Perfect 10, Inc. v. Visa International Service, Ass'n

494 F.3d 788 (2007)

PERFECT 10, INC., Plaintiff-Appellant,
v.
VISA INTERNATIONAL SERVICE, ASSOCIATION; First Data Corporation; Cardservice International, Inc.; Humboldt Bank; Mastercard International, Inc., Defendants-Appellees.

No. 05-15170.

United States Court of Appeals, Ninth Circuit.

Argued and Submitted December 4, 2006. Filed July 3, 2007.

*789 *790 *791 *792 Howard E. King (argued) and Stephen D. Rothschild, King, Holmes, Paterno & Berliner, LLP, Los Angeles, California, for the plaintiff-appellant.

Jeffrey N. Mausner, Berman, Mausner & Resser, Los Angeles, California, for the plaintiff-appellant.

Daniel J. Cooper, Los Angeles, California, for the plaintiff-appellant.

Andrew P. Bridges (argued), John C. Nishi, Winston & Strawn LLP, San Francisco, California, for defendant-appellee Mastercard International Incorporated.

Mark T. Jansen, Nancy L. Tompkins, Anthony J. Malutta, Townsend and Townsend and Crew LLP, San Francisco, California, for defendant-appellee Visa International Service Association.

Robert A. Van Nest, Michael H. Page, R. James Slaughter, Keker & Van Nest, LLP, San Francisco, California, for defendants-appellees First Data Corp., Cardservice International, Inc., and Humboldt Bank.

Before: STEPHEN REINHARDT, ALEX KOZINSKI, and MILAN D. SMITH, JR., Circuit Judges.

Opinion by Judge MILAN D. SMITH, JR.; Dissent by Judge KOZINSKI.

MILAN D. SMITH, JR., Circuit Judge:

Perfect 10, Inc. (Perfect 10) sued Visa International Service Association, Master-Card International Inc., and several affiliated banks and data processing services (collectively, the Defendants), alleging secondary liability under federal copyright and trademark law and liability under California statutory and common law. It sued because Defendants continue to process credit card payments to websites that infringe Perfect 10's intellectual property rights after being notified by Perfect 10 of infringement by those websites. The district court dismissed all causes of action under Federal Rule of Civil Procedure 12(b)(6) for failure to state a claim upon *793 which relief can be granted. We affirm the decision of the district court.

FACTS AND PRIOR PROCEEDINGS

Perfect 10 publishes the magazine "PERFECT10" and operates the subscription website www.perfect10.com., both of which "feature tasteful copyrighted images of the world's most beautiful natural models." Appellant's Opening Brief at 1. Perfect 10 claims copyrights in the photographs published in its magazine and on its website, federal registration of the "PERFECT 10" trademark and blanket publicity rights for many of the models appearing in the photographs. Perfect 10 alleges that numerous websites based in several countries have stolen its proprietary images, altered them, and illegally offered them for sale online.

Instead of suing the direct infringers in this case, Perfect 10 sued Defendants, financial institutions that process certain credit card payments to the allegedly infringing websites. The Visa and Master-Card entities are associations of member banks that issue credit cards to consumers, automatically process payments to merchants authorized to accept their cards, and provide information to the interested parties necessary to settle the resulting debits and credits. Defendants collect fees for their services in these transactions. Perfect 10 alleges that it sent Defendants repeated notices specifically identifying infringing websites and informing Defendants that some of their consumers use their payment cards to purchase infringing images. Defendants admit receiving some of these notices, but they took no action in response to the notices after receiving them.

Perfect 10 separately alleges that it formerly had a merchant account with defendant First Data Corporation (FDC) but that in the Spring of 2001 FDC terminated the account. FDC's stated reason for the termination is that the percentage of Perfect 10's customers who later disputed the charges attributed to them (the chargeback rate) exceeded contractual limits. Perfect 10 claims these chargeback rates were temporarily and substantially inflated because Perfect 10 was the "victim of hackers who were subsequently investigated by the Secret Service." Appellant's Opening Brief at 13. Perfect 10 claims that FDC was aware of this and was also aware that Perfect 10's chargeback rate dropped to within association limits once the hacking ceased, but that FDC nevertheless placed Perfect 10 on an industry-wide "black list" of terminated accounts.

Perfect 10 filed suit against Defendants on January 28, 2004 alleging contributory and vicarious copyright and trademark infringement as well as violations of California laws proscribing unfair competition and false advertising, violation of the statutory and common law right of publicity, libel, and intentional interference with prospective economic advantage. Defendants moved to dismiss the initial complaint under FRCP 12(b)(6). The district court granted the motion, dismissing the libel and intentional interference claims with prejudice but granting leave to amend the remaining claims. In its first amended complaint, Perfect 10 essentially repeated the allegations in its original complaint concerning the surviving causes of action and Defendants again moved to dismiss under FRCP 12(b)(6). The district court granted the Defendants' second motion in full, dismissing all remaining causes of action with prejudice. Perfect 10 appealed to this court.

JURISDICTION

The district court had original jurisdiction over the copyright and trademark claims pursuant to 28 U.S.C. §§ 1331 and *794 1338 and supplemental jurisdiction over the related state law claims pursuant to 28 U.S.C. § 1367. This court has appellate jurisdiction pursuant to 28 U.S.C. § 1291.

STANDARDS OF REVIEW

We review de novo the district court's dismissal for failure to state a claim upon which relief can be granted pursuant to FRCP 12(b)(6). Rodriguez v. Panayiotou, 314 F.3d 979, 983 (9th Cir.2002). On appeal, "we take all of the allegations of material fact stated in the complaint as true and construe them in the light most favorable to the nonmoving party. A complaint should not be dismissed unless it appears beyond doubt that plaintiff can prove no set of facts in support of his claim which would entitle him to relief." Id. (internal citations omitted).

Although a plaintiff's allegations are generally taken as true, the court need not accept conclusory allegations of law or unwarranted inferences, and dismissal is required if the facts are insufficient to support a cognizable claim. City of Arcadia v. U.S. Envtl. Prot. Agency, 411 F.3d 1103, 1106 n. 3 (9th Cir.2005); see also Pena v. Gardner, 976 F.2d 469, 471-72 (9th Cir.1992). The court may also affirm on any ground supported by the record even if the district court did not consider the issue. Fields v. Legacy Health Sys., 413 F.3d 943, 958 n. 13 (9th Cir.2005); ARC Ecology v. United States Dep't of the Air Force, 411 F.3d 1092, 1096 (9th Cir. 2005).

We review de novo the district court's interpretation of state law. Rodriguez, 314 F.3d at 983.

DISCUSSION

SECONDARY LIABILITY UNDER FEDERAL COPYRIGHT AND TRADEMARK LAW

A. Secondary Liability for Copyright Infringement

Perfect 10 alleges that numerous websites based in several countries—and their paying customers—have directly infringed its rights under the Copyright Act, 17 U.S.C. § 101, et seq.[1] In the present suit, however, Perfect 10 has sued Defendants, not the direct infringers, claiming contributory and vicarious copyright infringement because Defendants process credit card charges incurred by customers to acquire the infringing images.

We evaluate Perfect 10's claims with an awareness that credit cards serve as the primary engine of electronic commerce and that Congress has determined it to be the "policy of the United States—(1) to promote the continued development of the Internet and other interactive computer services and other interactive media [and] (2) to preserve the vibrant and competitive free market that presently exists for the Internet and other interactive computer services, unfettered by Federal or State regulation." 47 U.S.C. §§ 230(b)(1), (2).[2]

1. Contributory Copyright Infringement

Contributory copyright infringement is a form of secondary liability with *795 roots in the tort-law concepts of enterprise liability and imputed intent. See Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259, 264 (9th Cir.1996); Perfect 10, Inc. v. Amazon.com, Inc. et al., 487 F.3d 701 (9th Cir.2007). This court and the United States Supreme Court (Supreme Court) have announced various formulations of the same basic test for such liability. We have found that a defendant is a contributory infringer if it (1) has knowledge of a third party's infringing activity, and (2) "induces, causes, or materially contributes to the infringing conduct." Ellison v. Robertson, 357 F.3d 1072, 1076 (9th Cir.2004) (citing Gershwin Publ'g Corp. v. Columbia Artists Mgmt., Inc., 443 F.2d 1159, 1162 (2d Cir.1971)). In an Internet context, we have found contributory liability when the defendant "engages in personal conduct that encourages or assists the infringement." A & M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1019 (9th Cir.2001) (internal citations omitted). In Metro-Goldwyn-Mayer Studios, Inc. v. Grokster, Ltd., the Supreme Court adopted from patent law the concept of "inducement" and found that "[o]ne infringes contributorily by intentionally inducing or encouraging direct infringement." 545 U.S. 913, 930, 125 S. Ct. 2764, 162 L. Ed. 2d 781 (2005).[3] Most recently, in a case also brought by Perfect 10, we found that "an actor may be contributorily liable [under Grokster] for intentionally encouraging direct infringement if the actor knowingly takes steps that are substantially certain to result in such direct infringement." Amazon.com, 487 F.3d at 727.

We understand these several criteria to be non-contradictory variations on the same basic test, i.e., that one contributorily infringes when he (1) has knowledge of another's infringement and (2) either (a) materially contributes to or (b) induces that infringement. Viewed in isolation, the language of the tests described is quite broad, but when one reviews the details of the actual "cases and controversies" before the relevant court in each of the testdefining cases and the actual holdings in those cases, it is clear that the factual circumstances in this case are not analogous. To find that Defendants' activities fall within the scope of such tests would require a radical and inappropriate expansion of existing principles of secondary liability and would violate the public policy of the United States.

a. Knowledge of the Infringing Activity

Because we find that Perfect 10 has not pled facts sufficient to establish that Defendants induce or materially contribute to the infringing activity, Perfect 10's contributory copyright infringement claim fails and we need not address the Defendants' knowledge of the infringing activity.[4]

*796 b. Material Contribution, Inducement, or Causation

To state a claim of contributory infringement, Perfect 10 must allege facts showing that Defendants induce, cause, or materially contribute to the infringing conduct. See, e.g., Ellison, 357 F.3d at 1076. Three key cases found defendants contributorily liable under this standard: Fonovisa, 76 F.3d 259; Napster, 239 F.3d 1004; and Grokster, 545 U.S. 913, 125 S. Ct. 2764, 162 L. Ed. 2d 781. In Fonovisa, we held a swap meet operator contributorily liable for the sale of pirated works at the swap meet. In Napster, we held the operator of an electronic file sharing system liable when users of that system employed it to exchange massive quantities of copyrighted music. In Grokster, the Supreme Court found liability for the substantially similar act of distributing software that enabled exchange of copyrighted music on a peer-to-peer, rather than a centralized basis.[5] Perfect 10 argues that by continuing to process credit card payments to the infringing websites despite having knowledge of ongoing infringement, Defendants induce, enable and contribute to the infringing activity in the same way the defendants did in Fonovisa, Napster and Grokster. We disagree.

1. Material Contribution

The credit card companies cannot be said to materially contribute to the infringement in this case because they have no direct connection to that infringement. Here, the infringement rests on the reproduction, alteration, display and distribution of Perfect 10's images over the Internet. Perfect 10 has not alleged that any infringing material passes over Defendants' payment networks or through their payment processing systems, or that Defendants' systems are used to alter or display the infringing images. In Fonovisa, the infringing material was physically located in and traded at the defendant's market. Here, it is not. Nor are Defendants' systems used to locate the infringing images. The search engines in Amazon.com provided links to specific infringing images, and the services in Napster and Grokster allowed users to locate and obtain infringing material. Here, in contrast, the services provided by the credit card companies do not help locate and are not used to distribute the infringing images. While Perfect 10 has alleged that Defendants make it easier for websites to profit from this infringing activity, the issue here is reproduction, alteration, display and distribution, which can occur without payment. Even if infringing images were not paid for, there would still be infringement. See Napster, 239 F.3d at 1014 (Napster users infringed the distribution right by uploading file names to the search index for others to copy, despite the fact that *797 no money changed hands in the transaction).

Our analysis is fully consistent with this court's recent decision in Perfect 10 v. Amazon.com, where we found that "Google could be held contributorily liable if it had knowledge that infringing Perfect 10 images were available using its search engine, could take simple measures to prevent further damage to Perfect 10's copyrighted works, and failed to take such steps." 487 F.3d at 729. The dissent claims this statement applies squarely to Defendants if we just substitute "payment systems" for "search engine." Dissent at 811. But this is only true if search engines and payment systems are equivalents for these purposes, and they are not. The salient distinction is that Google's search engine itself assists in the distribution of infringing content to Internet users, while Defendants' payment systems do not. The Amazon.com court noted that "Google substantially assists websites to distribute their infringing copies to a worldwide market and assists a worldwide audience of users to access infringing materials." Id. Defendants do not provide such a service. They in no way assist or enable Internet users to locate infringing material, and they do not distribute it. They do, as alleged, make infringement more profitable, and people are generally more inclined to engage in an activity when it is financially profitable. However, there is an additional step in the causal chain: Google may materially contribute to infringement by making it fast and easy for third parties to locate and distribute infringing material, whereas Defendants make it easier for infringement to be profitable, which tends to increase financial incentives to infringe, which in turn tends to increase infringement.[6]

The dissent disagrees with our reading of Amazon.com and charges us with wishful thinking, dissent at 811, and with "draw[ing] a series of ephemeral distinctions," dissent at 825. We respectfully disagree and assert that our construction of the relevant statutes and case law is completely consistent with existing federal law, is firmly grounded in both commercial and technical reality and conforms to the public policy of the United States. Helping users to locate an image might substantially assist users to download infringing images, but processing payments does not. If users couldn't pay for images with credit cards, infringement could continue on a large scale because other viable funding mechanisms are available. For example, a website might decide to allow users to download some images for free and to make its profits from advertising, or it might develop other payment mechanisms that do not depend on the credit card companies.[7] In either case, the unlicensed use of Perfect 10's copyrighted images would still be infringement.[8] We acknowledge that Defendants' *798 payment systems make it easier for such an infringement to be profitable, and that they therefore have the effect of increasing such infringement, but because infringement of Perfect 10's copyrights can occur without using Defendants' payment system, we hold that payment processing by the Defendants as alleged in Perfect 10's First Amended Complaint does not constitute a "material contribution" under the test for contributory infringement of copyrights.[9]

Our holding is also fully consistent with and supported by this court's previous holdings in Fonovisa and Napster. While there are some limited similarities between the factual scenarios in Fonovisa and Napster and the facts in this case, the differences in those scenarios are substantial, and, in our view, dispositive. In Fonovisa, we held a flea market proprietor liable as a contributory infringer when it provided the facilities for and benefitted from the sale of pirated works. 76 F.3d 259. The court found that the primary infringers and the swap meet were engaged in a mutual enterprise of infringement and observed that "it would be difficult for the infringing activity to take place in the massive quantities alleged without the support services provided by the swap meet. These services include, among other things, the provision of space, utilities, parking, advertising, plumbing, and customers." 76 F.3d at 264. But the swap meet owner did more to encourage the enterprise. In 1991, the Fresno County Sheriff raided the swap meet and seized 38,000 counterfeit recordings. Id. at 261. The Sheriff sent a letter to the swap meet operator the following year notifying it that counterfeit sales continued and reminding it that it had agreed to provide the Sheriff with identifying information from each vendor, but had failed to do so. Id. The Fonovisa court found liability because the swap meet operator knowingly provided the "site and facilities" for the infringing activity. Id. at 264.

In Napster, this court found the designer and distributor of a software program liable for contributory infringement. 239 F.3d 1004. Napster was a file-sharing *799 program which, while capable of non-infringing use, was expressly engineered to enable the easy exchange of pirated music and was widely so used. See Napster, 239 F.3d at 1020 n. 5 (quoting document authored by Napster co-founder which mentioned "the need to remain ignorant of users' real names and IP addresses `since they are exchanging pirated music'"). Citing the Fonovisa standard, the Napster court found that Napster materially contributes to the users' direct infringement by knowingly providing the "site and facilities" for that infringement. 239 F.3d at 1022.

Seeking to draw an analogy to Fonovisa and, by extension, Napster, Perfect 10 pleads that Defendants materially contribute to the infringement by offering services that allow it to happen on a larger scale than would otherwise be possible. Specifically, because the swap meet in Fonovisa created a commercial environment which allowed the frequency of that infringement to increase, and the Napster program increased the frequency of infringement by making it easy, Perfect 10 argues that the Defendants have made available a payment system that allows third-party infringement to be profitable, and, consequently, more widespread than it otherwise might be. This analogy fails.

The swap meet operator in Fonovisa and the administrators of the Napster and Grokster programs increased the level of infringement by providing a centralized place, whether physical or virtual, where infringing works could be collected, sorted, found, and bought, sold, or exchanged.[10] The provision of parking lots, plumbing and other accoutrements in Fonovisa was significant only because this was part of providing the environment and market for counterfeit recording sales to thrive.

Defendants, in contrast, do no such thing. While Perfect 10 has alleged that it is easy to locate images that infringe its copyrights, the Defendants' payment systems do not cause this. Perfect 10's images are easy to locate because of the very nature of the Internet—the website format, software allowing for the easy alteration of images, high-speed connections allowing for the rapid transfer of high-resolution image files, and perhaps most importantly, powerful search engines that can aggregate and display those images in a useful and efficient manner, without charge, and with astounding speed. Defendants play no role in any of these functions.

Perfect 10 asserts otherwise by arguing for an extremely broad conception of the term "site and facilities" that bears no relationship to the holdings in the actual "cases and controversies" decided in Fonovisa and Napster. Taken literally, Perfect 10's theory appears to include any tangible or intangible component related to any transaction in which infringing material is bought and sold. But Fonovisa and Napster do not require or lend themselves to such a construction. The actual display, location, and distribution of infringing images in this case occurs on websites that organize, display, and transmit information over the wires and wireless instruments that make up the Internet. The websites are the "site" of the infringement, not Defendants' payment networks. Defendants do not create, operate, *800 advertise, or otherwise promote these websites. They do not operate the servers on which they reside. Unlike the Napster (and Grokster) defendants, they do not provide users the tools to locate infringing material, nor does any infringing material ever reside on or pass through any network or computer Defendants operate.[11] Defendants merely provide a method of payment, not a "site" or "facility" of infringement. Any conception of "site and facilities" that encompasses Defendants would also include a number of peripherally-involved third parties, such as computer display companies, storage device companies, and software companies that make the software necessary to alter and view the pictures and even utility companies that provide electricity to the Internet.

Perfect 10 seeks to side-step this reality by alleging that Defendants are still contributory infringers because they could refuse to process payments to the infringing websites and thereby undermine their commercial viability.[12] Even though we must take this factual allegation as true, that Defendants have the power to undermine the commercial viability of infringement does not demonstrate that the Defendants materially contribute to that infringement. As previously noted, the direct infringement here is the reproduction, alteration, display and distribution of Perfect 10's images over the Internet. Perfect 10 has not alleged that any infringing material passes over Defendants' payment networks or through their payment processing systems, or that Defendants designed or promoted their payment systems as a means to infringe. While Perfect 10 has alleged that Defendants make it easier for websites to profit from this infringing activity, the infringement stems from the failure to obtain a license to distribute, not the processing of payments.

2. Inducement

In Grokster, the Supreme Court applied the patent law concept of "inducement" to a claim of contributory infringement against a file-sharing program. 545 U.S. 913, 125 S. Ct. 2764, 162 L. Ed. 2d 781. The court found that "one who distributes a device with the object of promoting its use to infringe copyright, as shown by clear expression or other affirmative steps taken to foster infringement, is liable for the resulting acts of infringement by third parties." Id. at 936-37, 125 S. Ct. 2764. Perfect 10 claims that Grokster is analogous because Defendants induce customers to use their cards to purchase goods and services, and are therefore guilty of specifically inducing infringement if the cards are used to purchase images from sites that have content stolen from Perfect 10. This is mistaken. Because Perfect 10 alleges no "affirmative steps taken to foster infringement" and no facts suggesting that Defendants promoted their payment system as a means to infringe, its claim is premised on a fundamental misreading of *801 Grokster that would render the concept of "inducement" virtually meaningless.

The Grokster court announced that the standard for inducement liability is providing a service "with the object of promoting its use to infringe copyright." Id. "[M]ere knowledge of infringing potential or actual infringing uses would not be enough here to subject [a defendant] to liability." Id. at 937, 125 S. Ct. 2764. Instead, inducement "premises liability on purposeful, culpable expression and conduct, and thus does nothing to compromise legitimate commerce or discourage innovation having a lawful promise." Id. Moreover, to establish inducement liability, it is crucial to establish that the distributors "communicated an inducing message to their . . . users," the classic example of which is an "advertisement or solicitation that broadcasts a message designed to stimulate others to commit violations." Id. The Grokster court summarized the "inducement" rule as follows:

In sum, where an article is good for nothing else but infringement, there is no legitimate public interest in its unlicensed availability, and there is no injustice in presuming or imputing an intent to infringe. Conversely, the doctrine absolves the equivocal conduct of selling an item with substantial lawful as well as unlawful uses, and limits liability to instances of more acute fault than the mere understanding that some of one's products will be misused. It leaves breathing room for innovation and a vigorous commerce.

545 U.S. at 932-33, 125 S. Ct. 2764 (internal citations and quotation marks omitted).

Perfect 10 has not alleged that any of these standards are met or that any of these considerations are present here. Defendants do, of course, market their credit cards as a means to pay for goods and services, online and elsewhere. But it does not follow that Defendants affirmatively promote each product that their cards are used to purchase. The software systems in Napster and Grokster were engineered, disseminated, and promoted explicitly for the purpose of facilitating piracy of copyrighted music and reducing legitimate sales of such music to that extent. Most Napster and Grokster users understood this and primarily used those systems to purloin copyrighted music. Further, the Grokster operators explicitly targeted then-current users of the Napster program by sending them ads for its OpenNap program. Id. at 925-26, 125 S. Ct. 2764. In contrast, Perfect 10 does not allege that Defendants created or promote their payment systems as a means to break laws. Perfect 10 simply alleges that Defendants generally promote their cards and payment systems but points to no "clear expression" or "affirmative acts" with any specific intent to foster infringement.

The Amazon.com court recognized this distinction and applied it in a matter fully consistent with our analysis in this case. While the Amazon.com court did not bifurcate its analysis of contributory liability into "material contribution" liability and "inducement" liability, it did recognize that contributory liability "may be predicated on actively encouraging (or inducing) infringement through specific acts." Amazon.com, 487 F.3d at 726 (quoting Grokster, 545 U.S. at 942, 125 S. Ct. 2764 (Ginsburg, J., concurring)). It also found that Google could be held contributorily liable if it has "actual knowledge that specific infringing material is available using its system, and can take simple measures to prevent further damage," but does not. Id. at 728 (internal citations and quotation marks omitted). While this test is read more naturally as a test for "material contribution" than as a test for "inducement," *802 under an "inducement" analysis Defendants are not within its scope. As discussed above, Perfect 10 has not alleged any "specific acts" intended to encourage or induce infringement. And moreover, Defendants are distinguishable under the Amazon.com test because, unlike Google, infringing material is not "available using [their] system" of payment processing. Id. That system does not "facilitate access to websites," id.; infringers do not use it to copy, alter, distribute or display infringing material; and consumers do not use it to locate, view or download the infringing images. Rather, all parties involved simply use Defendants' system to process payments for that infringing material.

Finally, we must take as true the allegations that Defendants lend their names and logos to the offending websites and continue to allow their cards to be used to purchase infringing images despite actual knowledge of the infringement—and perhaps even bending their association rules to do so. But we do not and need not, on this factual basis, take as true that Defendants "induce" consumers to buy pirated content with their cards. "Inducement" is a legal determination, and dismissal may not be avoided by characterizing a legal determination as a factual one. We must determine whether the facts as pled constitute a "clear expression" of a specific intent to foster infringement, and, for the reasons above noted, we hold that they do not.

2. Vicarious Copyright Infringement

Vicarious infringement is a concept related to, but distinct from, contributory infringement. Whereas contributory infringement is based on tort-law principles of enterprise liability and imputed intent, vicarious infringement's roots lie in the agency principles of respondeat superior. See Fonovisa, 76 F.3d at 261-62. To state a claim for vicarious copyright infringement, a plaintiff must allege that the defendant has (1) the right and ability to supervise[13] the infringing conduct and (2) a direct financial interest in the infringing activity. Ellison, 357 F.3d at 1078; Napster, 239 F.3d at 1022 (citations omitted). The Supreme Court has recently offered (in dictum) an alternate formulation of the test: "One . . . infringes vicariously by profiting from direct infringement while declining to exercise a right to stop or limit it." Grokster, 545 U.S. at 930, 125 S. Ct. 2764 (internal citations omitted). Perfect 10 alleges that Defendants have the right and ability to control the content of the infringing websites by refusing to process credit card payments to the websites, enforcing their own rules and regulations, or both. We hold that Defendants' conduct alleged in Perfect 10's first amended complaint fails to state a claim for vicarious copyright infringement.

a. Right and Ability to Supervise the Infringing Activity

In order to join a Defendant's payment network, merchants and member banks must agree to follow that Defendant's rules and regulations. These rules, among other things, prohibit member banks from providing services to merchants engaging in certain illegal activities and require the members and member banks to investigate merchants suspected of engaging in such illegal activity and to terminate their participation in the payment network if certain illegal activity is *803 found. Perfect 10 has alleged that certain websites are infringing Perfect 10's copyrights and that Perfect 10 sent notices of this alleged infringement to Defendants. Accordingly, Perfect 10 has adequately pled that (1) infringement of Perfect 10's copyrights was occurring, (2) Defendants were aware of the infringement, and (3) on this basis, Defendants could have stopped processing credit card payments to the infringing websites. These allegations are not, however, sufficient to establish vicarious liability because even with all reasonable inferences drawn in Perfect 10's favor, Perfect 10's allegations of fact cannot support a finding that Defendants have the right and ability to control the infringing activity.

In reasoning closely analogous to the present case, the Amazon.com court held that Google was not vicariously liable for third-party infringement that its search engine facilitates. In so holding, the court found that Google's ability to control its own index, search results, and webpages does not give Google the right to control the infringing acts of third parties even though that ability would allow Google to affect those infringing acts to some degree. Amazon.com, 487 F.3d at 730-32. Moreover, and even more importantly, the Amazon.com court rejected a vicarious liability claim based on Google's policies with sponsored advertisers, which state that it reserves "the right to monitor and terminate partnerships with entities that violate others' copyright[s]." Id. at 730 (alteration in original). The court found that

Google's right to terminate an AdSense partnership does not give Google the right to stop direct infringement by third-party websites. An infringing third-party website can continue to reproduce, display, and distribute its infringing copies of Perfect 10 images after its participation in the AdSense program has ended.

Id. This reasoning is equally applicable to the Defendants in this case. Just like Google, Defendants could likely take certain steps that may have the indirect effect of reducing infringing activity on the Internet at large. However, neither Google nor Defendants has any ability to directly control that activity, and the mere ability to withdraw a financial "carrot" does not create the "stick" of "right and ability to control" that vicarious infringement requires. A finding of vicarious liability here, under the theories advocated by the dissent, would also require a finding that Google is vicariously liable for infringement—a conflict we need not create, and radical step we do not take.

Perfect 10 argues that this court's decision in Napster compels a contrary result. The Napster court found a likelihood of vicarious liability because Napster "had the right and ability to police its system and failed to exercise that right to prevent the exchange of copyrighted material." 239 F.3d at 1023. The Napster program created a forum for the exchange of digital music files and the program administrators had the ability to block certain users from accessing that forum to upload or download such files. As pled by Perfect 10, Defendants also provide a system that allows the business of infringement for profit to operate on a larger scale than it otherwise might, and Defendants have the ability to deny users access to that payment system.

This argument fails. The Napster program's involvement with—and hence its "policing" power over—the infringement was much more intimate and directly intertwined with it than Defendants' payment systems are. Napster provided users with the tools to enable the easy reproduction and distribution of the actual infringing content and to readily search *804 out and identify infringing material. Defendants' payment systems do not. Napster also had the right and ability to block user access to its program and thereby deprive particular users of access to their forum and use of their location and distribution tools. Defendants can block access to their payment system, but they cannot themselves block access to the Internet, to any particular websites, or to search engines enabling the location of such websites. Defendants are involved with the payment resulting from violations of the distribution right, but have no direct role in the actual reproduction, alteration, or distribution of the infringing images.[14] They cannot take away the tools the offending websites use to reproduce, alter, and distribute the infringing images over the Internet. They can only take away the means the websites currently use to sell them.[15]

Perfect 10 offers two counter-arguments. Perfect 10 first claims that Defendants' rules and regulations permit them to require member merchants to cease illegal activity—presumably including copyright infringement—as a condition to their continuing right to receive credit card payments from the relevant Defendant entities. Perfect 10 argues that these contractual terms effectively give Defendants contractual control over the content of their merchants' websites, and that contractual control over content is sufficient to establish the "right and ability" to control that content for purposes of vicarious liability. In the sense that economic considerations can influence behavior, these contractual rules and regulations do give Defendants some measure of control over the offending websites since it is reasonable to believe that fear of losing access to credit card payment processing services would be a sufficient incentive for at least some website operators to comply with a content-based suggestion from Defendants. But the ability to exert financial pressure does not give Defendants the right or ability to control the actual infringing activity at issue in this case. Defendants have no absolute right[16] to stop that activity—they cannot stop websites *805 from reproducing, altering, or distributing infringing images. Rather, the credit card companies are analogous to Google, which we held was not liable for vicarious copyright infringement even though search engines could effectively cause a website to disappear by removing it from their search results, and reserve the right to do so. Like Google, the credit card companies "cannot stop any of the third-party websites from reproducing, displaying, and distributing unauthorized copies of Perfect 10's images because that infringing conduct takes place on the third-party websites." Amazon.com, 487 F.3d at 731. Defendants can only refuse to process credit card payments to the offending merchant within their payment network, or they can threaten to do so if the merchant does not comply with a request to alter content. While either option would likely have some indirect effect on the infringing activity, as we discuss at greater length in our analysis of the Grokster "stop or limit" standard below, so might any number of actions by any number of actors. For vicarious liability to attach, however, the defendant must have the right and ability to supervise and control the infringement, not just affect it, and Defendants do not have this right or ability.

Perfect 10 relies heavily on the reasoning of Fonovisa and Napster to support this argument, but that reliance is misplaced. The swap meet operator in Fonovisa and the software operator in Napster both had the right to remove individual infringers from the very place the infringement was happening. Defendants, like the defendants in Amazon.com, have no such right. As already discussed, Defendants cannot take away the software the offending websites use to copy, alter, and distribute the infringing images, cannot remove those websites from the Internet, and cannot themselves block the distribution of those images over the Internet. Defendants can refuse to process credit card payments for those images, but while this refusal would reduce the number of those sales, that reduction is the result of indirect economic pressure rather than an affirmative exercise of contractual rights.[17]

Perfect 10 also argues that were infringing websites barred from accepting the Defendants' credit cards, it would be impossible for an online website selling adult images to compete and operate at a profit.[18] While we must take this allegation as *806 true, it still fails to state a claim because it conflates the power to stop profiteering with the right and ability to control infringement. Perfect 10's allegations do not establish that Defendants have the authority to prevent theft or alteration of the copyrighted images, remove infringing material from these websites or prevent its distribution over the Internet. Rather, they merely state that this infringing activity could not be profitable without access to Defendants' credit card payment systems. The alleged infringement does not turn on the payment; it turns on the reproduction, alteration and distribution of the images, which Defendants do not do, and which occurs over networks Defendants do not control.

The Supreme Court's recent decision in Grokster does not undermine the validity of this distinction. As we held in Amazon.com, 487 F.3d at 728-31, Grokster does not stand for the proposition that just because the services provided by a company help an infringing enterprise generate revenue, that company is necessarily vicariously liable for that infringement. Numerous services are required for the third party infringers referred to by Perfect 10 to operate. In addition to the necessity of creating and maintaining a website, numerous hardware manufacturers must produce the computer on which the website physically sits; a software engineer must create the program that copies and alters the stolen images; technical support companies must fix any hardware and software problems; utility companies must provide the electricity that makes all these different related operations run, etc. All these services are essential to make the businesses described viable, they all profit to some degree from those businesses, and by withholding their services, they could impair—perhaps even destroy—the commercial viability of those business. But that does not mean, and Grokster by no means holds, that they are all potentially liable as vicarious infringers. Even though they have the "right" to refuse their services, and hence the literal power to "stop or limit" the infringement, they, like Defendants, do not exercise sufficient control over the actual infringing activity for vicarious liability to attach.

b. Obvious and Direct Financial Interest in the Infringing Activity

Because Perfect 10 has failed to show that Defendants have the right and ability to control the alleged infringing conduct, it has not pled a viable claim of vicarious liability. Accordingly, we need not reach the issue of direct financial interest.

B. Secondary Liability for Trademark Infringement

The tests for secondary trademark infringement are even more difficult to satisfy than those required to find secondary copyright infringement. See Sony Corp. v. Universal City Studios, 464 U.S. 417, 439 n. 19, 104 S. Ct. 774, 78 L. Ed. 2d 574 (1984); Fonovisa, 76 F.3d at 265 (noting that "trademark infringement liability is more narrowly circumscribed than copyright infringement"). While the tests for such infringement are somewhat different in the trademark context, Perfect 10's factual allegations in support of these claims are essentially identical to those alleged in Perfect 10's copyright claims, and they fail to state a claim for similar reasons.

*807 1. Contributory Trademark Infringement

To be liable for contributory trademark infringement, a defendant must have (1) "intentionally induced" the primary infringer to infringe, or (2) continued to supply an infringing product to an infringer with knowledge that the infringer is mislabeling the particular product supplied. Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 855, 102 S. Ct. 2182, 72 L. Ed. 2d 606 (1982). When the alleged direct infringer supplies a service rather than a product, under the second prong of this test, the court must "consider the extent of control exercised by the defendant over the third party's means of infringement." Lockheed Martin Corp. v. Network Solutions, Inc., 194 F.3d 980, 984 (9th Cir.1999). For liability to attach, there must be "[d]irect control and monitoring of the instrumentality used by a third party to infringe the plaintiff's mark." Id.

Perfect 10 has failed to plead a viable claim under either prong of Inwood Labs — and, by extension, Lockheed Martin. First, it has not pled facts showing that Defendants "intentionally induced" infringement of Perfect 10's mark. Perfect 10 has alleged that Defendants are providing critical support to websites that are using the PERFECT 10 mark in a manner that is likely to cause the public to believe that they are authorized by Perfect 10. Its factual allegations in support of this claim are identical to those it made in support of its copyright claims. These allegations, however, cite no affirmative acts by Defendants suggesting that third parties infringe Perfect 10's mark, much less induce them to do so.

Second, Perfect 10 has failed to allege facts sufficient to show "[d]irect control and monitoring of the instrumentality used by a third party to infringe the plaintiff's mark." Lockheed Martin, 194 F.3d at 984. Perfect 10 claims that the "product" or "instrumentality" at issue here is the credit card payment network through which Defendants process payments for infringing material. Appellant's Opening Brief at 39. As discussed at length above, this network is not the instrument used to infringe Perfect 10's trademarks; that infringement occurs without any involvement of Defendants and their payment systems. Perfect 10 has not alleged that Defendants have the power to remove infringing material from these websites or directly stop their distribution over the Internet. At most, Perfect 10 alleges that Defendants can choose to stop processing payments to these websites, and that this refusal might have the practical effect of stopping or reducing the infringing activity. This, without more, does not constitute "direct control." See Lockheed Martin, 194 F.3d at 985 ("While the landlord of a flea market might reasonably be expected to monitor the merchandise sold on his premises, [defendant] NSI cannot reasonably be expected to monitor the Internet.") (citation omitted).

2. Vicarious Trademark Infringement

Vicarious liability for trademark infringement requires "a finding that the defendant and the infringer have an apparent or actual partnership, have authority to bind one another in transactions with third parties or exercise joint ownership or control over the infringing product." Hard Rock Café Licensing Corp. v. Concession Servs., Inc., 955 F.2d 1143, 1150 (7th Cir.1992) (internal quotations omitted), followed by Symantec Corp. v. CD Micro, Inc., 286 F. Supp. 2d 1265, 1275 (D.Or.2003).

Perfect 10 argues that Defendants are liable as follows: "Defendants and the *808 Stolen Content Websites are in a symbiotic financial partnership pursuant to which the websites operate their businesses according to defendants' rules and regulations and defendants share the profits, transaction by transaction." Appellant's Opening Brief at 40. For the same reasons that this relationship does not establish "right and ability to control" for copyright purposes, neither does it establish such a "symbiotic" relationship or "joint ownership or control" for trademark purposes. Defendants process payments to these websites and collect their usual processing fees, nothing more.

Perfect 10 further argues that "Defendants' acceptance of a charge binds the Stolen Content Website to provide the infringing images to third parties." Appellant's Opening Brief at 40. Even if legally relevant, Perfect 10's allegation is legally incorrect. It is the websites' contracts with the consumers that bind the websites to provide the infringing images, not the websites' relationship with Defendants.[19] The websites' contracts with Defendants are merely a means of settling the resulting debits and credits among the websites and the relevant consumers. We hold that Perfect 10 fails to state a claim for vicarious trademark infringement.

CALIFORNIA STATUTORY AND COMMON LAW CLAIMS

In addition to its federal copyright and trademark claims, Perfect 10 pled causes of action for unfair competition, false advertising, violation of the right of publicity, libel, and intentional interference with economic relations. We hold that the district court properly dismissed all these claims with prejudice.

A. California State Law Claims of Unfair Competition hand False Advertising

Defendants do not dispute Perfect 10's claims that the websites themselves are potentially violating California state and common law prohibiting unfair competition and false advertising. See Cal. Bus. & Prof.Code §§ 17200, et seq., and 17500, et seq. Defendants do, however, argue that Emery v. Visa International Service Association, 95 Cal. App. 4th 952, 116 Cal. Rptr. 2d 25 (2002), precludes liability for Defendants in this case, both under secondary liability and aiding and abetting theories. Defendants are correct on both counts.

In Emery, a California appellate court affirmed a grant of summary judgment in favor of Visa, finding that Visa did not exercise the requisite control over merchants marketing foreign lottery tickets to impose secondary liability under the state's unfair competition or false advertising laws. Id. at 959-964, 116 Cal. Rptr. 2d 25. Emery found that an "unfair practices claim under section 17200 cannot be predicated on vicarious liability. . . . A defendant's liability must be based on his personal participation in the unlawful practices and unbridled control over the practices that are found to violate section 17200 or 17500." Id. at 960, 116 Cal. Rptr. 2d 25 (internal citations omitted). Because "Visa itself played no part in preparing or sending any `statement' that might be construed as untrue or misleading under the unfair business practices *809 statutes," it could not be liable for unfair competition. Id. at 964, 116 Cal. Rptr. 2d 25. The false advertising claim also necessarily failed because "even if Visa allowed the merchants to use its logo, trade name, or trademark, it would not be liable for false advertising. There is no duty to investigate the truth of statements made by others." Id. (citations omitted). Emery is dispositive of Perfect 10's claims that the Defendants are secondarily liable under California unfair competition and false advertising laws and the district court properly dismissed them.[20]

In an attempt to avoid the impact of Emery, Perfect 10 argues on appeal that it alleged aiding and abetting theories of liability in its complaints, and further, that the district court improperly dismissed these civil claims under a criminal standard of aiding and abetting. Perfect 10 fails to establish a viable claim on these theories as well. The only authority offered by Perfect 10 in support of such liability is an opinion which is now uncitable in California: Schulz v. Neovi Data Corporation, 28 Cal. Rptr. 3d 46 (Cal. App.4th Dist.2005), superceded by 32 Cal. Rptr. 3d 758, 117 P.3d 475 (Cal.2005), cause transferred by 56 Cal. Rptr. 3d 471, 154 P.3d 998 (Cal.2007), transferred to, 152 Cal. App. 4th 86, 60 Cal. Rptr. 3d 810 (4th Dist. Jun 15, 2007).

Furthermore, even under the standards announced in the superceding Schulz opinion, Defendants would not be liable. The Schulz court found a credit card company potentially liable for its role in facilitating an illegal online lottery because that company "went far beyond merely processing credit cards." 152 Cal.App.4th at 95, 60 Cal. Rptr. 3d 810. In support, the court cited specific statements from a company representative in which he "personally assured" an agent of the website that the defendant company "did not have any problem with the operation of the [illegal] lottery site" and had a "stronger stomach" than other payment processors. Id. Perfect 10 alleges no similar conduct here— Defendants merely process credit card payments.

B. Aiding and Abetting the Websites' Violations of Perfect 10's Right of Publicity

Perfect 10 alleges that Defendants aided and abetted the websites' violations of Perfect 10's rights of publicity, acquired by assignment from its models, in violation of Cal. Civil Code § 3344 and the common law right of publicity. This aiding and abetting claim fails for the same reasons as the aiding and abetting claims under unfair competition and false advertising. Even if such liability is possible under California law—a proposition for which Perfect 10 has provided no clear authority—Defendants lack sufficient control or personal involvement in the infringing activities to be so liable. See Schulz, 152 Cal.App.4th at 93-94, 60 Cal. Rptr. 3d 810; Emery, 95 Cal.App.4th at 962-63, 116 Cal. Rptr. 2d 25.

C. Libel and Intentional Interference with Prospective Economic Advantage

The district court dismissed Perfect 10's claims of libel and intentional *810 interference with prospective economic advantage with prejudice on multiple grounds. We affirm on the ground that both are time-barred. Under California law, a libel claim must be filed within one year of publication of the allegedly libelous statement, Cal. Civ. Proc. § 340(c), and an intentional interference claim must be filed within two years of the underlying harmful act, Cal. Civ. Proc. § 339. Perfect 10 claims the same underlying wrongful act as the basis for both claims: its placement on the industry "black list" in the Spring of 2001. However, Perfect 10 failed to file suit until January 2004—well beyond the statute of limitations applicable to each claim—and has failed to show any possible exception under either statute. Those claims are time-barred.

CONCLUSION

We decline to create any of the radical new theories of liability advocated by Perfect 10 and the dissent and we affirm the district court's dismissal with prejudice of all causes of action in Perfect 10's complaint for failure to state a claim upon which relief can be granted.

AFFIRMED.

KOZINSKI, Circuit Judge, dissenting for the most part:[1]

Federal law gives copyright owners the exclusive right to "distribute copies [of their works] . . . to the public by sale." 17 U.S.C. § 106(3). Plaintiff alleges that certain third parties it refers to as the "Stolen Content Websites" unlawfully copy its protected images and sell them to the public, using defendants' payment systems as financial intermediaries. According to plaintiff, the Stolen Content Websites "maintain no physical presence in the United States in order to evade criminal and civil liability for their illegal conduct." First Am. Compl. at 8 ¶ 26. Plaintiff also claims that "Defendants do not enforce their own rules against [the] Stolen Content Websites because Defendants do not want to lose the substantial revenues and profits they receive from the websites." Id. at 10 ¶ 35. Plaintiff has repeatedly notified defendants that they are abetting the sale of stolen merchandise by "knowingly providing crucial transactional support services for the sale of millions of stolen photos and film clips worth billions of dollars," id. at 1 ¶ 5, but to no avail. Frustrated in its effort to protect the rights Congress has given it, plaintiff turns to the federal courts for redress. We should not slam the courthouse door in its face.

Accepting the truth of plaintiff's allegations, as we must on a motion to dismiss, the credit cards[2] are easily liable for indirect copyright infringement: They knowingly provide a financial bridge between buyers and sellers of pirated works, enabling them to consummate infringing *811 transactions, while making a profit on every sale. If such active participation in infringing conduct does not amount to indirect infringement, it's hard to imagine what would.[3] By straining to absolve defendants of liability, the majority leaves our law in disarray.

Contributory Infringement

We have long held that a defendant is liable for contributory infringement if it "materially contributes to the infringing conduct." A & M Records, Inc. v. Napster, Inc., 239 F.3d 1004, 1019 (9th Cir. 2001) (internal quotations omitted) (citing Gershwin Publ'g Corp. v. Columbia Artists Mgmt., Inc., 443 F.2d 1159, 1162 (2d Cir.1971)). Our recent opinion in Perfect 10, Inc. v. Amazon.com, Inc., 487 F.3d 701 (9th Cir.2007), canvasses the caselaw in this area and concludes that Google "could be held contributorily liable if it had knowledge that infringing Perfect 10 images were available using its search engine, could take simple measures to prevent further damage to Perfect 10's copyrighted works, and failed to take such steps." Amazon, 487 F.3d at 729. Substitute "payment systems" for "search engine" in this sentence, and it describes defendants here: If a consumer wishes to buy an infringing image from one of the Stolen Content Websites, he can do so by using Visa or MasterCard, just as he can use Google to find the infringing images in the first place. My colleagues engage in wishful thinking when they claim that "Google's search engine itself assists in the distribution of infringing content to Internet users, while Defendants' payment systems do not" and that "[h]elping users to locate an image might substantially assist users to download infringing images, but processing payments does not." Maj. op. at 797, 797.[4]

The majority struggles to distinguish Amazon by positing an "additional step in the causal chain" between defendants' activities and the infringing conduct. Id. at 797. According to the majority, "Google may materially contribute to infringement by making it fast and easy for third parties to locate and distribute infringing material, whereas Defendants make it easier for infringement to be profitable, which tends to increase financial incentives to infringe, which in turn tends to increase infringement." Id. The majority is mistaken; there is no "additional step." Defendants participate in every credit card sale of pirated images; the images are delivered to the buyer only after defendants approve the transaction and process the payment. *812 This is not just an economic incentive for infringement; it's an essential step in the infringement process.

In any event, I don't see why it matters whether there is an "additional step." Materiality turns on how significantly the activity helps infringement, not on whether it's characterized as one step or two steps removed from it. The majority recognizes that "Defendants make it easier for websites to profit from this infringing activity," maj. op. at 796; that defendants' conduct "tends to increase infringement," id. at 797; that defendants "have the effect of increasing . . . infringement," id. at 798; that "Defendants have the power to undermine the commercial viability of" the Stolen Content Websites and that they "make it easier for websites to profit from this infringing activity," id. at 800; that "Defendants could likely take certain steps that may have the indirect effect of reducing infringing activity on the Internet," id. at 803-04; and that defendants could "reduce the number of those [infringing] sales," id. at 805. Taking the majority at its word, it sounds like defendants are providing very significant help to the direct infringers.

My colleagues recognize, as they must, that helping consumers locate infringing content can constitute contributory infringement,[5] but they consign the means of payment to secondary status. Maj. op. at 799 ("Defendants merely provide a method of payment. . . ."); id. at 802 ("[A]ll parties involved simply use Defendants' system to process payments for that infringing material."); id. at 804 ("They can only take away the means the websites currently use to sell [the infringing images]."); id. at 805 ("Defendants can only refuse to process credit card payments to the offending merchant within their payment network. . . ."). But why is locating infringing images more central to infringement than paying for them? If infringing images can't be found, there can be no infringement; but if infringing images can't be paid for, there can be no infringement either. Location services and payment services are equally central to infringement; the majority's contrary assertion is supported largely by disparaging use of "merely," "simply" and "only." See also id. at 803 ("[M]ere ability to withdraw a financial'carrot' does not create the `stick' of `right and ability to control'. . . . ").[6]

The majority dismisses the significance of credit cards by arguing that "infringement could continue on a large scale [without them] because other viable funding mechanisms are available." Maj. op. at 797.[7] Of course, the same could be said *813 about Google. As the majority admits, if Google were unwilling or unable to serve up infringing images, consumers could use Yahoo!, Ask.com, Microsoft Live Search, A9.com or AltaVista instead. Id. at 797-98 n. 8. Even if none of these were available, consumers could still locate websites with infringing images through e-mails from friends, messages on discussion forums, tips via online chat, "typo-squatting," peer-to-peer networking using BitTorrent or eDonkey, offline and online advertisements (see p. 820 infra), disreputable search engines hosted on servers in far-off jurisdictions or even old-fashioned word of mouth. The majority's claim that search engines "could effectively cause a website to disappear by removing it from their search results," maj. op. at 805, is quite a stretch.

If the test for contributory infringement really were whether "infringement could continue on a large scale[without the aid of the defendant] because other viable . . . mechanisms are available," Amazon should have absolved Google of liability because of the availability of such obvious alternatives. But Amazon held that Google could be liable for contributory infringement because it "substantially assists" users in finding infringing materials; the existence of other means of infringement was not even considered because no case has suggested this to be a relevant consideration. The majority's "other viable . . . mechanisms" test conflicts with Amazon, Napster, Grokster and every other material assistance case that I know of.

The majority does even worse when it tries to describe the "other viable funding mechanisms" that could serve as alternatives to credit cards: According to the majority, the Stolen Content Websites "might . . . make [their] profits from advertising" or "might develop other payment mechanisms that do not depend on the credit card companies." Maj. op. at 797 (emphasis added). This shows that my colleagues have a healthy imagination but contravenes our responsibilities, the most fundamental of which is that we must work with the facts the parties presented below, not invent new facts on appeal. Defendants have presented no evidence that the pirates could survive without credit cards, nor could they, as the case is still at the motion to dismiss stage. Even if speculation as to what the Stolen Content Websites "might" do were admissible evidence, which I seriously doubt, we must still wait for one of the parties to present it, not conjure it up ourselves.[8] At the pleadings stage, we must accept plaintiff's allegations that credit cards are indispensable to the operation of the Stolen Content Websites, and that these websites would be forced out of business without them. See First Am. Compl. at 2 ¶ 7 ("Stolen Content Websites cannot exist without the knowledge and direct participation of [defendants]."); id. at 10 ¶ 35 ("[T]he Stolen Content Websites would be eradicated."). If my colleagues can't justify their result without contradicting plaintiff's allegations, this is a pretty good hint that they're *814 wrong. See also p. 812-13 n. 7 supra; pp. 818 n. 15, 800-02 infra.

The majority's attempt to distinguish location services from payment services by trying to show that there are viable alternatives for the latter but not the former cuts entirely against them. As plaintiff alleges, and experience tells us, there are numerous ways of locating infringing images on the Internet, but there are no adequate substitutes for credit cards when it comes to paying for them. A few consumers might use checks or money orders to pay for infringing images, but this would be far more cumbersome, time-consuming and risky than using credit cards. See pp. 799-800 & n. 14 infra. If it mattered whether search engines or credit cards are more important to peddling infringing content on the Internet, the cards would win hands down.

But it doesn't matter. Material assistance turns on whether the activity in question "substantially assists" infringement. Amazon, 487 F.3d at 729. It makes no difference whether the primary infringers might do without it by finding a workaround, which is why the majority can cite no case supporting its analysis. We presume that primary infringers have good reasons for selecting a particular means to infringe, and that other ways to do so will be more costly, more cumbersome and less efficient. Moreover, infringement can always be carried out by other means; if the existence of alternatives were a defense to contributory infringement then there could never be a case of contributory infringement based on material assistance. The majority makes some very new—and very bad—law here.

The majority also makes a slightly different argument: "While Perfect 10 has alleged that Defendants make it easier for websites to profit from this infringing activity, the issue here is reproduction, alteration, display and distribution, which can occur without payment. Even if infringing images were not paid for, there would still be infringement." Maj. op. at 796-97. What the majority seems to be arguing here is that helping an infringer get paid cannot materially assist infringement because the actual process of infringement— "reproduction, alteration, display and distribution"—does not include payment. There are two problems with this argument. The first is that the Stolen Content Websites are alleged to infringe plaintiff's right of distribution "by sale." 17 U.S.C. § 106(3). It's not possible to distribute by sale without receiving compensation, so payment is in fact part of the infringement process. Second, this argument runs head-on into Amazon, where we held that helping to find infringing images materially assists infringement, even though locating infringing images also isn't "reproduction, alteration, display [or] distribution." To be sure, locating images, like paying for them, makes it a lot easier to infringe, but neither is intrinsic to the infringement process, as the majority conceives it.

Nor can today's opinion be squared with Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259 (9th Cir.1996). In Fonovisa, defendant allowed known infringers to sell pirated works from stalls at its swap meet. We found material assistance based on the fact that "it would [have been] difficult for the infringing activity to take place in the massive quantities alleged without the support services provided by the swap meet." 76 F.3d at 264. The pivotal role played by the swap meet in Fonovisa is played by the credit cards in cyberspace, in that they make "massive quantities" of infringement possible that would otherwise be impossible. Indeed, the assistance provided here is far more material than in Fonovisa. A pirate kicked out of a swap meet could still peddle his illicit wares through newspaper *815 ads or by word of mouth, but you can't do business at all on the Internet without credit cards. Plaintiff thus plausibly alleges that the "Stolen Content Websites would be eradicated" if defendants withdrew their support. First Am. Compl. at 10 ¶ 35.

The majority rejects Fonovisa by pointing out that the swap meet there provided a "centralized place" for the infringement to take place, maj. op. at 799, whereas defendants here "have no direct connection to [the] infringement," id. at 796.[9] But material assistance does not depend on physical contact with the infringing activity. If you lend money to a drug dealer knowing he will use it to finance a drug deal, you materially assist the transaction, even if you never see the drugs. Or, if you knowingly drive a principal to the scene of the crime, you provide material assistance, even if nothing happens during the ride. See United States v. Lopez, 482 F.3d 1067, 1076-79 (9th Cir.2007). Material assistance turns on whether the conduct assists infringement in a significant way, not on pedantic factual distinctions unrelated to how much the activity facilitates infringement.

Sure, a marketplace for pirated works (as in Fonovisa) or an index for such works (as in Napster and Grokster) is important to infringement, but so is a means of getting paid. Defendants are directly involved in every infringing transaction where payment is made by credit card, which (according to plaintiff) amounts to virtually every sale of pirated works. First Am. Compl. at 9 ¶ 35. Credit cards don't provide some tangential service that marginally affects sales; they are the financial lifeblood of the Stolen Content Websites.

The majority's concern that imposing liability on defendants here would implicate vast numbers of other actors who provide incidental services to infringers, maj. op. at 800, is unfounded. Line-drawing is always a bit tricky, but courts have shown themselves adept at dealing with it from time out of mind, in resolving such issues as proximate causation and reasonable suspicion. Contributory infringement requires material assistance to the infringing activity, and those the majority worries about would doubtless be absolved of liability because their contribution to the infringing activity is insufficiently material.

Courts have, in fact, had no difficulty in distinguishing those who are materially involved in copyright infringement from those who are not. As Fonovisa explains, two lines of cases developed in the first part of the last century: the absentee landlord cases and the dance hall cases. The first line involved landlords who "lacked knowledge of the infringing acts of [their] tenant[s] and who exercised no control over the leased premises." Fonovisa, 76 F.3d at 262. These were held not liable for the infringement committed by tenants on the premises. See, e.g., Deutsch v. Arnold, 98 F.2d 686, 688 (2d Cir.1938). In the second line of cases, "the operator of an entertainment venue was held liable for infringing performances when the operator (1) could control the premises and (2) obtained a direct financial benefit from the audience, who paid to enjoy the infringing performance." 76 F.3d at 262 (citing Buck *816 v. Jewell-LaSalle Realty Co., 283 U.S. 191, 51 S. Ct. 410, 75 L. Ed. 971 (1931), and Dreamland Ball Room, Inc. v. Shapiro, Bernstein & Co., 36 F.2d 354 (7th Cir. 1929)).[10]

These cases show that courts are able to forestall the majority's parade of horribles. But our case does not present a close or difficult question: Defendants here are alleged to provide an essential service to infringers, a service that enables infringement on a massive scale. Defendants know about the infringements; they profit from them; they are intimately and causally involved in a vast number of infringing transactions that could not be consummated if they refused to process the payments; they have ready means to stop the infringements. Were we to rule for plaintiff, as we should, I have every confidence that future courts would be able to distinguish this case when and if they are confronted with lawsuits against utility companies, software vendors and others who provide incidental services to infringers.

Vicarious Infringement

A party "infringes vicariously by profiting from direct infringement while declining to exercise a right to stop or limit it." Amazon, 487 F.3d at 729 (quoting Grokster, 545 U.S. at 930, 125 S. Ct. 2764) (internal quotation marks omitted).[11] There is no doubt that defendants profit from the infringing activity of the Stolen Content Websites; after all, they take a cut of virtually every sale of pirated material. First Am. Compl. at 4 ¶ 13, 7 ¶ 25. The majority does not dispute this point so I need not belabor it. Maj. op. at 806-07.

Defendants here also have a right to stop or limit the infringing activity, a right they have refused to exercise. As the majority recognizes, "Perfect 10 . . . claims that Defendants' rules and regulations permit them to require member merchants to cease illegal activity—presumably including copyright infringement—as a condition to their continuing right to receive credit card payments from the relevant Defendant entities." Maj. op. at 804.[12] Assuming *817 the truth of this allegation,[13] the cards have the authority, given to them by contract, to force the Stolen Content Websites to remove infringing images from their inventory as a condition for using defendants' payment systems. If the merchants comply, their websites stop peddling stolen content and so infringement is stopped or limited. If they don't comply, defendants have the right—and under copyright law the duty—to kick the pirates off their payment networks, forcing them to find other means of getting paid or go out of business. In that case, too, infringement is stopped or limited. The swap meet in Fonovisa was held vicariously liable precisely because it did not force the pirates to stop infringing or leave; there is no reason to treat defendants here differently.

That the pirates might find some other way of doing business is of no consequence; our cases make this perfectly clear. It didn't matter in Fonovisa that the infringers there could have continued their illegal sales by mail order or by hawking their unlawful merchandise on street corners. Nor did it matter in Napster or Grokster that the direct infringers might find some other means of illegally sharing their protected content with others. Indeed, there is no case involving secondary infringement, going back to the dance hall cases of the last century, where the secondary infringer's refusal to do business with the direct infringer could have stopped infringement altogether and forever. Yet, courts have presumed that removing the particular means of infringement challenged in each case would make direct infringement more difficult and thereby diminish the scale of infringing activity.

Here, the Stolen Content Websites have chosen credit cards as a form of payment, and for good reason. Credit cards are ubiquitous and permit the transfer of funds electronically in a matter of seconds. Consumers need not wait days or weeks for a check to reach its destination and clear before gaining access to the salacious pictures they crave. Consumers also know that, if goods bought by credit card are not delivered, the cards will likely reverse the transaction.[14] Credit cards thus act as informal escrow agents, effectively guaranteeing that their merchants will deliver the goods. Blocking the ability to accept credit cards would be a heavy blow to the Stolen Content Websites because cards are "overwhelmingly the primary way by which customers pay to view Stolen Content Websites." First Am. Compl. at 9 ¶ 35. Even if the pirates could find an alternative way of plying their illegal trade, being denied their preferred means of doing business *818 would sharply curtail their unlawful activities.

The majority toils to resist this obvious conclusion but its arguments are not persuasive. For example, it makes no difference that defendants control only the means of payment, not the mechanics of transferring the material. Maj. op. at 802, 805, 806. In a commercial environment, distribution and payment are (to use a quaint anachronism) like love and marriage—you can't have one without the other. If cards don't process payment, pirates don't deliver booty. The credit cards, in fact, control distribution of the infringing material.

The majority also disparages defendants' ability to control the Stolen Content Websites as just "financial pressure" which doesn't give them an "absolute right to stop [the infringing] activity—they cannot stop websites from reproducing, altering, or distributing infringing images." Id. at 805 (footnote omitted).[15] But we have never required an "absolute right to stop [the infringing] activity" as a predicate for vicarious liability; it's enough if defendants have the "practical ability" to do so. Amazon, 487 F.3d at 729, 731. While proclaiming its fidelity to Amazon, maj. op. at 796, 803, the majority jettisons Amazon's "practical ability" standard and substitutes its own "absolute right to stop" standard. Id. at 804.[16]

It's perfectly clear that the cards do have the "practical ability" to force websites that display their logos and use their payment systems to remove unlawful merchandise. As the majority admits, "Defendants can . . . refuse to process credit card payments to the offending merchant within their payment network, or they can threaten to do so if the merchant does not comply with a request to alter content." Maj. op. at 805 (disparaging "only" omitted). Commercial websites are dependent on credit cards as a form of payment, and the Stolen Content Websites are uniquely so, as virtually all of their illicit sales are paid for by card. First Am. Compl. at 9 ¶ 35. A threat by credit card companies to withdraw use of their payment systems couldn't be ignored. After all, how many consumers would be willing to send a check or money order to a far-off jurisdiction in the hope that days or weeks later they will be allowed to download some saucy pictures? If the Stolen Content Websites cannot get paid for their unlawful *819 products, or if payment is made more difficult or cumbersome, this will dramatically affect their operations. Some may lose customers who are unwilling to use alternative forms of payment;[17] others may go out of business; still others may remove the infringing content from their websites. Even the majority admits that "fear of losing access to credit card payment processing services would be a sufficient incentive for at least some website operators to comply with a content-based suggestion from Defendants." Maj. op. at 804-05.[18] As a consequence, infringing activity would be "stop[ped] or limit[ed]." See Amazon, 487 F.3d at 729.

The majority also reads the complaint for less than it's worth by "understand[ing]" plaintiff to allege "that the `Stolen Content Websites' could not continue to exist as websites offering images for sale online should defendants withdraw their services, not [to allege] that the websites would completely vanish or that infringement by these sites in all its forms would necessarily cease." Maj. op. at 805-06 n. 18. But plaintiff expressly alleges what the majority "understand[s]" it not to allege, namely that the sites "cannot exist" without defendants, First Am. Compl. at 2 ¶ 7, and that "the Stolen Content Websites would be eradicated" if they could not use credit cards, id. at 9-10 ¶ 35. It is horn-book law that we must construe complaints liberally on a motion to dismiss. See Glus v. Brooklyn Eastern Dist. Terminal, 359 U.S. 231, 235, 79 S. Ct. 760, 3 L. Ed. 2d 770 (1959); Conley v. Gibson, 355 U.S. 41, 47-48, 78 S. Ct. 99, 2 L. Ed. 2d 80 (1957). A liberal construction means reading ambiguous provisions in a way that would save the complaint from dismissal, and sometimes even reading between the lines to give plaintiff the benefit of every reasonable inference. I have never heard of reading a complaint liberally by ignoring allegations that are clearly present.

But let's say the majority "understand[s]" plaintiff's allegations correctly: So what? To sustain a vicarious infringement claim, plaintiff need not allege that the Stolen Content Websites "would completely vanish or that infringement by these sites in all its forms would necessarily cease." Maj. op. at 805-06 n. 18. The standard is "stop or limit" the infringing conduct. Amazon, 487 F.3d at 725 (emphasis added) (quoting Grokster, 545 U.S. at 930, 125 S. Ct. 2764). And my colleagues admit that plaintiff has alleged that "at least some website operators [would] comply with a content-based suggestion from Defendants." Maj. op. at 804-05. Q.E.D.

To resolve this case, however, we need not adopt a rule holding all credit cards responsible for all infringing Internet sales because plaintiff has alleged far more than the ordinary credit card/merchant relationship. *820 According to plaintiff, defendants have adopted special rules and practices that apply only to the Stolen Content Websites, and that are designed to make it easier for these websites to ply their illegal trade. First Am. Compl. at 9-11 ¶¶ 33-37. Plaintiff claims that the credit cards have singled out the Stolen Content Websites for preferential treatment because of the unusual and substantial profits they make on such transactions. Read fully and fairly, the complaint alleges that defendants are not merely passive providers of services available on equal terms to legal and illegal businesses alike; they are actually in cahoots with the pirates to prop up their illegal businesses and share their ill-gotten gains. If this is not vicarious infringement, nothing is.

The majority claims that Amazon employs "reasoning closely analogous" to its own, maj. op. at 803, but it is mistaken. Amazon addressed two questions of vicarious infringement, one involving third-party websites whose images are picked up by Google's search engine, the other involving websites that participate in its AdSense program. As to the first, Google could not be vicariously liable because "Perfect 10 ha[d] not shown that Google has contracts with third-party websites that empower Google to stop or limit them from reproducing, displaying, and distributing infringing copies of Perfect 10's images on the Internet." 487 F.3d at 730.[19] In the absence of such a contractual relationship, there could be no vicarious infringement, because Google lacked "the legal right to stop or limit the direct infringement of third-party websites." Id. at 730. Why the majority believes this is in any way analogous, or even remotely instructive, to our situation, where the credit cards do have contracts giving them a right to control what merchants sell on their websites, is a mystery.

Google's relationship with websites that participate in its AdSense program presents a somewhat closer analogy because Google did have contracts that would have allowed it to kick websites out of AdSense for displaying infringing images. But that's as far as the similarity goes: AdSense is an advertising program; Google pays participating merchants to host third-party ads on their websites. This is the cyberspace analogue of renting out space on your land for a billboard. The ads have no effect on the operation of the host websites; users can download infringing content whether or not ads are present. Being excluded from AdSense would thus mean some loss of revenue, but would have no effect on the operation of the business itself. It is therefore far from certain that merchants would be induced to modify their businesses to avoid being excluded from AdSense.[20]

Because plaintiff had not presented proof that any third-party websites would stop infringing if they were threatened *821 with exclusion from AdSense, Amazon concluded that plaintiff there had not met its burden for a preliminary injunction. Our case is presented on a motion to dismiss and plaintiff here need only make allegations. And plaintiff alleges that the infringing websites could not continue doing business at all without the use of credit cards. Amazon's reasoning on this point gives the majority no help.

The majority's attempt to distinguish Napster is equally thin. My colleagues argue that "[t]he Napster program's involvement with . . . the infringement was much more intimate and directly intertwined with it than Defendants' payment systems are." Maj. op. at 803-04. But I don't see how much more "directly intertwined" you can get in a purchase transaction than carrying the payment from buyer to seller. If this were a drug deal, for example, we would never say that the guy entrusted with delivery of the purchase money is less involved in the transaction than the guy who helps find the seller. Both would be held equally culpable.

Thus, the majority's insistence that defendants "cannot themselves block access to the Internet, to any particular websites, or to search engines enabling the location of such websites," maj. op. at 804, is beside the point. Physical control over the infringing activity is one way to stop infringers, but it's certainly not the only way. Withdrawing crucial services, such as financial support, can be just as effective, and sometimes more effective, than technical measures that can often be circumvented.[21]

Finally, the majority dismisses the Supreme Court's opinion in Grokster by suggesting that the Court could not have meant what it said because the standard it announced (and which we adopted in Amazon) would sweep in too many goods and services that contribute to infringing activity. See maj. op. at 806 (listing hardware manufacturers, software engineers, technical support companies and utilities). The majority misreads the Court's opinion. Providing a crucial service to an infringer may give someone the practical ability to stop infringement, but that's only half of what it takes to be a vicarious infringer. The other half is a right, found in contract, to control the infringer's actions. See Amazon, 487 F.3d at 730 (requiring "contracts with [direct infringers] that empower [defendant] to stop or limit them from reproducing, displaying, and distributing infringing copies"). Those third parties the majority worries about could not be held vicariously liable because they lack the legal right to stop the infringement. So far as I know, utilities are provided by public franchise, not by contract, and a utility has no right to stop providing electricity or phone service because it learns that its electrons are being put to illegal use.[22] Computer manufacturers don't usually *822 retain the right to reclaim computers they have sold because they are being used unlawfully. Ditto for software producers and repairmen. Having no contract that authorizes them to stop providing services on account of illegality, these actors do not meet the first prong of the test for vicarious infringement. See p. 799 n. 10 supra.[23]

Trademark Infringement

For precisely the same reasons, I disagree with the majority when it claims that defendants do not contributorily infringe on Perfect 10's trademark because they lack "[d]irect control and monitoring of the instrumentality used by a third party to infringe the plaintiff's mark." Maj. op. at 807 (internal quotation marks omitted). The Lanham Act forbids "use in commerce . . . of a registered mark in connection with the sale, offering for sale, distribution, or advertising of any goods or services on or in connection with which such use is likely to cause confusion." 15 U.S.C. § 1114(1)(a). Plaintiff alleges that the Stolen Content Websites sell images marked with Perfect 10's trademark. First Am. Compl. at 17 ¶ 65. Without defendants' payment systems, the infringers would find it much harder to peddle their infringing goods. Plaintiff thus pled facts sufficient to state a claim for contributory trademark infringement.

The cases on which the majority relies are not to the contrary. Inwood Laboratories, Inc. v. Ives Laboratories, Inc., 456 U.S. 844, 102 S. Ct. 2182, 72 L. Ed. 2d 606 (1982), involved a manufacturer and says nothing of consequence bearing on our situation. Lockheed Martin Corp. v. Network Solutions, Inc., 194 F.3d 980, 984 (9th Cir.1999), turned on the fact that the defendant there, a registrar of Internet domain names, lacked sufficient control "over the third party's means of infringement," because it lacked "control and monitoring of the instrumentality used by a third party to infringe the plaintiff's marks." Id. By contrast, credit cards are directly involved in every infringing transaction; not only do they process the payment for virtually every sale of pirated images by the Stolen Content Websites, they control whether such transactions will go forward. This is more than enough to establish the "control and monitoring" that Lockheed Martin requires for contributory trademark infringement.

As to vicarious trademark infringement, the majority claims that there is neither a symbiotic partnership between the direct infringers and defendants, nor authority to bind one another in transactions with third parties. Maj. op. at 807 (citing Hard Rock Cafe Licensing Corp. v. Concession Servs., Inc., 955 F.2d 1143 (7th Cir.1992)). But plaintiff alleges that the Stolen Content *823 Websites cannot operate without the use of credit cards, First Am. Compl. at 2 ¶ 7, while defendants make huge profits by processing these illegal transactions. See also p. 820 supra. If this is not symbiosis, what is? Likewise, while "the websites' contracts with the consumers . . . bind the websites to provide the infringing images," maj. op. at 808 (emphasis removed), it is defendants' actions that bind the websites to that contract. Only after the credit cards approve and process the payment does the obligation to deliver the stolen content come into existence. The majority thus errs in absolving defendants of vicarious trademark infringement.

State Law Claims

The majority disposes of the state law claims on the same theory as the copyright claims, namely, that defendants are not directly involved in the infringing activity. Maj. op. at 808-09. But defendants are involved, because they provide the means to pay for the infringing content and thus make "massive quantities" of infringement possible. First Am. Compl. at 18 ¶ 73.

The case on which the majority relies, Emery v. Visa International Service Association, 95 Cal. App. 4th 952, 116 Cal. Rptr. 2d 25 (2002), is not on point because, in that case, plaintiff sued only Visa, not the merchant banks that had a direct relationship with the alleged wrongdoer or the consumers. Id. at 956, 962, 116 Cal. Rptr. 2d 25. Plaintiff there also based his theory of liability on advertising letters bearing the credit card logo. Emery held that plaintiff hadn't proven Visa could police the use of its logo in letters peddling an illegal lottery sent by merchants directly to consumers. By contrast, plaintiff here alleges that defendants are knowing participants in thousands of transactions that amount to unfair trade practices and infringe on the right of publicity of the women depicted in the stolen images. I see nothing in Emery that would preclude plaintiff's state law claims, as alleged in the complaint.

* * *

It would certainly be much easier for us if plaintiff were suing the Stolen Content Websites rather than the credit cards. No doubt, they would if they could.[24] But direct infringers are sometimes too ubiquitous, too small or too difficult to find. That's why we have cases such as Fonovisa, Napster, Aimster, Grokster and Amazon. Here, plaintiff alleges that many direct infringers have no physical presence in the United States. They operate from far-off jurisdictions, where lawsuits are difficult to bring and remedies impossible to enforce because the infringers can easily move their operations to servers in other remote jurisdictions.

The weak link in the pirates' nefarious scheme is their need to get paid; for this they must use the services of legitimate financial institutions. If plaintiff's allegations are to be believed, the financial institutions (defendants here) collect billions for sellers of stolen merchandise; in a very real sense, they profit from making piracy possible. I can see no reason they should not be held responsible.

The majority's refrain that imposing liability on defendants here would violate "the public policy of the United States," maj. op. at 795, 798, is equally off base. While the majority correctly identifies that *824 policy as facilitating the development of electronic commerce, id. at 794 n. 2, that solicitude does not extend to commerce in illegal merchandise. I am aware of no policy of the United States to encourage electronic commerce in stolen goods, illegal drugs or child pornography. When it comes to traffic in material that violates the Copyright Act, the policy of the United States is embedded in the FBI warning we see at the start of every lawfully purchased or rented video: Infringers are to be stopped and prosecuted. Preventing financial intermediaries from servicing such shady transactions is entirely consistent with that policy. If Congress believes that this places too heavy a burden on credit cards, it can grant the cards immunity (along with corresponding responsibilities), as it did for ISPs in passing the DMCA.[25]

The majority's solicitude for "credit cards . . . as the primary engine of electronic commerce," and for preserving "the vibrant and competitive free market that presently exists for the Internet," maj. op. at 794, is understandable but misguided. It does not serve the interests of a free market, or a free society, to abet marauders who pilfer the property of law-abiding, tax-paying rights holders, and who turn consumers into recipients of stolen property. Requiring defendants to abide by their own rules, which "strictly prohibit members from servicing illegal businesses," First Am. Compl. at 6 ¶ 20, will hardly impair the operation of a "vibrant and competitive free market," any more than did the recent law prohibiting the use of credit cards for Internet gambling. See 31 U.S.C. § 5364.

Nor does plaintiff seek to hold the credit cards responsible for illegal activities of which they are unaware. Plaintiff claims that it has repeatedly written to defendants, "putting them on notice of more than 240 specifically identified Celebrity Porn Websites with obvious stolen content that they were supporting." First Am. Compl. at 19 ¶ 75. Plaintiff has also sent defendants "[d]eclarations from celebrities [such as Britney Spears, Christina Aguilera, Anna Kournikova and Yasmine Bleeth] stating that they have not authorized the use of their name, likeness, or identity on pornographic websites and that they do not want their images and names so used. . . ." Id. at 19 ¶ 77. Credit cards already have the tools to police the activities of their merchants, which is why we don't see credit card sales of illegal drugs or child pornography. According to plaintiff, "defendants inspect websites and business premises, and obtain and review merchants' bank statements, tax returns, credit reports, and a merchant's other financial information. . . ." Id. at 7 ¶ 26. Plaintiff is not asking for a huge change in the way credit cards do business; they ask only that defendants abide by their own rules and stop doing business with crooks. Granting plaintiff the relief it seeks would not, I am confident, be the end of Capitalism as we know it.

This is an easy case, squarely controlled by our precedent in all material respects. Fairly applying our cases to the facts alleged by Perfect 10, we should reverse the district court and give plaintiff an opportunity to prove its case through discovery and trial. In straining to escape the strictures *825 of our caselaw, the majority draws a series of ephemeral distinctions that are neither required nor permitted; the opinion will prove to be no end of trouble.

NOTES

[1] While Perfect 10's complaint does not clearly specify which of Perfect 10's rights are being infringed, it appears that at least four such rights are potentially at issue: reproduction (17 U.S.C. § 106(1)); derivative works (17 U.S.C. § 106(2)); distribution of copies (17 U.S.C. § 106(3)); and public display (17 U.S.C. § 106(5)).

[2] Congress expressed similar sentiments when it enacted the Digital Millennium Copyright Act (DMCA), 17 U.S.C. § 512, one of the stated purposes of which was to "facilitate the robust development and worldwide expansion of electronic commerce, communications, research, development, and education in the digital age." S. Rep. 105-190, at 1-2 (1998).

[3] In her concurring opinion in Grokster, Justice Ginsburg identified another strand of contributory liability in the Supreme Court's jurisprudence, i.e., liability based on "distributing a product distributees use to infringe copyrights, if the product is not capable of `substantial' or `commercially significant' noninfringing uses." Grokster, 545 U.S. at 942, 125 S. Ct. 2764 (citing Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417, 442, 104 S. Ct. 774, 78 L. Ed. 2d 574 (1984)). Even assuming Defendants offer a "product" for these purposes, Perfect 10 does not claim that the "product" of credit card services is incapable of substantial and commercially significant noninfringing uses.

[4] We note that an anomaly exists regarding the concept of notice in secondary copyright infringement cases outside a FRCP 12(b)(6) context. Congress addressed the issue of notice in the DMCA, which grants a safe harbor against liability to certain Internet service providers, even those with actual knowledge of infringement, if they have not received statutorily-compliant notice. See Perfect 10 v. CCBill LLC, 481 F.3d 751 (9th Cir.2007), amended and superceded, 488 F.3d 1102 (9th Cir.2007); 17 U.S.C. § 512(c)(3). Because Defendants are not "service providers" within the scope of the DMCA, they are not eligible for these safe harbors. The result, under Perfect 10's theories, would therefore be that a service provider with actual knowledge of infringement and the actual ability to remove the infringing material, but which has not received a statutorily compliant notice, is entitled to a safe harbor from liability, while credit card companies with actual knowledge but without the actual ability to remove infringing material, would benefit from no safe harbor. We recognize that the DMCA was not intended to displace the development of secondary liability in the courts; rather, we simply take note of the anomalous result Perfect 10 seeks.

[5] Because the Grokster court focused primarily on an "inducement" theory rather than a "material contribution" theory, our primary discussion of Grokster is located below in the "inducement" section of this opinion.

[6] As discussed in note 11, infra, the dissent's claims that payment processing is "an essential step in the infringement process," dissent at 812, and that "Defendants are directly involved in every infringing transaction where payment is made by credit card," dissent at 815, suggests that the dissent believes that the Defendants are directly infringing when they process these payments.

[7] As discussed more fully in the vicarious infringement section, infra, Perfect 10's factual allegations are not to the contrary.

[8] We recognize that Google is not the only search engine available to Internet users, and that users do not necessarily need Google to locate infringing images. The distinction we draw, however, is not specific to Google; it is between location services and payment services. Because location services lead Internet users directly to infringing images and often display them on the website of the service itself, we find that location services are more important and more essential—indeed, more "material"—to infringement than payment services are.

[9] Our dissenting colleague assures us that we would not jeopardize Internet commerce by finding Defendants liable because he has "every confidence" that this court will simply find that other providers of essential services may contribute to infringement, but not materially so. Dissent at 816. We take little comfort in his assurances because the predicate of our colleague's optimistic view of future judicial refinement of his new world of secondary liability is a large number of expensive and drawn-out pieces of litigation that may, or may not, ever be filed. Meanwhile, what would stop a competitor of a web-site from sending bogus notices to a credit card company claiming infringement by its competitor in the hope of putting a competitor out of business, or, at least, requiring it to spend a great deal of money to clear its name? Threatened with significant potential secondary liability on a variety of fronts under the dissent's proposed expansion of existing secondary liability law, perhaps the credit card companies would soon decline to finance purchases that are more legally risky. They, after all, are as moved by Adam Smith's "invisible hand" as the next set of merchants. If that happened, would First Amendment rights of consumers be trampled? Would Perfect 10 itself be adversely impacted because no credit card company would want to take a chance on becoming secondarily liable?

We similarly take little comfort in the dissent's resurrection of the "dance-hall-owner/absentee-landlord" cases as a source of any principled distinction in this area. Dissent at 815-16. Those tests were developed for a brick-and-mortar world, and, as the Napster and Grokster courts implicitly recognized by paying little attention to them, they do not lend themselves well to application in an electronic commerce context. In deciding this case, we are well-advised to follow the lead of the Supreme Court's and our own court's cases confronting online commerce issues.

[10] In fact, as virtually every interested college student knew—and as the program's creator expressly admitted-the sole purpose of the Napster program was to provide a forum for easy copyright infringement. See Napster, 239 F.3d at 1020 n. 5. Perfect 10 does not contend that Defendants' payment systems were engineered for infringement in this way, and we decline to radically expand Napster's cursory treatment of "material contribution" to cover a credit card payment system that was not so designed.

[11] Moreover, if the processing of payment for an infringing transaction were as central to the infringement as the dissent believes it to be—see, e.g., dissent at 811 (payment processing is "an essential step in the infringement process"), dissent at 815 ("Defendants are directly involved in every infringing transaction where payment is made by credit card")—it is difficult to see why Defendants would be not be direct infringers of the distribution right. Not even Perfect 10 has gone so far as to allege that theory here—Perfect 10 would undoubtedly be quite surprised to learn, after years of litigation attempting to expand the scope of secondary liability, that Defendants are direct infringers after all.

[12] This allegation is considered below under vicarious infringement, but we also address it here in terms of contributory infringement.

[13] Fonovisa essentially viewed "supervision" in this context in terms of the swap meet operator's ability to control the activities of the vendors, 76 F.3d at 262, and Napster essentially viewed it in terms of Napster's ability to police activities of its users, 239 F.3d at 1023.

[14] The same analysis of Defendants' role in any violation of the distribution right under 17 U.S.C. § 106(3), discussed in note 11, supra, is equally applicable here. While the Napster program allowed its operators to block users from violation of the distribution right, Defendants' "policing" power is limited to refusing to process payments resulting from such violations and does not extend to directly stopping the violations themselves.

[15] The conclusion that the Defendants operate outside the scope of the Napster rule is further bolstered by consideration—though as persuasive authority only—of this court's opinion in Metro-Goldwyn-Mayer Studios, Inc. v. Grokster Ltd., 380 F.3d 1154 (9th Cir. 2004), which the Supreme Court vacated on other grounds, 545 U.S. 913, 125 S. Ct. 2764, 162 L. Ed. 2d 781 (2005). In Grokster, we found the defendants not vicariously liable in part because they could not block individual users or remove copyrighted material from the network. Id. at 1165. Similarly, because none of the infringing images resides on or passes through present Defendants' own systems or any systems over which Defendants exercise direct control, Defendants have no ability to actually remove infringing material from the Internet or directly block its distribution. This distinguishes credit card companies from Napster, which could block access to the tools needed for the easy reproduction and distribution of the actual infringing content.

[16] We do not, as the dissent suggests, hold that an absolute right to stop the infringement is a prerequisite for vicarious liability. Dissent at 818-19. Rather, we consider the Defendants' inability to directly control the actual infringing activities of third-party websites—reproduction, alteration, display, and distribution over the Internet, not over Defendants' payment systems—as evidence that they, much like Google, lack the right and ability to control those activities.

[17] We do not hold, as the dissent suggests, that the ability to exert financial pressure is categorically insufficient to establish sufficient control for vicarious liability. We recognize that financial pressure is often very powerful, but it is precisely for this reason that we hesitate to expand the law of vicarious liability to encompass the sort of financial pressure Defendants may exert. The dissent believes that the gravamen of "right and ability to control" is the "practical ability" to limit infringement. Dissent at 818-19. But if this were true, despite the dissent's protestations to the contrary, there are many providers of essential services who could limit infringement by refusing to offer those services. If "practical ability" is the test, it does not matter if software operators, network technicians, or even utility companies do not have a contractual right to affect the websites' content. It is an article of faith of the free market that, subject to certain limited exceptions, one can refuse to deal with anyone for any reason, and by refusing to deal with the offending websites, these providers could limit infringement.

[18] Specifically, Perfect 10 defines "Stolen Content Websites" as "websites . . . that routinely offer for sale to the public stolen [images]," First Am. Compl. at 2, ¶ 6 (emphasis added), and alleges that "Stolen Content Websites cannot exist without the knowledge and direct participation of the financial institutions that process the credit card transactions for such unlawful material," id. at 2, ¶ 7. We do acknowledge that at this procedural stage, Perfect 10 is entitled to all reasonable inferences, but we understand this to be a factual allegation that the "Stolen Content Websites" could not continue to exist as websites offering images for sale online should defendants withdraw their services, not an allegation that the websites would completely vanish or that infringement by these sites in all its forms would necessarily cease.

[19] The dissent claims that no contractual relationship arises between the infringers and consumers until Defendants process a payment. Dissent at 822-23. Even if true as a factual and legal matter—and given the absence of any citation, it is difficult to know whether this is true—this results from a decision of the websites to delay formation of the relationship, not from any requirement Defendants impose on the transaction.

[20] The dissent argues that Emery does not preclude Perfect 10's claims because the only defendant in Emery was Visa International Service Association, whereas Perfect 10 has also sued the member merchant banks who issue cards and process payments from merchants. Dissent at 822-23. This distinction is only relevant if the activities of the member banks constitute personal participation in the infringing activity, and for all the reasons discussed above, those banks are not personally involved in the reproduction, alteration, or distribution of the infringing images. Rather, they merely process payments related to those activities.

[1] I join part C of the "California Statutory and Common Law Claims" section of the opinion, dealing with plaintiff's libel and prospective economic advantage claims.

[2] Throughout this dissent, I refer to defendants collectively as credit card companies or credit cards. In so doing, I am adopting the same simplifying assumptions as the majority. I am aware that Visa and MasterCard don't deal directly with merchants; rather, merchants obtain credit card accounts from banks, which are in turn authorized by Visa or MasterCard to use their respective payment systems. Some of the other defendants are involved in clearing these transactions. For a description of how the system works, see Emery v. Visa Int'l Serv. Ass'n, 95 Cal. App. 4th 952, 956, 116 Cal. Rptr. 2d 25 (2002). It may well be that some of the defendants will be absolved of liability because they have no direct contact with merchants or consumers, but that is a matter to be sorted out after discovery.

[3] As the majority points out, maj. op. at 797 n. 6, 800 n. 11, plaintiff's allegations might also support a theory of direct infringement. See First Am. Compl. at 8 ¶ 30 ("Defendants, jointly with the Stolen Content Websites, are engaged in . . . the willful and systematic infringement of the intellectual property rights of" plaintiff and others). Because plaintiff has not argued this theory on appeal, we have no occasion to address it. But the fact that defendants may also be committing direct infringement does not diminish their responsibility as indirect infringers for providing essential services to buyers and sellers of stolen merchandise. A defendant can be liable for both direct and indirect infringement based on the same conduct. See, e.g., Alcatel USA, Inc. v. DGI Technologies, Inc., 166 F.3d 772, 791 (5th Cir.1999).

[4] Neither Google nor the credit cards here were designed for infringement. The majority tries to distinguish this case from Napster and Metro-Goldwyn-Mayer Studios, Inc. v. Grokster, Ltd., 545 U.S. 913, 125 S. Ct. 2764, 162 L. Ed. 2d 781 (2005), where defendants' services were designed for no other purpose. Maj. op. at 799 n. 10, 801. But Napster and Grokster are not the endpoint of this court's caselaw: Even though Google has many legitimate, noninfringing uses, Amazon held that it would be guilty of contributory infringement if it could modify its service to avoid helping infringers.

[5] Amazon, as well as Napster and Grokster, hold as much.

[6] The majority argues that "[b]ecause location services lead Internet users directly to infringing images, and often display them on the website of the service itself, we find that location services are more important and more essential—indeed, more `material'—to infringement than payment services are." Maj. op. at 797-98 n. 8. Skipping lightly over the fact that we lack the power to "find" anything, the majority admits that payment services are important, essential and material. That location services may — or may not — be more so, is of no consequence; this is not a race where there can be only one winner.

[7] The majority's claim that "Perfect 10's factual allegations are not to the contrary," maj. op. at 797 n. 7, is simply not accurate. Indeed, elsewhere in the opinion, the majority concedes that plaintiff has made "a factual allegation" that the Stolen Content Websites "could not continue to exist as websites offering images for sale online." Id. at 805-06 n. 18. How then can the majority hold here, apparently as a matter of law, that defendants are absolved of liability because "other viable funding mechanisms are available"? Maj. op. at 797. If we accept as true, as the majority says it does, that the Stolen Content Websites will no longer be able to sell their images, how can we hold that they could still do so by developing other (unknown and unsuspected) ways to get paid?

[8] I note in passing that, even if we were to accept the majority's speculations, they would be insufficient. That the Stolen Content Websites "might" change the way they do business or develop alternative payment mechanisms hardly proves that "other viable funding mechanisms are available." Maj. op. at 797 (emphasis added). The majority's prognostication as to what "might" happen in the future leaves open the likelihood that it will not happen, and positively admits that there are no viable alternative payment mechanisms today.

[9] The majority seeks to distinguish Napster and Grokster on similar grounds by arguing that the defendants do not provide the "tools to locate infringing material," id. at 800, and that the infringing material "[n]ever reside[s] on or pass[es] through any network or computer Defendants operate," id.

[10] The majority consigns the dance hall/absentee landlord cases to oblivion by holding that they have no relevance to the Internet. Maj. op. at 798 n. 9. It is true that these cases were developed in a brick and mortar world, but the distinction they draw between those who materially assist infringement (and are therefore liable) and those who are more remotely involved (and are therefore not liable) is equally important—perhaps even more important—in cyberspace than in real space. That Napster and Grokster did not consider these cases is hardly significant. The defendants there were centrally involved in the infringing transactions—indeed, as the majority reminds us, their systems were created solely to promote infringement, maj. op. at 799 n. 10, 801 — and thus there could be no argument that their involvement in the infringing transactions was too peripheral to give rise to a claim of secondary infringement. The Seventh Circuit managed to apply the dance hall cases to the Internet, see In re Aimster Copyright Litig., 334 F.3d 643, 654 (7th Cir.2003), and I'm confident that federal judges west of the Rockies could have figured out how to do the same.

[11] Amazon interprets the "stop or limit" language as requiring "a legal right to stop or limit the allegedly infringing conduct, as well as the practical ability to do so." Amazon, 487 F.3d at 730.

[12] Plaintiff's allegation on this point, as on many others, is very specific:

When MasterCard or Visa learns of a merchant engaged in illegal, fraudulent, or otherwise improper business practices, their own regulations require them to cause member banks to investigate and, depending on the nature of the misconduct, terminate the merchants from the Visa and MasterCard systems. The rules of both associations strictly prohibit members from servicing illegal businesses.

First Am. Compl. at 6 ¶ 20.

[13] In fact, there can be no doubt that it's true. For example, the MasterCard Merchant Rules Manual provides that "[a] Payment Transaction may not be effected for any of the following reasons: . . . to transfer gambling winnings or funds related to chips, currency, or other value usable for gambling that were purchased in connection with gambling; for any illegal purpose or any other purpose deemed by MasterCard to be impermissible." MasterCard International, Merchant Rules Manual § 2.1.11.3(6) (2006) (emphasis added), available at http://www.mastercard.com/ us/wce/PDF/12999—MERC-Entire—Manual. pdf.

[14] Visa's website, for example, explains that "Visa and its card issuers and acquirers have in place an efficient dispute resolution process." Visa USA, Chargebacks & Dispute Resolution, http://www.usa.visa.com/ merchants/operations/chargebacks—dispute— resolution/index.html (last visited March 24, 2007). It also notes that "[c]hargebacks arise for many reasons, primary among which are customer disputes, fraud, processing errors, authorization issues, and non-fulfillment of copy requests." Id. (emphasis added).

[15] The majority tries to take back in a footnote what it says in text by claiming that an "absolute right to stop" is not "a prerequisite" to vicarious liability, but that its absence is "evidence that[defendants], much like Google, lack the right and ability to control those [infringing] activities." Maj. op. at 804 n. 16. Alas, it won't work. If not having an "absolute right to stop" is merely "evidence" that defendants lack sufficient control for vicarious infringement, then this can be offset by other evidence that they do have such control. Conflicts in the evidence must be resolved after discovery and trial, not on a motion to dismiss.

"Practical ability," the standard announced in Amazon, is a capacious concept, far broader than "absolute right to stop." Even if the majority were right that defendants lack the "absolute right to stop" the infringements, plaintiff would be entitled to show that defendants have the "practical ability" to do so. If the majority means what it says in its footnote, then what it says in text is beside the point. In fact, there can be no doubt that the majority means what it says in text, because it upholds dismissal of the complaint on the ground that defendants lack the "absolute right to stop" the infringers; the footnote is merely an unpersuasive attempt to sweep the conflict with Amazon under the rug.

[16] The conflict with Amazon is clearly drawn in footnote 17, where the majority explicitly disavows "practical ability" as the standard for vicarious infringement. Maj. op. at 805 n. 17. The majority is free to disagree with the standard adopted by our caselaw, but it is not free to reject it.

[17] Those customers may take their patronage to plaintiff's website.

[18] The majority disparages this as mere "financial pressure," but I am aware of no prior case holding that the legal right to exercise "financial pressure" over infringing activity is insufficient to support a finding of vicarious infringement. This is a dangerous precedent. We live in a commercial world and economic incentives are often the strongest possible motivators — far stronger than the often empty threat of litigation. As this case demonstrates, litigation can be costly and protracted, and its results uncertain. By contrast, the threat of stopping an essential service can be implemented at once, without hiring an army of lawyers or persuading judges and juries of the rightness of one's cause. In an economy marked by competition, financial pressure which raises costs or diminishes patronage can be a powerful weapon. By holding that the legal right to exercise financial pressure is an insufficient basis for establishing vicarious infringement, my colleagues take a hasty and unwise step in the development of the law.

[19] Amazon also relied on the district court's finding that Google "lacks the practical ability to police the third-party websites' infringing conduct" because the technical means for doing so suggested by plaintiff "were not workable." Id. at 731 (citing district court's opinion, Perfect 10 v. Google, Inc., 416 F. Supp. 2d 828, 857-58 & n. 25 (C.D.Cal.2006)). There is not, and cannot be, such a finding here as the case is presented on a 12(b)(6) motion.

[20] Anecdotal evidence suggests that the AdSense program produces vastly less revenue for most program members than what they earn through their businesses. One poll found that 45% of AdSense members surveyed earned less than $30 per month from the program, and only a small percentage earned a substantial amount. Darren Rowse, AdSense Earnings for November-Poll Results, Problogger (Dec. 19, 2005), http://www. problogger.net/archives/ 2005/12/19/adsense-earnings-for-november-poll-results/.

[21] Providing financial support has long been held to be a basis for vicarious infringement, where that financial support carries with it the contractual right to approve the infringing activity. See Davis v. E.I. DuPont de Nemours & Co., 240 F. Supp. 612 (S.D.N.Y.1965). In Davis, DuPont sponsored a dramatization of "Ethan Frome," which was alleged to infringe several copyrights. DuPont was held vicariously liable, even though it did not own the studio or the broadcast facilities, and could not have prevented airing of the show with another sponsor.

[22] See, e.g., Rules and General Orders of the Vermont Public Service Board § 3.302, available at http://www.state.vt.us/psb/rules/Official AdoptedRules/RulesComplete.pdf (last visited May 14, 2007) ("[N]o utility shall disconnect residential service of gas, electric, or water unless payment of a valid bill or charge is delinquent."); State of New Hampshire, Consumer Rights and Responsibilities, at 5 (1996), available at http://services.unitil.com/ content/info/consumer—rights.html.

[23] The majority is also mistaken when it suggests that parties would be held vicariously liable for infringement simply because, in a market economy, they are free not to deal with one another. Maj. op. at 805 n. 17. Our cases have been very clear that more is required for vicarious infringement; defendant must have a formal contractual or principal-agent relationship with the infringer. It is that contract or relationship that forms the predicate for vicarious liability, and plaintiff must point to some term in the contract or formal relationship that gives defendant a right to stop the infringing activity. See, e.g., Aimster, 334 F.3d at 654; Fonovisa, 76 F.3d at 262; Amazon, 487 F.3d at 729-31. Responding to a service call, in the absence of a contract which provides that the service may be discontinued in the event of illegal conduct, cannot form the basis of vicarious liability and thus the fact that the technician is free to leave can't render him vicariously liable. The requirement that plaintiffs point to a relationship explicitly spelled out in a contract or other legal arrangement is an important limitation on who may be held to answer for vicarious infringement. It should not be casually discarded.

[24] In fact, Perfect 10 has brought suit against some direct infringers. See Perfect 10, Inc. v. CCBill, LLC, 481 F.3d 751 (9th Cir.2007) (reversing summary judgment on direct infringement claim); Perfect 10, Inc. v. Talisman Commc'ns Inc., No. CV99-10450, 2000 WL 364813 (C.D.Cal. Mar.27, 2000).

[25] The majority finds it "anomalous" to hold credit cards liable without DMCA-compliant notice, while ISPs are immune unless they receive such a notice. Maj. op. at 795-96 n. 4. But there is no anomaly in treating parties that are covered by the statute differently from those that are not. Plaintiff here did give ample notice to the credit cards, see p. 824 infra, and should not have its claim dismissed for failing to allege compliance with a statute that does not apply to them.