dissenting:

This much is clear: American International Group and its lawyers knew defendant was calling itself American International Bank and did nothing at all for better than two and a half years. AIG sat on its hands while the Bank — a small, struggling business — built up a clientele, opened a new branch and finally turned the corner on profitability. AIG offers no excuse for its inordinate delay; yet the majority allows it to proceed with its infringement action, imposing the daunting expense of a trial on a party who has been operating under the American International mark for twelve years. This is precisely the type of case laches is meant to bar.

The majority reaches the wrong result because it applies the wrong test. An issue of fact is not enough to defeat summary judgment; there must be a genuine issue of material fact, a dispute capable of affecting the outcome of the case. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248-50 , 106 S.Ct. 2505, 2510-11 , 91 L.Ed.2d 202 (1985). Here, AIG raises only quibbles; each and every E-Systems factor favors the Bank and will favor the Bank *834 regardless of AIG’s proof at trial. The learned district judge got it just right.

A. The first E-Systems factor is the strength of the mark. E-Systems, Inc. v. Monitek, Inc., 720 F.2d 604, 607 (9th Cir.1983). A fanciful or made-up mark is normally strong; one that is merely descriptive or suggestive is weak. See Accuride Int’l, Inc. v. Accuride Corp., 871 F.2d 1531, 1536 (9th Cir.1989). American International is, as the district court found, largely descriptive, “a common phrase used by many businesses.” Statement of Un-controverted Facts and Conclusions of Law 7 (Oct. 22, 1987), ER 32 [hereinafter “Findings”]. It’s hard to disagree: There are few terms more generic than American; it describes someone or something originating in the United States. International is an equally descriptive word, suggesting a connection with foreign trade or commerce. Nor does it take much creativity to juxtapose the two; what else does a domestic company looking for some cosmopolitan panache call itself? 1

Not surprisingly, dozens of businesses across the country use American International as part of their name. The Manhattan phone book lists 24 entities operating under the name American International something; in Washington, D.C., there are 12 listings; in Miami 28; in Los Angeles 20; San Francisco 12; Philadelphia 10; Chicago 11; San Diego 10. Kodak or Xerox this ain’t. 2

It’s true that intrinsically weak marks “may be strengthened” by “advertising, length of exclusive use, [and] public recognition.” Majority at 832 (quoting Accuride, 871 F.2d at 1536 ) (emphasis added). It’s also irrelevant: AIG has presented no evidence that it managed to elevate its mark above the cacophony of other marks that use American or International or a combination of the two. AIG’s advertising may show an effort at strengthening its mark, but what counts is success. See First Brands Corp. v. Fred Meyer, Inc., 809 F.2d 1378, 1383 (9th Cir.1987); 1 J. Gilson, Trademark Protection & Practice § 209[5], at 2-123 (1990). This AIG has not demonstrated.

B. The second E-Systems factor, diligence in enforcing the mark, also strongly favors the Bank; in fact, it is conclusive. The Bank began operating under the name American International in 1978. For five years it used the name openly and notoriously in one of the country’s largest business communities, but AIG failed to identify the alleged conflict. Worse, in 1982 an AIG member company underwrote the Bank’s directors’ and officers’ insurance; no one was troubled by the similarity in the marks. Findings 4, ER 29.

Companies expecting judicial enforcement of their marks must conduct an effective policing effort, which may include training employees to detect infringers. Were we, nevertheless, to excuse AIG’s underwriters for failing to ring the bell on the Bank’s potential infringement, we surely cannot shut our eyes to similar inaction by plaintiff’s lawyers. AIG’s general counsel noticed the Bank’s name in the Los Angeles telephone book in June 1983 but did nothing at all for six months, at which time he sent the information to AIG’s trademark counsel. Id. at 5, E.R. 30. Trademark counsel must not have seen a menace to AIG’s mark either, as he took no action for two more years — not even so much as a letter of inquiry.

The majority admits that AIG offers “no excuse” for this two and a half year delay, majority at 830, but nevertheless concludes that AIG has raised a genuine issue of fact as to diligence. Comyare majority at 830 with id. at 832. But the only evidence of diligence the majority can muster is that *835 AIG has enforced its mark against other parties. Id. at 832. My colleagues do not explain how diligence against unrelated entities can defeat the Bank’s laches defense once the Bank conclusively establishes that AIG was not diligent as to it. Diligence as to third parties might have some relevance where, despite vigorous enforcement efforts, a particular infringer remains undetected. But here we are dealing with a two and a half year post-discovery delay. How enforcement efforts against third parties can show diligence as to the Bank under these circumstances is beyond me.

The fundamental premise of laches is that those who sleep on their rights surrender them; if you snooze, you lose. Here, the claimed infringement took place under AIG’s nose for eight years; three and a half years before it filed suit, AIG began doing business with the Bank; and for two and a half years, AIG’s lawyers knew of the infringement but did nothing. Having chosen two words that accurately describe thousands of businesses, AIG could expect others to be equally uninspired in their selection of a trade name. To preserve its rights, AIG was required to be especially diligent in policing its mark; instead, AIG was especially lackadaisical.

C. The third and fifth E-Systems factors, harm to the senior user and the existence of competition between the senior and junior users, go to one issue: whether the public is likely to be confused by the similarity in names. The majority concedes there’s no proof of actual confusion. Majority at 832. This is especially significant given that AIG and the Bank operated in Los Angeles, side-by-side, for eight years. In all that time, AIG has been able to document not a single instance of confusion, not so much as a misdirected phone call. 3 Lack of actual confusion during such a substantial period of concurrent use in the same geographic market establishes there is no likelihood of confusion; if confusion were a real problem, it would have happened already. Plus Prods. v. Plus Discount Foods, Inc., 722 F.2d 999 , 1006 (2d Cir.1983); 4 3A R. Callman, The Law of Unfair Competition, Trademarks & Monopolies § 20.06, at 27, § 20.61, at 498 (4th ed. 1983); see Prudential Ins. Co. v. Gibraltar Fin. Corp., 694 F.2d 1150, 1156 (9th Cir.1982), cert. denied, 463 U.S. 1208 , 103 S.Ct. 3538 , 77 L.Ed.2d 1389 (1983). Not even AIG’s employees were confused, unless they thought they were selling insurance to themselves. See SER 7-9. 5

This lack of confusion is not surprising, as AIG and the Bank are not competitors: AIG is not authorized to act as a commercial bank and has no plans to diversify in that direction, Findings at 6, 9, ER 31, 34; the Bank does not sell insurance, id. at 5, ER 30. Moreover, providers of commercial banking and insurance services do not compete even indirectly; having a bank account does not materially reduce one’s need for insurance nor vice versa.

The majority nonetheless suggests there might be confusion because the Bank and AIG both provide “financial services.” Majority at 832. All this proves is that, if you define the product market broadly enough, you can encompass any number of businesses, no matter how little they compete with each other. American Cab, American Airlines and American Motors all *836 provide “transportation services,” but no one is likely to call American Cab for a ride from New York to London; American Telephone & Telegraph and American Broadcasting Company both provide “mass communication services,” but ABC cheerfully carries AT & T’s advertising; the American Civil Liberties Union, the American Legion and the American College of Foot Surgeons are all fairly characterized as “public service organizations,” yet I rather doubt the ACLU gets a lot of calls about podiatry. Most people know the difference between a bank and an insurance company; I doubt they will be confused just because my colleagues have come up with a term fuzzy enough to cover both institutions.

The crux of plaintiffs case seems to be, not that AIG and the Bank compete now, but that they may compete at some time in the future. This is pretty thin stuff given that the prejudice to AIG is hypothetical and contingent rather than immediate and concrete; it’s so thin that our caselaw rejects it outright. Prudential Ins. Co., 694 F.2d 1150 , involved a similar struggle between an insurance company and a financial institution. At stake was Gibraltar Financial Corporation’s right to use the Rock of Gibraltar, Prudential’s longtime logo. Prudential challenged the district court’s finding of no competition, contending that it and Gibraltar were both selling insurance and that, given the prospect of deregulation, they would “soon compete in a wide variety of services.” Id. at 1155 . We held that “[t]he mere possibility of future competition is too tenuous a basis on which to reverse the district court.... [T]he growth of the companies and the changes in their services do not create a basis for relief.” Id. (emphasis added).

AIG has a far weaker claim than did Prudential. Because the Rock of Gibraltar has only a very tenuous relationship to financial services, it’s a relatively strong mark; it is descriptive only in a remote and metaphorical way. Strong marks are sometimes protected against use on noncompeting goods; weak marks never are. 3A R. Callman, supra p. 835, § 20.43, at 345 {“Only a strong mark will be protected against the defendant’s use on noncompeting goods.” (emphasis added)); id. (“If a mark is weak, its protection may have an extremely narrow scope and may even be limited to similar goods similarly marketed.”); see AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 350 (9th Cir.1979); 1 J.T. McCarthy, Trademarks and Unfair Competition § 11.26, at 512 (2d ed. 1984). As we explained over 50 years ago:

Treager v. Gordon-Alien, Ltd., 71 F.2d 766, 768 (9th Cir.1934), (quoting France Milling Co. v. Washburn-Crosby Co., 7 F.2d 304, 306 (2d Cir.1925)), quoted in 3A R. Callman, supra p. 835, § 20.43, at 345. Because AIG’s mark is weak, see p. 834 supra, the protection to which it is entitled is necessarily limited to the area in which it operates, namely insurance; it does not extend to banking.

Particularly significant is the district court’s uncontroverted finding that, when the Bank opened its doors, AIG’s registration listed insurance underwriting as its only vocation; AIG claimed no rights in the amorphous area of “financial services.” See Findings at 8-9, ER 33-34. Registration does not establish the exclusive right to use a mark; it is, however, evidence of that right. See 15 U.S.C. § 1057 (b); Mushroom Makers, Inc. v. R.G. Barry Corp., 580 F.2d 44, 48 (2d Cir.1978), cert. denied, 439 U.S. 1116 , 99 S.Ct. 1022 , 59 L.Ed.2d 75 (1979); 1 J. Gilson, supra p. 834, § 4.01. The absence of registration indicates that AIG itself did not consider its rights as extending beyond insurance underwriting. Moreover, the absence of registration deprived other businesses, such as defendant, of constructive notice of AIG’s claim. See id. § 4.02.

The majority is right, of course, that the party opposing summary judgment is entitled to the benefit of all reasonable infer- *837 enees. Majority at 832. But inferences cannot be drawn from thin air; they must be based on evidence which, if believed, would be sufficient to support a judgment for the nonmoving party. See Celotex v. Cattret, 477 U.S. 317, 322 , 106 S.Ct. 2548, 2552 , 91 L.Ed.2d 265 (1986) (party with burden of proof at trial must come forward with evidence sufficient to sustain a verdict in its favor). AIG has produced nothing to prove it competes with the Bank or is harmed by the Bank’s use of the mark; everything in the record is to the contrary. These E-Systems factors also favor the Bank.

D. Plaintiff didn’t deny the Bank was unaware of AIG’s claim that its mark extended to banking services. How could it have? Had the Bank performed a trademark search before selecting its name, it would have found that AIG had claimed rights only in insurance underwriting services. Findings at 8-9, ER 33-34; see p. 836, supra. The majority does not dispute that this fourth E-Systems factor favors the Bank.

E. Finally, the majority finds a triable issue of fact as to the sixth E-Systems factor, prejudice to the Bank caused by AIG’s delay. My colleagues suggest that, because the Bank was unaware of AIG’s existence, there was no “detrimental reliance.” Majority at 833. But the Bank is not claiming estoppel; it’s claiming laches —i.e., that it was harmed by AIG’s lack of diligence. Estoppel requires detrimental reliance; laches does not. Indeed, in E-Systems, we found laches notwithstanding the fact that the defendant was unaware of the senior user’s existence. E-Systems, 720 F.2d at 607 .

Unlike my colleagues, the district court properly looked to the harm occasioned by AIG’s unjustified delay, finding that the Bank “is a struggling entity that only recently turned the corner into profitability in the highly competitive Southern California banking world, and it would be dealt a severe blow if all of its hard-earned good will were stripped from it in this litigation.” Findings at 9, ER 34. The majority dismisses this finding as “unsubstantiated” but gives no citation to the record. Majority at 833. In fact, the record contains much that supports, and nothing that controverts, the district court’s finding.

In 1983, the Bank was — in AIG’s words —“another lousy small California Bank that lost over a million last year [with] negative undivided profits.” CR 38, at 123; see SER 7. But AIG did not warn the Bank of its infringement claim back then, when a name change would have cost the Bank relatively little by way of reputation or good will; instead AIG waited — and underwrote the Bank’s directors’ and officers’ liability insurance. CR 38, at 141; see SER 8-9. Did AIG say anything about its infringement claim in 1984, when the Bank lost another $2.3 million? No. Or in 1985, when the FDIC became concerned about the Bank’s stability and required it “to correct unsatisfactory conditions?” No. See CR 37, at 2-3. AIG, fully aware of the Bank’s name and its struggle, did nothing, year after year after year.

The Bank’s horizons had begun to brighten by the beginning of 1986: It turned the corner on profitability, got approval to open a new branch and corrected all its deficiencies to the FDIC’s satisfaction. More importantly, the Bank’s customer base swelled: Loans grew by 62% to $31 million and deposits increased by 25% to $43 million. CR 37, at 3. It was only then — after the Bank had been in business eight years and had spent untold dollars establishing itself, developing a customer base and opening a new branch — that AIG first asserted its claim.

If these facts, none of which are controverted, fail to establish prejudice, what will? Trademark law is, after all, premised on the notion that what a business calls itself makes a difference because, within its niche, the name becomes synonymous with the business itself. As the leading treatises on trademark law explain, the “[g]ood will of a business and its symbol, a trademark, are inseparable.” 1 J.T. McCarthy, supra p. 836, § 2.7, at 69; accord 1 J. Gilson, supra p. 834, § 1.03[5], at 27-28 (“[Trademarks] are inextricably bound up with [good will], and they are *838 protected because of it.” (emphasis in original)). Surely the district court was correct when it concluded that stripping the Bank of its “hard-earned good will” would deal it “a severe blow.” Findings at 9, ER 34.

Had AIG made its claim earlier, the Bank could have chosen to plough a different furrow. A change of name now, at the very least, will require a costly consumer reeducation campaign; it will also, almost certainly, undermine consumer confidence by suggesting the Bank is financially unstable and was forced to change names because it was taken over by another entity. Public confidence in our financial institutions has been severely shaken; changing the Bank’s name will no longer be as simple as making a new sign for the front door and ordering new stationery. This factor, like all the others, strongly favors the Bank.

A dispassionate reading of the record leads to only one conclusion: All the E-Systems factors favor the Bank. But even if some of the factors were to favor AIG, summary judgment would still be proper. E-Systems contemplates a balance, 720 F.2d at 607 ; to avoid summary judgment, AIG must point to disputes of sufficient weight and number that the entire balance could tip in its favor. It hasn’t even come close. My colleagues conclude that summary judgment was improper because “five of six E-Systems factors involve disputed issues of material fact,” majority at 833, but surely E-Systems contemplates more than simple bead-counting. The issue is not how many factors favor each party but their weight. The majority fails to explain how the balance could possibly tip in AIG’s favor once the Bank has shown that AIG’s lawyers sat around contemplating their navels for two and one half years while the Bank was struggling to build up its good will. Nothing AIG has proffered can overcome this hard fact. We should save everyone the time and expense of a trial by putting this case out of its misery now.

. This is exactly why the Bank chose it. As the Bank's founder and first president explained, American "reflect[s] the strength and integrity of our country” and International recalls "the cultural and ethnic diversity of our shareholders, our potential customers, and the Los Ange-les community in which we intended to do business.” CR 36, at 2.

. That AIG has sued a lot of other businesses cuts both ways. It suggests diligence but also demonstrates how nondistinctive the mark really is. See Plus Prods. v. Plus Discount Foods, Inc., 722 F.2d 999 , 1005 (2d Cir.1983).

. Misdirected inquiries by consumers, advertisers and suppliers are common signs of public confusion about the source of a trademark. See, e.g., International Kennel Club v. Mighty Star, Inc., 846 F.2d 1079 , 1090-91 (7th Cir.1988); Mustang Motels, Inc. v. Patel, 226 U.S.P.Q. 526 , 527 & n. 1 (C.D.Cal.1985).

. In Plus Products, the Second Circuit relied on the absence of confusion during a three year period of concurrent use. 722 F.2d at 1006. The case here is even more compelling: Over eight years went by without a single documented instance of confusion.

.On this issue, AIG is caught in a double bind from which it cannot possibly extricate itself at trial. If it comes up with proof that it received complaints from confused consumers during the eight years it and the Bank operated side by side, it will have to explain why it did not press its claim earlier. If it turns out that there were no complaints — as almost certainly there were not — there is no confusion. The Bank wins either way, so why put the parties through the agony of a trial?