Full Opinion

NOT FOR PUBLICATION FILED UNITED STATES COURT OF APPEALS OCT 29 2026 MOLLY C. DWYER, CLERK U.S. COURT OF APPEALS FOR THE NINTH CIRCUIT ROAR SPIRITS, LLC, a Delaware limited No. 25-2301 liability company; f/k/a El Bandido Yankee D.C. No. LLC, 4:23-cv-04809-HSG Plaintiff - Appellant, MEMORANDUM* v. SUTTER HOME WINERY, INC., a California corporation, Defendant - Appellee. Appeal from the United States District Court for the Northern District of California Haywood S. Gilliam, Jr., District Judge, Presiding Argued and Submitted October 21, 2026 San Francisco, California Before: S.R. THOMAS, KOH, and TUNG, Circuit Judges. Roar Spirits, LLC (“Roar”) appeals from the district court’s summary judgment for Sutter Home Winery, Inc. (“Sutter”) on the question of whether two of its trademarks were likely to cause confusion with Sutter’s registered mark. We * This disposition is not appropriate for publication and is not precedent except as provided by Ninth Circuit Rule 36-3. have jurisdiction under 28 U.S.C. § 1291, and we reverse. Roar and Sutter are two companies in the alcohol business. In 2020, Roar sought to register two trademarks: the word mark EL BANDIDO YANKEE and the composite mark EL BANDIDO YANKEE TEQUILA COMPANY BLANCO. Sutter opposed the registrations under 15 U.S.C. § 1052(d). It claimed that Roar’s marks were likely to cause confusion with Sutter’s BANDIT word mark. The Trademark Trial and Appeal Board sided with Sutter. Roar then sought de novo review in district court under 15 U.S.C. § 1071(b). The district court awarded Sutter summary judgment, prompting this appeal. We review a district court’s grant of summary judgment de novo. Metal Jeans, Inc. v. Metal Sport, Inc., 987 F.3d 1242, 1244 (9th Cir. 2021). We approve summary judgment only if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” FED. R. CIV. P. 56(a). We consider a dispute of fact to be “genuine” if a reasonable factfinder could side with the nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). We rarely approve summary judgment in likelihood-of-confusion cases. See JL Beverage Co. v. Jim Beam Brands Co., 828 F.3d 1098, 1105 (9th Cir. 2016). We treat likelihood of confusion as a question of fact. Quiksilver, Inc. v. Kymsta Corp., 466 F.3d 749, 759 (9th Cir. 2006). In a registration challenge, we evaluate the marks as described in the applications. Wells Fargo & Co. v. Stagecoach Props., Inc., 685 2 25-2301 F.2d 302, 306 (9th Cir. 1982). We analyze eight factors to decide whether consumers are likely to confuse the marks: (1) strength; (2) proximity of the goods; (3) similarity; (4) actual confusion; (5) marketing channels; (6) types of goods and degree of care likely to be exercised by the purchaser; (7) intent; and (8) likelihood of expansion of the product lines. AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 348– 49 (9th Cir. 1979); see B&B Hardware, Inc. v. Hargis Inds., Inc., 575 U.S. 138, 154 (2015) (“[T]he same likelihood-of-confusion standard applies to both registration and infringement.”). We decide the relative importance of each factor case by case. Brookfield Commc’ns, Inc. v. W. Coast Ent. Corp., 174 F.3d 1036, 1054 (9th Cir. 1999). 1. A reasonable factfinder could conclude that the marks are dissimilar. Similarity “has always been considered a critical question in the likelihood-of- confusion analysis.” GoTo.com, Inc. v. Walt Disney Co., 202 F.3d 1199, 1205 (9th Cir. 2000). Here, it proves decisive. A reasonable factfinder could side with Roar because of the difference in appearance, meaning, and sound. See id. at 1206. Start with the marks’ appearance and sound. We have found marks dissimilar because they differed by two letters. See, e.g., Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1145–46 (9th Cir. 2002); Yuga Labs, Inc. v. Ripps, 144 F.4th 1137, 1170 (9th Cir. 2025). Here, the differences are far more significant. Roar’s word mark exceeds Sutter’s word mark by two words and four syllables. “[T]his 3 25-2301 difference between the marks is a noticeable one” because the extra words make the EL BANDIDO YANKEE mark over “twice as long.” Entrepreneur Media, 279 F.3d at 1145. That rationale extends to the composite mark as well. The marks also diverge in meaning. Roar urges us to read the word YANKEE as a noun. Whether we read the word YANKEE as a noun or an adjective, it modifies the term BANDIDO in a salient way. YANKEE is far from “common” or “descriptive” of alcoholic drinks. Cf. Cunningham v. Laser Golf Corp., 222 F.3d 943, 947 (Fed. Cir. 2000) (holding that common and descriptive terms hold little weight in the analysis). The district court noted that YANKEE refers to the northeastern United States, whereas BANDIDO carries a “Mexican” connotation. By alluding to two distant geographic regions, Roar gave its marks unique meanings. See 5 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIR COMPETITION, §§ 14:7–8, 11. Finally, the marks create strikingly different overall impressions. See GoTo.com, 202 F.3d at 1206. Roar’s marks fuse English and Spanish words. And they refer to two distinct and unrelated geographic regions. Sutter’s mark has neither of those features. Based on these discrepancies, a reasonable factfinder could find that the marks are so dissimilar that they are unlikely to cause confusion in the marketplace. 2. The district court also erred by ignoring evidence of a lack of actual 4 25-2301 confusion. “Evidence that use of the two marks has already led to confusion is persuasive proof that future confusion is likely.” Sleekcraft Boats, 599 F.2d at 352. Conversely, the lack of actual confusion can favor the applicant. See Surfvivor Media, Inc. v. Survivor Prods., 406 F.3d 625, 633 (9th Cir. 2005). This factor could favor Roar for two reasons. First, the companies’ products coexisted for several years. See FocusVision Worldwide, Inc. v. Info. Builders, Inc., 859 F. App’x 573, 578 (Fed. Cir. 2021). Roar’s CFO attested that the company sold nearly 30,000 cases of tequila and earned millions of dollars in revenue from 2021 to 2024. Meanwhile, Sutter’s President and CEO declared that Sutter sold over a million cases between 2017 and 2021 and has continued to sell alcoholic beverages ever since. So there is at least some record evidence suggesting that the two companies sold large quantities of drinks during the same period. That is not all. Roar supplied an Eveready study finding that, out of 112 consumers, none identified Sutter as the source of Roar’s tequila. The district court erred when it set this survey aside. True, we must confine our legal analysis to the registration applications. See Wells Fargo & Co., 685 F.2d at 306. But a reasonable factfinder could find the survey evidence at least somewhat persuasive. They could infer that the lack of confusion in the marketplace bears some weight on the question of whether the products would cause confusion in the abstract. See J&J Snack Foods Corp. v. McDonald’s Corp., 932 F.2d 1460, 1463–64 (Fed. Cir. 1991) (stating that 5 25-2301 a survey was entitled to “appropriate weight” even if it “was not perfect in its content and execution”). To the extent this factor matters, a reasonable factfinder could conclude that it weighs in Roar’s favor. REVERSED AND REMANDED. 6 25-2301