any individual may reproduce a copyrighted work for a ‘fair use’; the copyright holder does not possess the exclusive right to such a use.” Id. at 433, 104 S.Ct. 774. Defendant bears the burden of proving these affirmative defenses. See Bateman v. Mnemonics, Inc., 79 F.3d 1532, 1542 n. 22 (11th Cir.1996) (“[I]t is clear the burden of proving fair use is always on the putative infringer.”). 2. For the reasons set forth below, the court finds that any potential non-infringing use of the Napster service is minimal or connected to the infringing activity, or both. The substantial or commercially significant use of the service was, and continues to be, the unauthorized downloading and uploading of popular music, most of which is copyrighted. 3. Section 107 of the Copyright Act provides a non-exhaustive list of fair use factors. These factors include: (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. 17 U.S.C. § 107. 4. In the instant action, the purpose and character of the use militates against a finding of fair use. Ascertaining whether the new work transforms the copyrighted material satisfies the main goal of the first factor. See Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579, 114 S.Ct. 1164, 127 L.Ed.2d 500 (1994). Plaintiff persuasively argues that downloading MP3 files does not transform the copyrighted music. See UMG Recordings, Inc. v. MRS.Com, Inc., 92 F.Supp.2d 349, 351 (S.D.N.Y.2000) (concluding that repackaging copyrighted recordings in MP3 format suitable for downloading “adds no ‘new aesthetics, new insights and understandings’ to the original”). 5. Under the first factor, the court must also determine whether the use is commercial. In Acuff-Rose, the Supreme Court clarified that a finding of commercial use weighs against, but does not preclude, a determination of fairness. See Acuff-Rose, 510 U.S. at 584, 114 S.Ct. 1164. 6. If a use is non-commercial, the plaintiff bears the burden of showing a meaningful likelihood that it would adversely affect the potential market for the copyrighted work if it became widespread. See Sony, 464 U.S. at 451, 104 S.Ct. 774. 7. Although downloading and uploading MP3 music files is not paradigmatic commercial activity, it is also not personal use in the traditional sense. Plaintiffs have not shown that the majority of Napster users download music to sell — that is, for profit. However, given the vast scale of Napster use among anonymous individuals, the court finds that downloading and uploading MP3 music files with the assistance of Napster are not private uses. At the very least, a host user sending a file cannot be said to engage in a personal use when distributing that file to an anonymous requester. Moreover, the fact that Napster users get for free something they would ordinarily have to buy suggests that they reap economic advantages from Napster use. See Sega Enters. Ltd. v. MAPHIA, 857 F.Supp. 679, 687 (N.D.Cal.1994) (“Sega /”) (holding that copying to save users expense of purchasing authorized copies has commercial character and thus weighs against finding of fair use); cf. American Geophysical Union v. Texaco, Inc., 60 F.3d 913, 922 (2d Cir.1994) (holding that for-profit enterprise which made unauthorized copies of scholarly articles to facilitate scientific research reaped indirect economic advantage from copying and, hence, that copying constituted commercial use). 8. The court finds that the copyrighted musical compositions and sound recordings are creative in nature; they constitute entertainment, which cuts against a finding of fair use under the second factor. See Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539, 563, 105 S.Ct. 2218, 85 L.Ed.2d 588 (1985); Sega I, 857 F.Supp. at 687; Playboy Enters., Inc. v. Frena, 839 F.Supp. 1552, 1558 (M.D.Fla.1993) (citing In re New Era Publications Int’l v. Carol Publ’g, 904 F.2d 152, 157-58 (2d Cir.), cert. denied, 498 U.S. 921, 111 S.Ct. 297, 112 L.Ed.2d 251 (1990)). 9. With regard to the third factor, it is undisputed that downloading or uploading MP3 music files involves copying the entirety of the copyrighted work. The Ninth Circuit held prior to Sony that “wholesale copying of copyrighted material precludes application of the fair use doctrine.” Marcus v. Rowley, 695 F.2d 1171, 1176 (9th Cir.1983). Even after Sony, wholesale copying for private home use tips the fair use analysis in plaintiffs’ favor if such copying is likely to adversely affect the market for the copyrighted material. See Sony, 464 U.S. at 449-50, 456, 104 S.Ct. 774. 10. The fourth factor, the effect on the potential market for the copyrighted work, also weighs against a finding of fair use. Plaintiffs have produced evidence that Napster use harms the market for their copyrighted musical compositions and sound recordings in at least two ways. First, it reduces CD sales among college students. See Jay Rep. at 4, 18; cf. Fine Rep. at 1. Second, it raises barriers to plaintiffs’ entry into the market for the digital downloading of music. See Teece Rep. at 12-18. 11. Defendant asserts several potential fair uses of the Napster service— including sampling, space-shifting, and the authorized distribution of new artists’ work. Sampling on Napster is not a personal use in the traditional sense that courts have recognized — copying which occurs within the household and does not confer any financial benefit on the user. See, e.g., Sony, 464 U.S. at 423, 449-50,104 S.Ct. 774. Instead, sampling on Napster amounts to obtaining permanent copies of songs that users would otherwise have to purchase; it also carries the potential for viral distribution to millions of people. Defendant ignores critical differences between sampling songs on Napster and VCR usage in Sony. First, while “time-shifting [TV broadcasts] merely enables a viewer to see ... a work which he ha[s] been invited to witness in its entirety free of charge,” plaintiffs in this action almost always charge for their music — even if it is downloaded song-by-song. Sony, 464 U.S. at 449-50, 104 S.Ct. 774; see e.g., Conroy Dec. ¶ 9; Eisenberg Dec. ¶ 16. They only make promotional downloads available on a highly restricted basis. See Conroy Dec. ¶ 9-17; Cottrell Dec. ¶ 15; Eisenberg Dec. ¶ 13; Kenswil Dec. ¶ 12; Vidich Dec. ¶¶ 7(d), 8. Copyright owners also earn royalties from streamed song samples on retail websites like Amazon.com. See Lambert Reply Dee. ¶ 3. Second, the majority of VCR purchasers in Sony did not distribute taped television broadcasts, but merely enjoyed them at home. See id. at 423, 104 S.Ct. 774. In contrast, a Napster user who downloads a copy of a song to her hard drive may make that song available to millions of other individuals, even if she eventually chooses to purchase the CD. So-called sampling on Napster may quickly facilitate unauthorized distribution at an exponential rate. Defendant’s argument that using Napster to sample music is akin to visiting a free listening station in a record store, or listening to song samples on a retail website, fails to convince the court because Napster users can keep the music they download. Whether or not they decide to buy the CD, they still obtain a permanent copy of the song. In contrast, many retail sites only offer thirty-to-sixty-second samples in streaming audio format, see Lambert Reply Dec. ¶ 2, and promotional downloads from the record company plaintiffs are often “timed-out.” See Cottrell Dec. ¶ 15; Eisenberg Dec. ¶ 13; Kenswil Dec. ¶ 12; Vidich Dec. ¶¶ 7(d), 8. The global scale of Napster usage and the fact that users avoid paying for songs that otherwise would not be free militates against a determination that sampling by Napster users constitutes personal or home use in the traditional sense. 12. Even if the type of sampling supposedly done on Napster were a non-commercial use, plaintiffs have demonstrated a substantial likelihood that it would adversely affect the potential market for their copyrighted works if it became widespread. See Sony, 464 U.S. at 451, 104 S.Ct. 774. Plaintiffs claim three general types of harm: a decrease in retail sales, especially among college students; an obstacle to the record company plaintiffs’ future entry into the digital downloading market; and a social devaluing of music stemming from its free distribution. With regard to sampling, twenty-one percent of the Jay survey respondents indicated that Napster helps them decide what music to purchase. See Jay Rep., Tbl. 4. Nevertheless, Jay reached the overarching conclusion that the more songs Napster users download, the more likely they are to reveal that such use reduces their music buying. See id. at 4, 18. Jay’s evidence suggests that sampling and building a free music library through unauthorized downloading are not mutually exclusive: it is likely that survey respondents who sample are primarily direct infringers. Napster users — not the record companies — control the music selection, the amount and the timing of the sampling activity, and they may keep many songs after deciding not to purchase the entire CD. Defendant maintains that sampling does not decrease retail music sales and may even stimulate them. To support this assertion, it relies heavily on the Fader Report, which concludes that consumers do not view MP3 files as perfect substitutes for CDs. See Lisi Dec. (Fader Rep.) ¶ 63. Fader cites a survey that he did not conduct for the assertion that “60% of online users who download free digital music do so to preview music before buying the CD.” Fader Rep. ¶ 74. Examining the results of a different survey that he purportedly designed, but did not carefully supervise, he reports that about twenty-eight percent of Napster users indicate that their music purchases have increased since they began using the Napster software. See id. ¶ 43. For reasons explained in the court’s evidentiary order, the Fader Report is unreliable and fails to rebut plaintiffs’ showing of harm. Plaintiffs have demonstrated a meaningful likelihood that the activity defendant calls sampling actually decreases retail sales of their music. 13. Any potential enhancement of plaintiffs’ sales due to sampling would not tip the fair use analysis conclusively in favor of defendant. Indeed, courts have rejected the suggestion that a positive impact on sales negates the copyright holder’s entitlement to licensing fees or access to derivative markets. See Ringgold v. Black Entertainment Television, 126 F.3d 70, 81 n. 16 (2d Cir.1997) (noting that, even if allegedly infringing use of plaintiffs poster in television program increased poster sales, plaintiff retained right to licensing fee); DC Comics, Inc. v. Reel Fantasy, Inc., 696 F.2d 24, 28 (2d Cir.1982) (stating that speculated increase in plaintiffs comic book sales due to unauthorized use of Batman and Green Arrow figures on advertising flyers did not establish fair use defense as matter of law); MP3.Com, 92 F.Supp.2d at 352 (holding that allegedly positive impact on plaintiffs’ prior market “in no way frees defendant to usurp a further market that directly derives from reproduction of [the] copyrighted works.”). The MP3.Com opinion is especially instructive. Although MP3.corn’s activities arguably stimulated CD sales, the plaintiffs “adduced substantial evidence that they ... [had] taken steps to enter [the digital downloading market].” MP3.Com, 92 F.Supp.2d at 352. The fourth factor thus weighed against a finding of fair use. Plaintiffs in the instant action similarly allege that Napster use impedes their entry into the online market. The record company plaintiffs have already expended considerable funds and effort to commence Internet sales and licensing for digital, downloads. See Conroy Dec. ¶¶ 9-18; Cottrell Dec. ¶¶ 6-17; Eisenberg Dec. ¶¶ 9-22; Vidich Dec. ¶¶ 7-10. Plaintiffs’ economic expert opined that the availability of free MP3 files will reduce the market for authorized, commercial downloading. See Teece Dec. at 14-18. This point is corroborated by the fact that all forty-nine songs available for purchase on Sony’s website can be obtained for free using Napster. See Eisenberg Dec. ¶ 16. If consumers choose to buy, rather than burn, entire CDs they are still more likely to obtain permanent copies of songs on Napster than buy them from Sony’s site or listen to streamed samples at other online locations. The court concludes that, even assuming the sampling alleged in this case is a noncommercial use, the record company plaintiffs have demonstrated a meaningful likelihood that it would adversely affect their entry into the online market if it became widespread. See Sony, 464 U.S. at 451, 104 S.Ct. 774. Moreover, it deprives the music publisher plaintiffs of royalties for individual songs. The unauthorized downloading of plaintiffs’ music to sample songs would not constitute a fair use, even if it enhanced CD sales. 14. The court is also unconvinced that Sony applies to space-shifting. Defendant erroneously relies on the Ninth Circuit’s assertion, in a case involving an inapplicable statute, that space-shifting constitutes non-commercial personal use. See Recording Indus. Ass’n of Am. v. Diamond Multimedia Sys., Inc., 180 F.3d 1072, 1079 (9th Cir.1999) (discussing the applicability of the Audio Home Recording Act of 1992 to the Rio MP3 player). Defendant also implies that space-shifting music is sufficiently analogous to time-shifting television broadcasts to merit the protection of Sony. According to the gravely flawed Fader Report, space-shifting — like time-shifting — leaves the value of the copyrights unscathed because it does not displace sales. See Fader Rep. ¶ 77; Sony, 464 U.S. at 421, 104 S.Ct. 774 (concluding that plaintiffs did not prove that time-shifting created any likelihood of harm). Defendant again cites Fader for the statistic that seventy percent of Napster users at least sometimes engage in space-shifting. See Lisi Dec. (Fader Rep.) ¶ 77. In contrast, Jay opined that approximately forty-nine percent of her college-student survey respondents previously owned less than ten percent of the songs they downloaded, and about sixty-nine percent owned less than a quarter. See Jay Rep. at 4, 21 & Tbl. 7. The court has already held that the Jay Report bears greater indicia of reliability than the Fader Report. Moreover, under either analysis, the instant matter is distinguishable from Sony because the Supreme Court determined in Sony that time-shifting represented the principal, rather than an occasional use of VCRs. See Sony, 464 U.S. at 421, 104 S.Ct. 774. 15. Defendant argues that, if space-shifting is deemed a fair use, the staple article of commerce doctrine precludes liability for contributory or vicarious infringement. Under Sony, the copyright holder cannot extend his monopoly to products “capable of substantial noninfringing uses.” Sony, 464 U.S. at 442, 104 S.Ct. 774. Defendant fails to show that space-shifting constitutes a commercially significant use of Napster. Indeed, the most credible explanation for the exponential growth of traffic to the website is the vast array of free MP3 files offered by other users — not the ability of each individual to space-shift music she already owns. Thus, even if space-shifting is a fair use, it is not substantial enough to preclude liability under the staple article of commerce doctrine. See Cable/Home Communication Corp. v. Network Prods., Inc., 902 F.2d 829, 846 (11th Cir.1990) (affirming finding of contributory infringement where defendant primarily promoted pirate computer chips and other devices capablé of descrambling pay-TV broadcasts as infringement aids); A&M Records v. General Audio Video Cassettes, Inc., 948 F.Supp. 1449, 1456 (C.D.Cal.1996) (rejecting Sony defense because counterfeiting was chief purpose of time-loaded cassettes that defendant sold). 16. This court also declines to apply the staple article of commerce doctrine because, as paragraphs (D)(6) and (E)(2) of the legal conclusions explain, Napster exercises ongoing control over its service. In Sony, the defendant’s participation did not extend past manufacturing and selling the VCRs: “[t]he only contact between Sony and the users of the Beta-max ... occurred at the moment of sale.” Sony, 464 U.S. at 438, 104 S.Ct. 774. Here, in contrast, Napster, Inc. maintains and supervises an integrated system that users must access to upload or download files. Courts have distinguished the protection Sony offers to the manufacture and sale of a device from scenarios in which the defendant continues to exercise control over the device’s use. See General Audio Video, 948 F.Supp. at 1456-57 (finding Sony doctrine inapplicable to seller of blank tapes who “acted as a contact between his customers and suppliers of other material necessary for counterfeiting”); RCA Records v. All-Fast Sys., Inc,, 594 F.Supp. 335, 339 (S.D.N.Y.1984) (holding that defendant in position to control cassette-copying machine could not invoke Sony); see also Columbia Pictures Indus., Inc. v. Aveco, Inc., 800 F.2d 59, 62 & n. 3 (3d Cir.1986) (holding that business which rented rooms where public viewed copyrighted video cassettes engaged in contributory infringement, even when it was not source of cassettes). Napster, Inc.’s facilitation of unauthorized file-sharing smacks of the contributory infringement in these cases, rather than the legitimate conduct of the VCR manufacturers. Given defendant’s control over the service, as opposed to mere manufacturing or selling, the existence of a potentially unobjectionable use like space-shifting does not defeat plaintiffs’ claims. 17. Nor do other potential non-infringing uses of Napster preclude contributory or vicarious liability. Defendant claims that it engages in the authorized promotion of independent artists, ninety-eight percent of whom are not represented by the record company plaintiffs. See Def.’s Opp. at 10 (citing, inter alia, Krause Dec. ¶ 16), 27. However, the New Artist Program may not represent a substantial or commercially significant aspect of Napster. The evidence suggests that defendant initially promoted the availability of songs by major stars, as opposed to “page after page of unknown artists.” See 1 Frackman Dec., Exh. C (Parker Dep.) at 104:16-105:10, Exh. 235. Its purported mission of distributing music by artists unable to obtain record-label representation appears to have been developed later. Other facts point to the conclusion that the New Artists Program was an afterthought, not a major aspect of the Napster business plan. Former CEO Eileen Richardson claimed in her deposition that she told the press Napster is not about known artists like Madonna. But, tellingly, discovery related to downloads by Napster executives reveals that Richardson’s own computer contained about five Madonna files obtained using Napster. See 1 Frack-man Dec., Exh. A (Richardson Dep.) at 238:2-240:25. Defendant did not launch the website aspect of its New Artist Program until after plaintiffs filed suit, and as recently as July 2000, bona fide new artists constituted a very small percentage of music available on Napster. See Krause Dec. ¶ 9, Exh. A; Olkin Reply Dec. ¶¶ 3-5; Hausman Reply Dec. ¶¶ 3-6. In any event, Napster’s primary role of facilitating the unauthorized copying and distribution established artists’ songs renders Sony inapplicable. See General Audio Video, 948 F.Supp. at 1456-57; RCA Records, 594 F.Supp. at 339. 18. Plaintiffs do not object to all of the supposedly non-infringing uses of Napster. They do not seek an injunction covering chat rooms or message boards, the New Artist Program or any distribution authorized by rights holders. See Pl.’s Reply at 19. Nor do they seek to enjoin applications unrelated to the music recording industry. See id. Because plaintiffs do not ask the court to shut down such satellite activities, the fact that these activities may be non-infringing does not lessen plaintiffs’ likelihood of success. The court therefore finds that plaintiffs have established a reasonable probability of proving third-party infringement. D. Contributory Copyright Infringement 1. Once they have shown direct infringement by Napster users, plaintiffs must demonstrate a likelihood of success on their contributory infringement claim. A contributory infringer is “one who, with knowledge of the infringing activity, induces, causes or materially contributes to the infringing conduct of another.” Gershwin Publ’g Corp. v. Columbia Artists Management, Inc., 443 F.2d 1159, 1162 (2d Cir.1971); see Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259, 264 (9th Cir.1996). Courts do not require actual knowledge; rather, a defendant incurs contributory copyright liability if he has reason to know of the third party’s direct infringement. See Cable/Home Communication Corp., 902 F.2d at 846; Sega Enter. Ltd. v. MAPHIA, 948 F.Supp. 923, 933 (N.D.Cal.1996) (“Sega II”). 2. Plaintiffs present convincing evidence that Napster executives actually knew about and sought to protect use of the service to transfer illegal MP3 files. For example, a document authored by co-founder Sean Parker mentions the need to remain ignorant of users’ real names and IP addresses “since they are exchanging pirated music.” 1 Frackman Dec., Exh. C (Parker Dep.) at 160:1-162:14, Exh. 254 at SF00100 (emphasis added). The same document states that, in bargaining with the RIAA, defendant will benefit from the fact that “we are not just making pirated music available but also pushing demand.” Id. at 160:1-162:14, Exh. 254 at SF00102 (emphasis added). These admissions suggest that facilitating the unauthorized exchange of copyrighted music was a central part of Napster, Inc.’s business strategy from the inception. Plaintiff also demonstrate that defendant had actual notice of direct infringement because the RIAA informed it of more than 12,000 infringing files. See Creighton 12/3/99 Dec., Exh. D. Although Napster, Inc. purportedly terminated the users offering these files, the songs are still available using the Napster service, as are the copyrighted works which the record company plaintiffs identified in Schedules A and B of their complaint. See Creighton Supp. Dec. ¶¶ 3-4. 3. The law does not require actual knowledge of specific acts of infringement. See Gershwin, 443