CUMULUS MEDIA NEW HOLDINGS INC v. NIELSEN COMPANY US LLC (2026)

Defendant-Appellant The Nielsen Company (US), LLC collects, compiles, and sells radio broadcasting data that tracks audience listenership and reach across the United States. It sells two different types of products: local radio data, which reflects listenership in specific geographic areas, and national radio data, which consists of all local data assembled into a comprehensive nationwide report. Nielsen is the only supplier of national radio data in the United States. Both types of data are critical for large radio broadcasters like Plaintiff-Appellee Cumulus Media New Holdings Inc., which uses national data to sell advertising time on its national radio networks, and local data to sell the same to advertisers on its local radio stations.

Nielsen once offered its national and local data products separately. But that changed in 2024, when it adopted a new policy barring national broadcasters like Cumulus from purchasing a functional version of its national report unless they also purchased all relevant local data products from Nielsen as well. That policy posed a problem for Cumulus, as it hoped to buy Nielsen's national report but only some of Nielsen's relevant local data. Cumulus wanted to buy at least some other local data from a competitor of Nielsen's.

After contract negotiations between the parties reached an impasse on those grounds, Cumulus sued Nielsen, claiming that its new policy is an anticompetitive tying arrangement in violation of the Sherman Act. Following discovery and a hearing, the district court agreed and granted Cumulus a preliminary injunction. We review for abuse of discretion and affirm that interlocutory order. The district court did not abuse its discretion when it concluded that Nielsen's new policy—enforced both expressly and constructively—was an unlawful tie, because Nielsen exploited its monopoly in the national radio data market in order to force customers like Cumulus to purchase its local data products. Nor did the district court abuse its discretion by finding that Nielsen's conduct caused anticompetitive effects in the relevant local data markets, that Cumulus would suffer irreparable harm as a result of the new policy, and that the public interest and the balance of equities tipped in Cumulus's favor. We further hold that the district court's injunction is sufficiently specific and tailored to remedy Cumulus's harm, and we agree with both parties that we may resolve this appeal despite Cumulus's interceding bankruptcy petition.

This interlocutory appeal raises several complicated issues of antitrust law. On this preliminary posture, we resolve them in favor of Cumulus, mindful of the deference we owe to the district court's factual findings and discretionary decisions. Accordingly, we affirm the district court's order in full.

BACKGROUND

A. Factual Background

This antitrust dispute concerns radio broadcasting data that drives the price and placement of advertisements on radio stations across the country. Cumulus, one of the largest audio media companies in the United States, sued Nielsen for its allegedly anticompetitive conduct in the sale of that crucial data.

The following facts are drawn from the district court's findings, to which we defer unless they constitute clear error. See D.L. Cromwell Invs., Inc. v. NASD Regul., Inc., 279 F.3d 155, 158 (2d Cir. 2002). In creating this factual record, the district court considered, among other things, “the Complaint, documents cited in the Complaint, and deposition testimony and declarations submitted by the parties.” Cumulus Media New Holdings, Inc. v. Nielsen Co. (US) LLC, No. 25-cv-8581, 2026 WL 63294, at *1 (S.D.N.Y. Jan. 8, 2026).

Both Nielsen and Cumulus are major players in the radio industry. Cumulus is a broadcaster that operates 395 radio stations nationwide and distributes audio content to over 9,500 other affiliated stations through its syndication network, Westwood One. Like other broadcasters, Cumulus generates revenue by selling advertising space—i.e., commercial time on its various radio stations—to advertisers. Nielsen, for its part, is a market research company that collects, analyzes, and sells radio ratings data. Nielsen collects audience listening statistics in more than 270 geographic areas; it then compiles that local information into a national ratings report known as the Nationwide Report (Nationwide). It sells both its Nationwide product as well as data in each discrete local area it surveys. Nielsen has one competitor in the local radio data space: Eastlan, which surveys listeners and produces radio data primarily in “smaller-sized” geographic markets as well as in geographic markets where Nielsen does not operate. Cumulus, 2026 WL 63294, at *3. But Nielsen has no competitors in the market for national radio ratings data, in which it enjoys a 100% market share.

Both national and local data are essential for broadcasters such as Cumulus. National audio networks and local radio stations alike generate revenue by selling advertisements, so broadcasters use both types of data to market their available advertising inventory. For example, while advertisers purchasing time slots on national radio reasonably seek to assess a broadcaster's national reach (that is where Nationwide comes in), advertisers purchasing time slots in only certain local markets seek data that reflects that area's specific listenership. Accordingly, Cumulus needs both types of data to adequately price and sell its time slots to different kinds of advertisers.

Nielsen has historically offered Nationwide and its individual local ratings data as separate products. However, in 2024, it announced a change to that practice. Nielsen's new “Network Policy” prohibits broadcast networks that operate “a local station in a Nielsen-measured market” and do not “subscribe” to Nielsen's local ratings data in that market from purchasing Nielsen's Nationwide product with the inclusion of data for “that specific market[.]” Cumulus, 2026 WL 63294, at *4. For example, if a national broadcasting network also operates local stations in 30 markets and purchases only Nielsen's Nationwide product—not Nielsen's local data—Nielsen will omit data for those 30 local markets from Nationwide.¹ Nielsen says it adopted the Network Policy in order to stop national broadcast customers from sharing Nationwide with their local affiliates for free, making decisions for those local affiliates based on information in Nationwide, or extracting the relevant local data from Nationwide without paying for it. According to an executive at Nielsen, the Policy was designed to “bring groups with non-subscribing markets back to the negotiation table,” “command subscriptions in local markets,” and “prevent networks from getting data through the back door.” Appellee's Suppl. Sealed App'x 440. But it is undisputed that a Nationwide product that exempts data from certain local markets is not useable. Indeed, in Nielsen's own words, the exempted Nationwide product is “like Swiss cheese”—and in Cumulus's case, that “Swiss cheese” national product would exclude data from numerous geographic markets where Cumulus operates, including major markets such as San Francisco, Los Angeles, and New York City. See Cumulus, 2026 WL 63294, at *6.

The Network Policy and its impact on the parties’ contract negotiations lie at the heart of this dispute. We briefly summarize those negotiations below. Cumulus's previous contract with Nielsen—executed before the development of the Network Policy—was set to expire at the end of 2025. Under that contract, Cumulus had purchased both Nationwide and Nielsen's local data in 76 discrete markets. However, before the parties began to negotiate a new contract in May 2025, Cumulus decided that it no longer wanted to subscribe to some of those 76 local markets. Instead, it determined that the cost of Nielsen's local ratings data had outstripped their value in most local radio markets, and it intended to switch to Eastlan—a lower-cost alternative—in at least some of those markets. Accordingly, at the outset of negotiations, Cumulus informed Nielsen that it wished to purchase only its local data in certain markets, not the full suite of local markets that Cumulus had previously purchased.

But the Network Policy impeded that possibility. In June, Nielsen made its first offer to Cumulus: the full Nationwide product, “plus local service in every local market in which Cumulus operates radio stations.” See id. at *5. Nielsen followed that up with another offer that also included all local markets in which Cumulus operates. Cumulus, however, did not want to subscribe to all of those local markets. So Cumulus asked Nielsen to provide a price for the Nationwide product alone, as well as market-by-market pricing for the local data it sought to purchase. Nielsen refused to provide either price, citing its Network Policy. See id. at *6 (quoting a Nielsen executive stating that a standalone price for Nationwide was “not possible to achieve in light of our policy on local subscription”). Instead, in line with its Network Policy, Nielsen offered Cumulus its desired subset of local markets but only the “Swiss cheese” version of Nationwide, which omitted data from all other local markets where Cumulus operates. Nielsen told Cumulus that it would sell the full version of Nationwide only if Cumulus also subscribed to all local markets in which Cumulus operates. Cumulus counteroffered for only its desired subset of local markets.

In August, the negotiations reached an impasse. Cumulus then sent a cease-and-desist letter, claiming that Nielsen's Network Policy violated the antitrust laws by unlawfully tying its products together (its Nationwide product and its data in the local markets) and threatening to file suit if Nielsen continued to enforce the Network Policy. About a month later, Nielsen responded by exempting Cumulus from the Network Policy by offering, for the first and only time in the negotiations, a price for Nationwide as a standalone product. The offered price was at least 150% more than any other national network paid for Nationwide, and it was also “ten times more than what Cumulus was paying for Nationwide under its existing contract.” See id. Cumulus did not accept. Finally, on October 16, 2025, Nielsen offered Cumulus its full Nationwide product plus local data in all 80 markets where Cumulus operates—which again included local markets that Cumulus did not want to purchase from Nielsen. That same day, Cumulus filed this lawsuit in the United States District Court for the Southern District of New York.

B. Procedural Background

Cumulus in its complaint accuses Nielsen of violating Section 2 of the Sherman Antitrust Act, 15 U.S.C. § 2, by creating an unlawful tying arrangement—what Cumulus calls “a textbook abuse of monopoly power.”² Compl. at 2. An unlawful tying arrangement occurs when a seller uses its economic power in one market to coerce a buyer into purchasing a different product in a separate market by refusing to sell the two products separately. Here, Cumulus claims that the Network Policy “ties” Nielsen's Nationwide product to its local data products and forces Cumulus into an illusory choice: either purchase Nielsen's local data in all markets, or use competitor Eastlan for local data, but receive only Nielsen's unusable “Swiss cheese” Nationwide Report. And though Nielsen later offered Cumulus a standalone price for full Nationwide, Cumulus argues that the exorbitant price of Nielsen's offer left Cumulus with the same illusory choice. Because Cumulus cannot buy national data from anyone else, and because Nielsen's exorbitant price for standalone Nationwide means it is not economically viable for Cumulus to separately purchase local data from its preferred vendor, Cumulus says the Network Policy has effectively foreclosed its ability to purchase Eastlan's local data and distorted competition in the relevant local data markets by fencing out competitors. All this, Cumulus says, constitutes an unlawful tying arrangement prohibited by the Sherman Act.

On the same day it filed its complaint, Cumulus moved for a preliminary injunction, seeking to bar Nielsen from imposing any tie related to Nationwide during the parties’ negotiations. After a three-day evidentiary hearing featuring testimony from economics experts and the parties’ representatives, the district court granted Cumulus's motion. We summarize the district court's relevant findings of fact and conclusions of law below. At the outset, the district court determined that Cumulus sought to alter the status quo and therefore subjected its motion to the more demanding standard for mandatory injunctions. But even under that heightened standard, the district court concluded that Cumulus established a clear likelihood of success on the merits of its tying claim as well as a strong showing of irreparable harm.

The district court first considered the merits of Cumulus's express and constructive tying claim. As an initial matter, it adopted Cumulus's definitions of the relevant product and geographic markets which, “[f]or purposes of this preliminary injunction motion, the parties are in agreement”: The product markets are those for “local radio ratings data and national radio ratings data,” Cumulus, 2026 WL 63294, at *10, and the geographic markets are the United States (national data) and “each local geographic area for which a ratings report is generated” (local data), id. at *11. It then concluded that “Nielsen's Network Policy is an anticompetitive tying policy,” reasoning that the Policy expressly conditions the availability of Nationwide on a buyer's purchase of Nielsen's local data, that Nielsen had used its monopoly power in the national data market to coerce customers into purchasing its local data, and that Nielsen's conduct had unlawfully restrained competition in the various local data markets. Cumulus, 2026 WL 63294, at *12–14.

Though it recognized that Nielsen had facially exempted Cumulus from the Network Policy, the district court found Nielsen's Nationwide-only offer to be “so exorbitant as to make it economically unfeasible to purchase Nationwide as a separate product.” Id. at 13. That high price, it said, “therefore served as a constructive tie.” Id. The district court also recognized that tying arrangements are per se unlawful—without regard to their anticompetitive effects—when the perpetrator is a monopolist in the tying market, like Nielsen is for national data. Id. But even absent such a per se rule, the district court went on to analyze the anticompetitive effects of the constructive tie and concluded that Nielsen's tying caused anticompetitive effects because “Eastlan is effectively foreclosed from competing” with Nielsen in many local markets. Id. The district court further rejected Nielsen's procompetitive justifications for its Network Policy on grounds that they were both unsupported by the record and pretextual. Id. at *14–15. All told, the district court concluded that Cumulus was substantially likely to succeed on its claim that Nielsen had unlawfully exploited its monopoly power in the national data market to restrain competition in local data markets via both express and constructive tying.

Next, the district court found that Cumulus made a strong showing of irreparable harm on three grounds. First, it held that “ ‘[t]hreatened economic harm to consumers is plainly sufficient to authorize injunctive relief’ under Section 16 [of the Clayton Act].” Id. at *15 (quoting New York ex rel. Schneiderman v. Actavis PLC, 787 F.3d 638, 661 (2d Cir. 2015)). Second, it found that Cumulus would likely experience “major disruption” of its business, loss of goodwill, and decreased market share without access to Nationwide because it would be unable to develop reasonable advertising proposals for the coming year. Id. at *15–16. Advertisers, the district court found, were therefore “all but guaranteed to shift some or all of their advertising inventory purchases to competitors or refrain from purchasing advertising inventory from [Cumulus] at all[.]” Id. at *16. Third, the district court held that distortion of the relevant local ratings markets due to anticompetitive behavior also constituted irreparable harm. Thus it concluded that Cumulus carried its burden on the irreparable harm prong. Finally, the district court determined that the balance of the equities and public interest favored Cumulus, reasoning that Nielsen would suffer little harm from the injunction and that barring anticompetitive behavior serves the public's interest.

The district court granted Cumulus a preliminary injunction. Its order barred Nielsen “from enforcing its Network Policy” and “from charging a commercially unreasonable rate for its Nationwide Report as a complete, standalone product.” Id. at *18. The order further provided that “a rate that is equal to or lower than the highest annual 2026 rate Nielsen charges any broadcaster (whether network or local) for Nationwide is presumptively reasonable.” Id.

Nielsen thereafter moved the district court to stay the preliminary injunction pending appeal. The district court denied Nielsen's motion for a stay pending appeal, but granted Nielsen's alternative motion for a short administrative stay to allow Nielsen to seek relief from this Court. Nielsen then noticed this interlocutory appeal and moved for a stay of the district court's order pending the appeal, a request that a motions panel granted in February 2026. In the meantime, Nielsen filed an answer in the district court asserting several counterclaims against Cumulus, all of which Cumulus moved to dismiss.

On March 10, 2026, Cumulus notified the district court that it had filed a petition for voluntary Chapter 11 bankruptcy. In the notice, Cumulus explained its view that, under Section 362(a)(1) of the Bankruptcy Code, its bankruptcy petition operated as an automatic stay of Nielsen's counterclaims because they were asserted “against [a] debtor.” 11 U.S.C. § 362(a)(1). And though Cumulus reserved its right to argue that its bankruptcy did not automatically stay its own claims against Nielsen (or this interlocutory appeal), it consented to a stay of its claims below as a matter of the district court's discretion. The next day, the district court stayed litigation of Nielsen's counterclaims under Section 362(a)(1) and stayed the remaining claims “pending further order[.]” By then, this expedited appeal was well underway.

STANDARD OF REVIEW

We review the district court's grant of a motion for a preliminary injunction for abuse of discretion. State Farm Mut. Auto. Ins. Co. v. Tri-Borough NY Med. Prac. P.C., 120 F.4th 59, 82 (2d Cir. 2024). “A district court ‘abuses’ or ‘exceeds’ the discretion accorded to it when (1) its decision rests on an error of law (such as application of the wrong legal principle) or a clearly erroneous factual finding, or (2) its decision—though not necessarily the product of a legal error or a clearly erroneous factual finding—cannot be located within the range of permissible decisions.” Zervos v. Verizon N.Y., Inc., 252 F.3d 163, 169 (2d Cir. 2001) (footnote omitted). And a “finding is clearly erroneous when[,] although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” Wu Lin v. Lynch, 813 F.3d 122, 126 (2d Cir. 2016) (quotation marks omitted). The abuse-of-discretion standard requires, in other words, that we take three different approaches to the district court's order: We assume the district court's factual findings are true unless we are firmly convinced they are wrong (clear error), we review the district court's legal rulings without deference (de novo), and we review the district court's discretionary decisions by asking whether they were within the range of permissible decisions.

DISCUSSION

Nielsen raises several challenges to the district court's grant of a preliminary injunction. It claims that the district court abused its discretion in applying all four prerequisites for obtaining interim relief, and that the resulting injunction is both too ambiguous and a poor match for Cumulus's asserted harms. In addition to reaching these issues, we consider an antecedent one: the impact of Cumulus's pending bankruptcy on the status of this appeal.

I. Section 362(a)(1) Automatic Bankruptcy Stay

Before resolving the merits, we must decide whether to hold this appeal until Cumulus's bankruptcy proceedings conclude. See Ostano Commerzanstalt v. Telewide Sys., Inc., 790 F.2d 206, 207 (2d Cir. 1986). Under Section 362(a)(1) of the Bankruptcy Code, also known as the automatic-stay provision, a bankruptcy petition ordinarily “operates as a stay, applicable to all entities, of ․ a judicial, administrative, or other action or proceeding against the debtor that was or could have been commenced” before the filing of the petition. 11 U.S.C. § 362(a)(1). The provision renders judicial proceedings, including appeals, “void and without vitality if they occur after the automatic stay takes effect.” Rexnord Holdings, Inc. v. Bidermann, 21 F.3d 522, 527 (2d Cir. 1994); see also Ostano, 790 F.2d at 207. The statute limits the reach of the automatic stay to proceedings brought “against the debtor,” which we assess based on the parties’ status at the outset of an action rather than by who is “ahead” or who has appealed. Teachers Ins. & Annuity Ass'n of Am. v. Butler, 803 F.2d 61, 64–65 (2d Cir. 1986). That said, a counterclaim asserted against a debtor, like those Nielsen asserted against Cumulus in this case, also counts as a “proceeding against the debtor” that must be stayed upon the filing of a bankruptcy petition. See Koolik v. Markowitz, 40 F.3d 567, 568 (2d Cir. 1994). That raises the question whether the assertion of a counterclaim against a debtor triggers an automatic stay of the entire action, including the plaintiff-debtor's original claims. We hold that it does not. Because an action must be “disaggregated” into claims brought by and against a debtor, see Mar. Elec. Co. v. United Jersey Bank, 959 F.2d 1194, 1204–05 (3d Cir. 1991), Section 362(a)(1) does not stay litigation over a debtor's original claims.³

Our interpretation of Section 362(a)(1) begins with its text, which we read not “in isolation” but “as a whole.” United States v. Bulloch, 165 F.4th 676, 682 (2d Cir. 2026) (quotation marks omitted). As the Third Circuit has explained, “the clear language of section 362(a) indicates that it stays only proceedings against a ‘debtor.’ ” Mar. Elec. Co., 959 F.2d at 1204 (quoting 11 U.S.C. § 362(a)(2)). The operative statutory term is “proceeding,” which the Bankruptcy Code does not define. See 11 U.S.C. § 101. But in the bankruptcy context in particular, the term “proceeding” is used to refer to “[a] particular dispute or matter arising within a pending case—as opposed to the case as a whole.” Proceeding, Black's Law Dictionary (12th ed. 2024) (emphasis added). After all, it is typical for bankruptcy cases to comprise core and non-core proceedings, see In re Ben Cooper, Inc., 896 F.2d 1394, 1397–98 (2d Cir. 1990), which is strong evidence that Congress intended the latter term to carry a narrower, specialized meaning in this context, see United States v. Hansen, 599 U.S. 762, 774–76, 143 S.Ct. 1932, 216 L.Ed.2d 692 (2023). Indeed, Congress drafted at least one other bankruptcy provision with precisely this distinction in mind, permitting district courts to refer “proceedings ․ arising in or related to a case under title 11” to a bankruptcy judge. 28 U.S.C. § 157(a) (emphases added). To define “proceeding” to be coextensive with an entire “case” would collapse these careful distinctions.

Any lingering ambiguity in the statutory text is resolved by its evident purpose. Cf. Mar-Can Transp. Co., Inc. v. Loc. 854 Pension Fund, 167 F.4th 581, 593 (2d Cir. 2026). We have explained that the rationale behind the automatic stay is “to give the debtor time to organize its affairs—which includes protection from having to defend claims brought against the estate as well as continuing to pursue judicial proceedings on its own behalf.” Teachers, 803 F.2d at 65. “The stay is designed to provide the debtor with a breathing spell from his creditors,” Koolik, 40 F.3d at 568 (quotation marks omitted), and lest such proceedings “distract a debtor's attention from its primary goal of reorganizing,” Teachers, 803 F.2d at 65. Those goals are well-served by permitting debtors to invoke the protections of the automatic stay against counterclaims, but would be frustrated by an interpretation of Section 362(a)(1) that permitted defendants to delay a debtor's recovery simply by asserting a counterclaim. That approach would prejudice creditors, too, by preventing any eventual recovery on the debtor's claims from flowing to the bankruptcy estate. Cf. Martin-Trigona v. Champion Fed. Sav. & Loan Ass'n, 892 F.2d 575, 577 (7th Cir. 1989) (Posner, J.) (“The fundamental purpose of bankruptcy, from the creditors’ standpoint, is to prevent creditors from trying to steal a march on each other ․ There is, in contrast, no policy of preventing persons whom the bankrupt has sued from protecting their legal rights.”).

Reading Section 362(a)(1) to permit disaggregating cases into discrete proceedings against the debtor (which must be stayed) and those brought by the debtor (which need not) thus “best reflects the statute's text, structure, and legislative purpose.” Mar-Can Transp., 167 F.4th at 593 (typeface altered). Application of the rule to this case is straightforward: The district court's preliminary injunction order concerned only Cumulus's original claims. Nielsen did not even file its answer with counterclaims until after the notice of appeal had been docketed. Section 362(a)(1) thus does not require that we stay this appeal pending the conclusion of Cumulus's bankruptcy proceedings. And though we, like the district courts, retain “discretionary power to stay proceeding[s] in the interest of justice and in control of [our] dockets,” Teachers, 803 F.2d at 65, we discern no reason to delay decision in this appeal, which has been briefed and argued on an expedited schedule.⁴

II. Cumulus's Entitlement to a Preliminary Injunction

Nielsen's principal argument on appeal is that Cumulus did not satisfy the requirements for obtaining preliminary relief.

To obtain a preliminary injunction, plaintiffs must show(1) a likelihood of success on the merits or sufficiently serious questions going to the merits to make them a fair ground for litigation and a balance of hardships tipping decidedly in the plaintiffs’ favor, (2) that they are likely to suffer irreparable injury in the absence of an injunction, (3) that the balance of hardships tips in their favor, and (4) that the public interest would not be disserved by the issuance of a preliminary injunction.

Mendez v. Banks, 65 F.4th 56, 63–64 (2d Cir. 2023) (alterations accepted and quotation marks omitted). “The standard for obtaining preliminary injunctive relief is higher, however, where the movant seeks to modify the status quo by virtue of a mandatory preliminary injunction” or “where the injunction being sought will provide the movant with substantially all the relief sought and that relief cannot be undone even if the defendant prevails at a trial on the merits.” A.H. ex rel. Hester v. French, 985 F.3d 165, 176 & n.39 (2d Cir. 2021) (quotation marks omitted). The higher standard requires plaintiffs “show a clear or substantial likelihood of success on the merits and make a strong showing of irreparable harm.” Schneiderman, 787 F.3d at 650 (quotation marks and citation omitted).

At the outset, we agree with Nielsen and the district court that the higher standard for obtaining a mandatory preliminary injunction applies in this case. See Cumulus, 2026 WL 63294, at *8. “Because the proposed injunction's effect on the status quo drives the standard, we must ascertain the status quo—that is, the last actual, peaceable uncontested status which preceded the pending controversy.” N. Am. Soccer League, LLC v. U.S. Soccer Fed'n, Inc., 883 F.3d 32, 37 (2d Cir. 2018) (quotation marks omitted). Here, Cumulus sought, at the end of 2025, an order preliminarily enjoining Nielsen's Network Policy, which had been in effect since September 2024. Cumulus and Nielsen spent months negotiating under the policy until, in August 2025, Cumulus sent Nielsen a cease-and-desist letter alleging an antitrust violation. In other words, the last “uncontested status preceding the present controversy,” see Mastrio v. Sebelius, 768 F.3d 116, 121 (2d Cir. 2014), was the time the parties spent negotiating a new contract under the Network Policy, rather than before September 2024. Moreover, if Cumulus obtains interim relief but Nielsen ultimately prevails, it would be difficult if not impossible to render a meaningful remedy to Nielsen who would have had to negotiate its contract renewals this year without the Network Policy. See Tom Doherty Assocs. v. Saban Ent., Inc., 60 F.3d 27, 34–35 (2d Cir. 1995). Thus, in addition to being a mandatory preliminary injunction, the remedy Cumulus sought and obtained below may irreparably “alter th[e] relationship” between the parties. See N. Am. Soccer League, 883 F.3d at 37–38.

Cumulus thus sought, and the district court issued, a mandatory preliminary injunction that altered the status quo, potentially irreparably. So, we review the district court's decision to ensure Cumulus demonstrated “a clear or substantial likelihood of success on the merits” as well as “a strong showing of irreparable harm.” Schneiderman, 787 F.3d at 650 (quotation marks omitted).

A. Clear or Substantial Likelihood of Success on the Merits

The parties next dispute the district court's application of Section 2 liability for unlawful tying. The typical tying case involves “an agreement by a party to sell one product but only on the condition that the buyer also purchases a different (or tied) product, or at least agrees that he will not purchase that product from any other supplier.” Eastman Kodak Co. v. Image Tech. Servs., Inc., 504 U.S. 451, 461, 112 S.Ct. 2072, 119 L.Ed.2d 265 (1992) (quotation marks omitted). Motivating the prohibition is “[t]he fear ․ that a monopolist in one product market will seek to expand its monopoly by conditioning the purchase of the monopolized product upon the purchase of a product in a separate market,” thereby distorting competition in the tied market. Kaufman v. Time Warner, 836 F.3d 137, 141 (2d Cir. 2016). Put another way, an unlawful tying policy forces a buyer to purchase a tied product “that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms”—and, accordingly, “competition on the merits in the market for the tied item is restrained and the Sherman Act is violated.” Jefferson Par. Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 12, 104 S.Ct. 1551, 80 L.Ed.2d 2 (1984), abrogated on other grounds, Ill. Tool Works Inc. v. Indep. Ink, Inc., 547 U.S. 28, 126 S.Ct. 1281, 164 L.Ed.2d 26 (2006).

This dispute began as a typical tying case. The district court found that Nielsen's Network Policy—which expressly prohibits Cumulus from buying Nielsen's useable Nationwide product without also buying its local market data—constituted an unlawful “agreement by a party to sell one product but only on the condition that the buyer also purchases a ․ tied[ ] product[.]” Eastman Kodak, 504 U.S. at 461, 112 S.Ct. 2072; see also Cumulus, 2026 WL 63294, at *12. The district court found that the “tying” product is national radio data (a market in which Nielsen has 100% market share), while the “tied” product is the data in certain local markets. Nielsen does not challenge the district court's conclusion as to the lawfulness of its Network Policy on appeal.

After Cumulus sent Nielsen a cease-and-desist letter accusing it of anticompetitive behavior, Nielsen exempted Cumulus from its Network Policy and offered Cumulus a standalone price for its Nationwide product. However, the district court reasoned that this standalone offer still constituted a “constructive tie,” because “[t]he price that Nielsen offered Cumulus for a standalone Nationwide—ten times more than it currently pays—is so exorbitant as to make it economically unfeasible to purchase Nationwide as a separate product.” Cumulus, 2026 WL 63294, at *13. As explained below, Nielsen takes issue with the district court's understanding of this “constructive” tie on both the law and the facts.

To state a prima facie tying claim, Cumulus must establish five elements:

(i) the sale of one product (the tying product) is conditioned on the purchase of a separate product (the tied product); (ii) the seller uses actual coercion to force buyers to purchase the tied product; (iii) the seller has sufficient economic power in the tying product market to coerce purchasers into buying the tied product; (iv) the tie-in has anticompetitive effects in the tied market; and (v) a not insubstantial amount of interstate commerce is involved in the tied market.

Kaufman, 836 F.3d at 141. Then, “if a plaintiff successfully establishes a prima facie case[,] ․ the monopolist may proffer a ‘procompetitive justification’ for its conduct.” United States v. Microsoft Corp., 253 F.3d 34, 59 (D.C. Cir. 2001) (citing Eastman Kodak, 504 U.S. at 483, 112 S.Ct. 2072); see also Schneiderman, 787 F.3d at 652. Once the defendant asserts a “nonpretextual” procompetitive justification for a purported tie, the burden shifts to the plaintiff either to rebut the justification or to “demonstrate that the anticompetitive harm outweighs the procompetitive benefit.” Schneiderman, 787 F.3d at 652 (citing Microsoft Corp., 253 F.3d at 58–59).

The parties largely dispute elements (i) and (ii) of the prima facie case ⁵ : Nielsen claims that it cannot have “conditioned” the sale of Nationwide on its local data because “constructive” tying is not a viable theory of Sherman Act liability, and even if it were, it did not “coerce” Cumulus into purchasing its local data.

We are unpersuaded. Instead, we agree with the district court that constructive tying is a valid theory of Sherman Act liability. Further, on these facts, we find that the district court did not abuse its discretion by determining that Nielsen effectively coerced Cumulus into purchasing its local data in certain markets that Cumulus did not otherwise want. We also find that the district court did not abuse its discretion in concluding that Nielsen's conduct had anticompetitive effects in the tied market or in rejecting Nielsen's proffered procompetitive justification.

1. The Availability of Constructive Tying

We consider first whether a “constructive” tying arrangement can ever be unlawful under Section 2 of the Sherman Act. Nielsen contends that the only actionable form of tying requires an express tie between two products—that is, the seller's explicit refusal to sell one product (the tying product) without the other (the tied product). We disagree. Case law and common sense indicate that unlawful tying liability is not solely limited to express tying policies. Tying may also occur when a seller de facto ties two products via exorbitantly high prices, leaving the buyer with only one economically rational choice: to purchase the products together.

Two cases lead us to that conclusion. We begin with the Supreme Court's decision in United States v. Loew's, Inc., in which the Court recognized the power of price to perpetuate an unlawful (yet constructive) tying arrangement. 371 U.S. 38, 83 S.Ct. 97, 9 L.Ed.2d 11 (1962), abrogated on other grounds, Ill. Tool Works, 547 U.S. 28, 126 S.Ct. 1281. In Loew's, the United States sued several major film distributors for their “block booking” policies, which prohibited TV stations from licensing their films individually. Id. at 39–40, 83 S.Ct. 97. The government claimed that block booking constituted an unlawful tying arrangement; in effect, if a TV station sought to license one popular movie, the block booking policy forced it to purchase the licenses for dozens of unpopular films as well. Id. at 40, 83 S.Ct. 97. The Supreme Court ultimately held that block booking violated the Sherman Act. Id. at 49–50, 83 S.Ct. 97. But before the Court issued its decision, the distributors allegedly rescinded their block booking policies and began to offer their films at individual prices. Id. at 50–51, 83 S.Ct. 97. In light of that change, the distributors urged the Court to vacate the district court's injunction. Instead, recognizing the risk of a constructive tie even absent a formal block booking policy, the Supreme Court affirmed an injunction that barred the distributors from charging “a price differential between a film offered individually and as part of a package which ‘has the effect of conditioning the sale or license of such film upon the sale or license of one or more other films,’ ” see id. at 54, 83 S.Ct. 97 (emphasis added), and modified the injunction to specifically prohibit “noncost-justified” price differentials, id. at 52–54, 83 S.Ct. 97. The Supreme Court in Loew's thus identified—and foreclosed—the distributors’ obvious workaround to an anti-block booking injunction: selling their most popular movie individually for an exorbitant price and their bundles for the same prices, which would replicate the harms of the express tying arrangement. In short, the Supreme Court rejected the distributors’ attempts to avoid liability by facially discarding their express tying policy but continuing to de facto enforce it via high price differentials. That is precisely the theory of constructive tying that the district court employed here.

We, too, have recognized the availability of constructive tying liability under the Sherman Act. In American Manufacturers Mutual Insurance Company v. American Broadcasting-Paramount Theatres, Inc. (American Manufacturers I), we considered a dispute between the TV network ABC and plaintiffs who sought to purchase advertising time on some of its local stations. 388 F.2d 272, 274–76 (2d Cir. 1967). The plaintiffs alleged that ABC required sponsors to purchase advertisements on all local stations during a given time slot in order to advertise on any local stations—what the plaintiffs called an unlawful tying arrangement. Id. When the plaintiffs sought to purchase advertising time on only some local ABC stations, they claimed that the network would only sell those individual stations at “an unreasonable cost[.]” Id. at 276. The district court granted ABC's motion for summary judgment, id. at 278, but we reversed, “find[ing] it impossible on the record before us to determine whether ABC was justified in reverting to its higher ․ prices” for the individual stations, id. at 283. Instead, “where there is no quality or distinguishing desideratum between a product offered singly or in a package,” we said that “the seller cannot charge substantially higher for the individual product if the price differential has the effect of conditioning the sale of the single product to the sale of the entire package and if the difference in price cannot be legitimately justified by cost considerations.” Id. (citing Loew's, 371 U.S. at 43, 83 S.Ct. 97). That statement in American Manufacturers I squarely recognizes that price differentials can, in some instances, create an unlawful constructive tie.

Nielsen resists this conclusion by claiming that we later abandoned that determinative holding in American Manufacturers Mutual Insurance Company v. American Broadcasting-Paramount Theatres, Inc. (American Manufacturers II), 446 F.2d 1131 (2d Cir. 1971). But Nielsen's reliance is misplaced. The second American Manufacturers appeal reached us after “a three-day trial” on the question of whether ABC had in fact coerced the plaintiff into purchasing bundled local stations. Id. at 1133. The district court answered in the negative and we affirmed, holding that, on those facts, plaintiff had not proven that it “fe[lt] any economic pressure from ABC” to purchase the bundled stations and thus was not coerced. Id. at 1137. But nothing in American Manufacturers II suggests that we revisited our earlier holding about the general availability of constructive tying liability. If anything, the pair of appeals follows a standard arc: We recognized the viability of a constructive tying claim in American Manufacturers I, but after trial, we held that the plaintiff did not in fact establish such a claim in American Manufacturers II. The latter holding is entirely consistent with the former.

In light of the Supreme Court's decision in Loew's and our own in American Manufacturers I, we conclude that constructive tying is an available theory of antitrust liability. That conclusion makes good sense: “[T]he essential characteristic of an invalid tying arrangement lies in the seller's exploitation of its control over the tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms.” Jefferson Par., 466 U.S. at 12, 104 S.Ct. 1551. That “essential characteristic” of tying—the compulsion to purchase an unwanted product—can occur via an express tie, like block booking. Id.; see also Loew's, 371 U.S. at 48–49, 83 S.Ct. 97. But it can also occur via a constructive tie, such as non-cost justified, steep price differentials between an individual product and the tied bundle that have “the effect of” coercing the buyer into accepting the bundle. American Manufacturers I, 388 F.2d at 283. Both types of tying arrangements can produce the same effect. See Jefferson Par., 466 U.S. at 14, 104 S.Ct. 1551 (“When ‘forcing’ occurs, our cases have found the tying arrangement to be unlawful.”).

Constructive tying is a viable theory of Sherman Act liability. The remaining question is whether the district court abused its discretion by finding on this preliminary posture that Cumulus is clearly or substantially likely to succeed on its constructive tying claim.

2. Nielsen's Constructive Tie

The district court determined that Nielsen's single standalone offer for Nationwide “was priced so exorbitantly that this offer was the effective equivalent” of its formally tied offers and thus constituted a constructive tie. Cumulus, 2026 WL 63294, at *6, *13. Put another way, the district court found that Nielsen had continued to de facto enforce its unlawful Network Policy by offering to sell Nationwide only at an unreasonable cost. Nielsen argues that its standalone price for Nationwide was merely its opening offer for the product, which it says categorically cannot produce tying liability because opening offers never “actual[ly] coerc[e]” a buyer into purchasing the tied product. Kaufman, 836 F.3d at 141. But the district court determined on this preliminary record that Nielsen's steep price for Nationwide alone was not merely an opening offer, but instead an ongoing enforcement of its Network Policy that left Cumulus with no other rational choice but to purchase Nielsen's “tie[d]” local markets data as well. See Cumulus, 2026 WL 63294, at *13.

We agree. Reviewing the district court's factfinding for clear error and its conclusions of law de novo, see Zervos, 252 F.3d at 169, we conclude that Nielsen's non-cost justified and “exorbitant” standalone offer for Nationwide within the context of negotiations here rises to the level of coercion by effectively forcing Cumulus to purchase unwanted local markets data from Nielsen. Cumulus, 2026 WL 63294, at *13. The preliminary record suggests that Nielsen's “exemption” of Cumulus from its unlawful Network Policy was no exemption at all.

Determining the existence of coercion is a fact-bound inquiry. On the one hand, a seller's “use of strong persuasion, encouragement, or cajolery to the point of obnoxiousness to induce [the purchaser] to buy its full line of products does not ․ amount to actual coercion.” Unijax, Inc. v. Champion Int'l, Inc., 683 F.2d 678, 685 (2d Cir. 1982) (quotation marks omitted). On the other, if the seller “goes beyond persuasion and conditions [the] purchase of one product on the purchase of another,” we will say that the buyer has shown “[a]ctual coercion supporting a finding of a tying violation.” Id. Though the line between the two is not always clear, the district court's decision as to actual coercion is supported by the preliminary record here.

We begin with a brief survey of the relevant facts as found by the district court, nearly all of which are unchallenged on appeal and to which we defer unless clearly erroneous. See D.L. Cromwell Invs., 279 F.3d at 158. First, the parties agree that the relevant geographic markets are “the United States” (for national data) and “each local geographic area for which a ratings report is generated” (for local data).⁶ Cumulus, 2026 WL 63294, at *11. Nielsen enjoys 100% market share in the market for national radio data, but in certain local markets, it competes with Eastlan. See id. at *5, *8. Neither party disputes that Cumulus sought to purchase local markets data from Nielsen in only a subset of the 80 markets that Cumulus operates in; Cumulus intended to purchase data in at least some of those remaining local markets from Eastlan. Therefore, the allegedly “tied” products are Nielsen's data in the unwanted local markets.

Turn now to the parties’ negotiations. At the outset of the talks, Cumulus informed Nielsen that it was interested in purchasing Nielsen's Nationwide data along with its data in only a subset of those local markets. See Cumulus, 2026 WL 63294, at *5. Nielsen replied that its Network Policy barred Cumulus from purchasing anything less than full Nationwide and all relevant local markets, id. at *5–6, and it offered Cumulus that package at [Redacted] (the June offer). Because Nielsen does not challenge the district court's finding that the Network Policy is an unlawful (and express) tying arrangement, id. at *12, that June offer enforcing the policy was also unlawful. Nielsen then proceeded to make more offers enforcing the Policy by either continuing to include all relevant local markets—with the unwanted markets—or offering only the unusable “Swiss cheese” version of Nationwide along with the desired local markets. Months later, and under threat of litigation, Nielsen exempted Cumulus from its Network Policy and offered its standalone Nationwide product at a price of [Redacted], the relevant offer for our purposes. But the district court credited Cumulus's expert's assertion that the standalone offer was the “effective equivalent” of the unlawful June offer, concluding that the standalone offer “was priced so exorbitantly” that it would be economically irrational for Cumulus to also purchase its desired local data from Eastlan in the relevant local markets. Id. at *6, *13.

Those undisputed facts support the district court's legal conclusion that Nielsen's standalone offer was coercive even in the absence of a formal tie. Here, the district court found that the best price Cumulus could get for Eastlan's data in the desired local markets was [Redacted]. Id. But if Cumulus purchased Nielsen's Nationwide data for [Redacted] and Eastlan's local data for [Redacted], it would pay a total of [Redacted] for both products—$1.2 million more than Nielsen's unlawfully tied June offer. Thus, Nielsen's standalone offer gave Cumulus only an illusory choice: accept Nielsen's tied June offer, or pay $1.2 million more to purchase the desired local data from a competitor. As the district court found, that is no choice at all. Instead, given the course of dealing here, it found that Nielsen went “beyond persuasion” and actually conditioned the purchase of Nationwide on the purchase of the local markets data, Unijax, 683 F.2d at 685, which coerced Cumulus into purchasing data in local markets that it “either did not want at all” or “preferred to purchase elsewhere on different terms,” Jefferson Par., 466 U.S. at 12, 104 S.Ct. 1551. That coercion “impair[ed] competition on the merits” in the relevant local data markets by “harm[ing] existing competitors” (like Eastlan) and “creat[ing] barriers to entry of new competitors”—both results that “a free market would not tolerate.” Id. at 14–15, 104 S.Ct. 1551 (quotation marks omitted).

As for whether Nielsen's standalone Nationwide offer was justified by the cost of producing that product, the district court noted that “Nielsen has made no effort to quantify these costs or correlate them with its price for standalone Nationwide.” Cumulus, 2026 WL 63294, at *15. But the district court did find that the Nationwide-only offer “was exponentially more than what any other network pays for Nationwide as a standalone product”—indeed, the offer to Cumulus was 150% higher than the highest price Nielsen had ever previously charged for the product. Id. at *6. Moreover, the standalone offer for Nationwide was “ten times more than what Cumulus was paying for Nationwide under its existing contract.” Id. Though Nielsen may later “quantify [its] costs or correlate them with its price for standalone Nationwide,” id. at *15, the record as it stands supports the conclusion that “the difference in price cannot be legitimately justified by cost considerations,” American Manufacturers I, 388 F.2d at 283.

On this preliminary posture, given the evidence that Nielsen exploited its economic power in the national data market to force Cumulus into the purchase of tied products—Nielsen's data in several local markets—that it did not otherwise want, the district court did not abuse its discretion by finding that Cumulus was clearly or substantially likely to succeed on its tying claim. We recognize, of course, that “antitrust law does not prohibit lawfully obtained monopolies from charging monopoly prices.” Pac. Bell Tel. Co. v. LinkLine Commc'ns, Inc., 555 U.S. 438, 454, 129 S.Ct. 1109, 172 L.Ed.2d 836 (2009). But the difference here is that the district court found that Nielsen's conduct—namely, the non-cost justified offer for Nationwide—gave Cumulus no rational choice but to accept its unlawfully tied June offer. That conduct coerced the purchase of the tied local data products and restricted competition in those markets. See Verizon Commc'ns. Inc. v. L. Offs. of Curtis V. Trinko, LLP, 540 U.S. 398, 407, 124 S.Ct. 872, 157 L.Ed.2d 823 (2004) (“[T]he possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct.”). And it is precisely the type of “economic pressure” we have said is the hallmark of tying liability. American Manufacturers II, 446 F.2d at 1137.

Nielsen does not meaningfully dispute the above facts, but instead argues that its standalone Nationwide offer was merely its opening salvo in the negotiations, to which Cumulus was required to respond before filing suit. Indeed, on Nielsen's telling, an opening offer can never establish antitrust liability. While we agree that a party's “preliminary bargaining position” will typically not establish tying liability, because a true opening offer will rarely be coercive, see id., we disagree that Nielsen's standalone offer here—its fourth offer of eight—falls into that category. Instead, on these facts, we think the district court did not err by concluding that coercion occurred.

Nielsen's “opening offer” argument is not a freestanding element of, or defense to, tying liability. Instead, in American Manufacturers II, we described the parties’ negotiation history as part of the broader question as to whether the buyer has been coerced. See id. at 1136–37. There, we held that the plaintiff had failed to demonstrate coercion because it “did not persevere long enough” in negotiations for individual TV stations “to feel any economic pressure from ABC[.]” Id. at 1137. Nielsen relies heavily on American Manufacturers II to argue that Cumulus, too, did not “feel any economic pressure” from Nielsen's standalone Nationwide offer. Id. But we think American Manufacturers II is factually distinguishable. In that case, the plaintiff “never sought or received a reasonably firm ABC offer” for individually-priced stations. Id. Indeed, the plaintiff negotiated with ABC for months without ever indicating that it might want less than the full slate of stations, id. at 1133–34, and when it finally made an offer for individual stations, the price of its offer “did not differ significantly from the price offered by ABC,” a fact that that we called “significant,” id. at 1135. Then, after ABC's “initial negative response” to the plaintiff's individually-priced offer, the plaintiff “abandoned [the] issue ․ without any coercion by ABC” and focused on negotiating other contractual terms. Id. at 1135–36. Only after the plaintiff eventually purchased the bundled stations did it later assert that it had been coerced into such a purchase. Id. at 1136–37. On those facts, we found that the plaintiff had not demonstrated that ABC's refusal to individually price “ever crystallized into any identifiable or reasonably definitive policy,” and thus the plaintiff had not demonstrated “actual exertion of economic muscle” that “influence[d] the buyer's choice.” Id. at 1135, 1137. After all, although “the law does not demand that [plaintiffs] joust with windmills,” American Manufacturers I, 388 F.2d at 285, the plaintiff in American Manufacturers did not press its negotiating position enough to “discover[ ] whether it faced a windmill or a bona fide giant,” American Manufacturers II, 446 F.2d at 1137.

The current record in this case tells a very different story. At the outset of contract negotiations, Cumulus informed Nielsen that it was no longer interested in purchasing certain local data from Nielsen. See Cumulus, 2026 WL 63294, at *5. When Nielsen repeatedly declined to craft an offer exempting those unwanted local markets, Cumulus asked Nielsen for Nationwide pricing alone, which Nielsen refused to provide given its Network Policy. See id. at *5–6. Later, Cumulus counteroffered Nielsen for only its desired local markets, an offer that Nielsen rejected. Id. Even after all the above negotiating—in which Cumulus made unmistakably clear its desire to purchase local data in only a subset of markets—Nielsen twice restated offers to sell Cumulus its local radio ratings data in all 80 markets, including the unwanted ones, and explicitly pointed to its Network Policy in doing so. Unlike the plaintiff in American Manufacturers II, the negotiation history between the parties here suggests that Cumulus repeatedly pressed its desire to purchase Nationwide with only a subset of the local data markets, and Nielsen steadfastly refused to make such an offer. In that context, Nielsen's eventual exorbitant offer for standalone Nationwide was not a “bartering ploy[ ],” but a continuation of the same coercion. See American Manufacturers II, 446 F.2d at 1137. Our holding in American Manufacturers II reflects the straightforward conclusion that plaintiffs cannot be coerced by a contract term that they did not seriously dispute; here, the record reflects that the unwanted local data markets were the central disputed term from the outset.

In sum, our case law illustrates that a seller's high price can, in certain circumstances, constitute unlawful coercion that effectively “conditions [the buyer's] purchase of one product on the purchase of another.” Unijax, 683 F.2d at 685. In other circumstances, of course, such an offer will merely constitute “strong persuasion[.]” Id. (quotation marks omitted). But this is not one of those circumstances. To the contrary, the facts support the conclusion that Nielsen “actual[ly] exert[ed] [its] economic muscle” to coerce Cumulus into purchasing unwanted local data in several markets. American Manufacturers II, 446 F.2d at 1137. We therefore hold that, on this preliminary record, the district court did not abuse its discretion by concluding that Cumulus was likely to succeed in proving that Nielsen's standalone offer constituted an unlawful tie.

3. Alleged Anticompetitive Effects

Next, Nielsen argues that the district court erred in concluding that its proscribed conduct has anticompetitive effects in the local ratings data markets.⁷ Reviewing the district court's preliminary factual findings for clear error, we disagree.

Prevailing on a tying claim requires proof of “anticompetitive effects in the tied market.” E & L Consulting, Ltd. v. Doman Indus. Ltd., 472 F.3d 23, 31 (2d Cir. 2006) (quotation marks omitted). The requirement is satisfied if a plaintiff proves “that the tie impairs competition in the tied market and forecloses a substantial volume of commerce in that market.” Gonzalez v. St. Margaret's House Hous. Dev. Fund Corp., 880 F.2d 1514, 1517 (2d Cir. 1989). In other words, illegal tying arrangements are those that “restrain competition on the merits by forcing purchases that would not otherwise be made.” Jefferson Par., 466 U.S. at 27, 104 S.Ct. 1551. In conducting this inquiry, we “must focus on the market or markets in which the two products are sold, for that is where the anticompetitive forcing has its impact.” Id. at 18, 104 S.Ct. 1551.

On this score, the district court found that “[t]he record evidence overwhelmingly indicates that the Network Policy poses a significant barrier to entry, preventing Eastlan from achieving any measure of scale or industry-wide acceptance.” Cumulus, 2026 WL 63294, at *13. Because the Network Policy applies to “the largest national broadcasters,” id., Eastlan is deprived of the customers most likely to purchase its products in the local ratings data markets. And though Eastlan can theoretically target smaller broadcasters in local markets that are not subject to the Network Policy, “the record indicates that many such stations would not earn enough in advertising revenue to make purchase of Eastlan's product feasible.” Id. at *14. To illustrate the Network Policy's anticompetitive effects, the district court relied in part on evidence from the New Orleans market, where one smaller broadcaster not subject to the Network Policy “earned less in annual revenue in 2024 than the cost of Eastlan's local ratings data.” Id. On those findings, the district court concluded that Nielsen's Network Policy “has an adverse effect on the competitive process in the market for local radio ratings data.” Id. On appeal, Nielsen presses two objections.

First, Nielsen argues that the district court improperly focused on the Network Policy, which it claims to have exempted Cumulus from, and that “[t]he district court never found that Nielsen's standalone Nationwide offer ․ had anticompetitive effects.” Appellant's Reply Br. at 13. But the district court's analysis, albeit preliminary, was more comprehensive. It explained that “[t]he fact that Nielsen, after being accused by Cumulus of violating the antitrust laws, finally offered to sell Nationwide to Westwood One as a standalone product does not negate the anticompetitive effects of its Network Policy.” Cumulus, 2026 WL 63294, at *13 (citations omitted). That the standalone price was “so exorbitant as to make it economically unfeasible to purchase Nationwide as a separate product,” id., meant that the standalone offer was part of Nielsen's continued coercing of Cumulus into accepting Nielsen's combined offer. In other words, Nielsen's conduct amounted to continued enforcement of its Network Policy, and it follows that such conduct had the same anticompetitive effect on Cumulus's ability to purchase from Eastlan in the local ratings data markets. Nielsen neither challenges that factual finding on appeal nor explains why its pricing structure would cause different competitive effects than those the district court found for the Network Policy.

Second, Nielsen argues that even if the district court was right to focus on the Network Policy in looking for anticompetitive effects, it erred in considering evidence from only one local market—New Orleans—in doing so. The record says otherwise. Before referring to the New Orleans market as an “example” of anticompetitive effects, see Cumulus, 2026 WL 63294, at *14, the district court cited evidence about the “many markets in America” where large broadcasters subject to the Network Policy were effectively forced to purchase Nielsen's local ratings data, cutting out Eastlan, id. at *13 (quotation marks omitted). The dynamic has “ma[de] it difficult [for Eastlan] to grow in local markets” or to target them, the district court found, “preventing Eastlan from achieving any measure of scale or industry-wide acceptance.” Id. That the district court then further explained its findings with reference to a particular market hardly qualifies as clear error.

Thus, the district court did not abuse its discretion in concluding that Cumulus made a substantial showing of a prima facie tying claim in violation of Section 2.

4. Nielsen's Procompetitive Justification

Because we find no reversible error in the district court's preliminary analysis of Cumulus's prima facie claim, we turn to Nielsen's argument that it adequately “proffer[ed] nonpretextual procompetitive justifications for its conduct.” Schneiderman, 787 F.3d at 652 (quotation marks omitted). When a “monopolist asserts a procompetitive justification—a nonpretextual claim that its conduct is indeed a form of competition on the merits because it involves, for example, greater efficiency or enhanced consumer appeal—then the burden shifts back to the plaintiff to rebut that claim.” Microsoft, 253 F.3d at 59. If a defendant's procompetitive justifications are found to be pretextual, however, “we need not weigh them against the anticompetitive harms” of the defendant's conduct. Schneiderman, 787 F.3d at 658.

The district court rejected Nielsen's purported procompetitive justifications—preventing unlicensed sharing of national ratings data, generating efficiency for consumers in the form of a bundled deal, and recouping the primarily local-driven costs of producing Nationwide—on two grounds. First, it held that all three justifications were insufficiently supported by the record, and thus outweighed by the anticompetitive harms generated by the Network Policy. See Cumulus, 2026 WL 63294, at *14–15. Second, the district court held that the justifications were pretextual in light of Nielsen's motive to exclude competition. See id. On appeal, Nielsen retains only its third, cost-recoupment justification, arguing that it was legitimately motivated to recoup its costs and that the record does not support finding pretext. We disagree.

First, the district court did not clearly err in concluding that Nielsen did not sufficiently quantify its cost-recoupment justification. Instead, it concluded based on the preliminary record before it that Nielsen “made no effort to quantify these costs or correlate them with its price for standalone Nationwide.” Id. at *15. In this deferential posture, “[t]he weight of the evidence is not a ground for reversal on appeal, and the fact that there may have been evidence to support an inference contrary to that drawn by the trial court does not mean that the findings are clearly erroneous.” Ceraso v. Motiva Enters., LLC, 326 F.3d 303, 316 (2d Cir. 2003) (citation omitted). Nor has Nielsen mustered any record evidence to support its contention on appeal. Instead, it merely quotes the district court's acknowledgments that its “data-collection costs are sizable” and thus “aligning pricing with costs was a legitimate concern.” Appellant's Br. at 28 (quotation marks omitted). At this stage of the proceedings, that absence of evidence well supports the district court's conclusion. See Schneiderman, 787 F.3d at 659; see also LePage's Inc. v. 3M, 324 F.3d 141, 164 (3d Cir. 2003) (en banc) (“3M cites to no testimony or evidence in the 55 volume appendix that would support any actual economic efficiencies in having single invoices and/or single shipments.”). The district court thus did not err in holding that the “unquantified and hypothetical” procompetitive justifications proffered by Nielsen were outweighed by the “imminent and substantial” anticompetitive harms. Cumulus, 2026 WL 63294, at *15.

Second, and in any event, the district court did not clearly err in finding that Nielsen intended to exclude competition in the local ratings data markets, and thus that its purported procompetitive justifications were pretextual. See Cumulus, 2026 WL 63294, at *15. On this score, the district court cited an email from Rich Tunkel, Managing Director of Nielsen Audio, claiming that the Network Policy was intended to “command subscriptions in local markets” and “bring groups with non-subscribing markets back to the negotiation table.” Id. (quotation marks omitted); see also Appellee's Suppl. Sealed App'x 440. The district court also cited other record evidence containing similar statements from Tunkel, including that the Network Policy would “help reinforce the value of Nielsen local market measurement and secure local subscription.” Cumulus, 2026 WL 63294, at *15 (quotation marks omitted). From these statements, the district court did not clearly err in concluding that Nielsen intended to “put up barriers or obstacles to ․ competition” in the local ratings data markets. See Schneiderman, 787 F.3d at 658 (quotation marks omitted).

We are likewise unpersuaded by Nielsen's final argument that, at most, the record shows “mixed motives,” which it claims are insufficient to establish pretext. Appellant's Br. at 29. True, the district court acknowledged that some record evidence supported Nielsen's intention to recoup costs as well as “prevent networks from getting data through the back door.” Cumulus, 2026 WL 63294, at *15 (quotation marks omitted). But “[t]he fact that [Nielsen] acted to benefit its own economic interests is hardly a reason to overturn the ․ finding that it violated § 2 of the Sherman Act” precisely because “[i]t can be assumed that a monopolist seeks to further its economic interests and does so when it engages in exclusionary conduct.” LePage's, 324 F.3d at 164. Though Nielsen may have had some economic interest in the Network Policy and pricing for standalone Nationwide separate from excluding competition, the district court crediting such evidence does not undermine its findings of pretext, let alone preclude such findings. Indeed, Nielsen musters no argument that its conduct in any way enhanced competition or efficiency in the local ratings data markets. See Microsoft, 253 F.3d at 59 (noting “greater efficiency or enhanced consumer appeal” as examples of acceptable, nonpretextual procompetitive justifications).

Nor do Nielsen's cases advance its mixed-motive argument. Watson Laboratories, for example, concerned the pleading standard for a plaintiff to allege an “unjustified” reverse payment in violation of the antitrust laws, not the degree of anticompetitive intent sufficient to establish that a proffered procompetitive justification is pretextual. See Watson Lab'ys, Inc. v. Forest Lab'ys Inc., 101 F.4th 223, 239 & n.7 (2d Cir. 2024); see also AD/SAT v. AP, 181 F.3d 216, 231 (2d Cir. 1999) (describing insufficient allegations for a monopoly leveraging claim—without having alleged tying—as “normal business development” (quotation marks omitted)). And Ostrowski, concerning the standard for retaliation in an employment discrimination case, is even further afield. See Ostrowski v. Atl. Mut. Ins. Cos., 968 F.2d 171, 185 (2d Cir. 1992). Though we do not hold that the coexistence of procompetitive and anticompetitive motives are always insufficient to shift the burden back to the plaintiff in a Section 2 case, see Microsoft, 253 F.3d at 59, here the district court's reliance on Nielsen's repeated statements about excluding competition in the local ratings data markets well support the pretext finding.

Thus, we hold that the district court did not abuse its discretion in concluding that Cumulus made a substantial showing it was likely to succeed on the merits of its Section 2 tying claim.

B. Strong Showing of Irreparable Harm

Past likelihood of success, Nielsen argues Cumulus failed to make a strong showing of irreparable harm, and that at most Cumulus is likely to suffer “financial” injuries. Appellant's Br. at 29 (quotation marks omitted and alteration accepted). Though we do not agree with all of the district court's reasoning on this score, we conclude that Cumulus made such a showing.

A showing of irreparable harm is “the single most important prerequisite for the issuance of a preliminary injunction,” and for the issuance of a mandatory preliminary injunction, the showing must be “strong.” Daileader v. Certain Underwriters at Lloyds Lond. Syndicate 1861, 96 F.4th 351, 358 (2d Cir. 2024) (quotation marks omitted). “To establish irreparable harm, a party ․ must show that there is a continuing harm which cannot be adequately redressed by final relief on the merits and for which money damages cannot provide adequate compensation.” Kamerling v. Massanari, 295 F.3d 206, 214 (2d Cir. 2002) (quotation marks omitted). But though we often say in shorthand that “financial loss” does not constitute irreparable harm, see, e.g., Borey v. Nat'l Union Fire Ins. Co. of Pittsburgh, 934 F.2d 30, 34 (2d Cir. 1991), “a perhaps more accurate description of the circumstances that constitute irreparable harm is that where, but for the grant of equitable relief, there is a substantial chance that upon final resolution of the action the parties cannot be returned to the positions they previously occupied,” Brenntag Int'l Chems., Inc. v. Bank of India, 175 F.3d 245, 249 (2d Cir. 1999). It is against this backdrop that Nielsen's financial-injury-only argument comes apart.

The district court found credible Cumulus's representations that it would suffer “immediate consequences” were it to lose access to current national ratings data, including an inability to “develop credible advertising proposals” causing “long-standing” customers to “shift some or all of their advertising inventory purchases to competitors.” Cumulus, 2026 WL 63294, at *16. This, the district court determined, would “result[ ] in an immediate decrease in Westwood One's market share.” Id. In the district court's view, these types of harms can be irreparable for purposes of obtaining a preliminary injunction. Id. at *15 (quotation marks omitted). Nielsen disagrees, arguing that all of these harms “still trace back to a compensable overcharge that damages can address,” and that any risks of major disruption, loss of goodwill, or loss of market share are “merely speculative.” Appellant's Br. 30–31 (quotation marks omitted). We find Nielsen's arguments unavailing.

First, as a factual matter, we find no clear error in the district court crediting Cumulus's expectations of imminent loss of customers, good will, and market share. In particular, the district court credited Jones's declaration that switching some local ratings data purchases from Nielsen to Eastlan would be “crucial for maintaining sufficient cash for continued operations.” Appellee's Suppl. Sealed App'x 201. If Cumulus lost access to affordable, up-to-date ratings data, it would face both imminent risks from its creditors and an inability to retain longtime customers in a highly competitive advertising market with thin margins. See id. at 202–04. Nielsen's complaints that this evidence came “on the eve of an evidentiary hearing” from “the self-serving statements of an executive responsible for negotiating pricing with Nielsen,” Appellant's Br. at 31, go to weight, rather than clear error.

Second, on de novo review, we affirm the district court's legal conclusions that “loss of good will[,] customers, ․ [and] current or future market share suffices to show irreparable harm.” Cumulus, 2026 WL 63294, at *15 (quotation marks omitted). Though Nielsen protests that these types of injuries flow from financial loss, financial losses can be irreparable if “there is a substantial chance that upon final resolution of the action the parties cannot be returned to the positions they previously occupied.” Brenntag Int'l Chems., 175 F.3d at 249. Thus, we have held that loss of good will, loss of longtime customers, and a substantial loss in market share—although theoretically merely financial—can constitute irreparable harms. See, e.g., Reuters Ltd. v. United Press Int'l, Inc., 903 F.2d 904, 907–08 (2d Cir. 1990) (loss of good will and loss of customers); Grand River Enter. Six Nations, Ltd. v. Pryor, 481 F.3d 60, 67 (2d Cir. 2007) (loss of meaningful market share).

Nielsen's remaining arguments are equally meritless. Even though Cumulus may not have collapsed in the time since filing suit, the district court did not err in crediting evidence showing that Cumulus's business would face irreparable losses without interim relief. See Cumulus, 2026 WL 63294, at *15–16. And though Nielsen argues in reply that Cumulus's subsequent bankruptcy renders the preliminary injunction either insufficient or unnecessary, “the [d]istrict [c]ourt should determine in the first instance the effect of this supervening event upon the appropriateness of injunctive relief.” McLeod v. Gen. Elec. Co., 385 U.S. 533, 535, 87 S.Ct. 637, 17 L.Ed.2d 588 (1967).⁸ Nielsen is free to move to modify or vacate the preliminary injunction in the district court on the basis of Cumulus's intervening bankruptcy.

Finally, though we affirm the district court's finding of a strong likelihood of irreparable harm in the form of loss of customers, good will, and market share, we conclude that it erred in finding two other forms of irreparable harm. First, we do not agree with the district court that mere “[t]hreatened economic harm to consumers is plainly sufficient to authorize injunctive relief”—a conclusion the district court drew from an antitrust enforcement action maintained by state officials. Cumulus, 2026 WL 63294, at *15 (alteration adopted) (quoting Schneiderman, 787 F.3d at 661). In that context, consumer economic harm may well be irreparable because non-party consumers cannot recover damages, see, e.g., California v. Am. Stores Co., 495 U.S. 271, 295–96, 110 S.Ct. 1853, 109 L.Ed.2d 240 (1990), but the same is not necessarily true in cases between private parties. Second, we disagree with the district court that “a reduction in competition due to an antitrust injury also constitutes irreparable harm”—a conclusion drawn from cases about the public interest in antitrust enforcement more generally. Cumulus, 2026 WL 63294, at *16 (citing Consol. Gold Fields PLC v. Minorco, S.A., 871 F.2d 252, 257–58 (2d Cir. 1989), and F. & M. Schaefer Corp. v. C. Schmidt & Sons, Inc., 597 F.2d 814, 819 (2d Cir. 1979)). As the Supreme Court has held more recently, a “private litigant ․ must have standing—in the words of § 16, he must prove ‘threatened loss or damage’ to his own interests in order to obtain relief.” Am. Stores, 495 U.S. at 296, 110 S.Ct. 1853 (emphasis added). These errors notwithstanding, however, the district court did not ultimately err in concluding that Cumulus made a strong showing of irreparable harm in the form of lost consumers, decreased good will, and diminished market share.

C. Balance of the Equities and the Public Interest

Nielsen's next set of arguments concerns the remaining requirements for preliminary injunctive relief—“that the balance of equities tips in [the movant's] favor, and that an injunction is in the public interest.” Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20, 129 S.Ct. 365, 172 L.Ed.2d 249 (2008). Nielsen argues that it will be irreparably harmed by enforcement of the preliminary injunction because it will be forced to negotiate, with Cumulus and others, “at a competitive disadvantage.” Appellant's Br. at 35 (quotation marks omitted). Nielsen claims, also, that it will be unable to recoup the costs it incurs producing its local ratings data, and that the “breadth and vagueness” of the preliminary injunction risks “overcompliance.” Id. at 37, 129 S.Ct. 365 (quotation marks omitted). And Nielsen contends that the public interest will be served by vacating the preliminary injunction because, under the district court's order, “Nielsen cannot offer better terms to others without risking contempt, even if those customers stand in very different situations than Cumulus.” Id. at 39, 129 S.Ct. 365. We are unpersuaded.

First, in balancing the potential harms faced by both parties, the district court concluded that Nielsen's fears about losing negotiating leverage and cost recoupment “are purely speculative,” compared to the “short term” inability for Cumulus to “absorb the costs of Nielsen's policies.” Cumulus, 2026 WL 63294, at *16. While Cumulus demonstrated that it was likely to suffer irreparable harms in a matter of months, Nielsen musters only generalized concerns about its operations disconnected from any loss of business or irreparable financial strain. Even if Nielsen's “arguments evince plausible disagreement with the district court's balancing of the hardships,” “the balance reached by the district court falls squarely within its discretion.” State Farm Mut. Auto. Ins. Co. v. Tri-Borough NY Med. Prac. P.C., 120 F.4th 59, 85 (2d Cir. 2024).

Second, the district court did not err in concluding that the public interest is served by preliminary enforcement of the antitrust laws. See Cumulus, 2026 WL 63294, at *16. We regularly hold that the public benefits from preliminarily enjoining conduct found likely to be anticompetitive. See, e.g., Schneiderman, 787 F.3d at 662 (collecting cases). Nielsen's arguments to the contrary assume that its conduct is not anticompetitive; Nielsen does not dispute that enforcing the antitrust laws serves the public. And Nielsen's backup argument, that its other customers will be required to pay higher prices for Nationwide else Nielsen risks setting the presumptively reasonable price too low, is entirely speculative and unsupported by any record evidence. Whatever influence the preliminary injunction has on Nielsen's other contract negotiations is hypothetical, unquantified, and outweighed by the public's interest in the maintenance of a competitive market for local ratings data.

We hold that the district court did not abuse its discretion in concluding that Cumulus satisfied the requirements to obtain a mandatory preliminary injunction.

III. Propriety of the Preliminary Injunction

Finally, Nielsen takes issue with the design of the district court's injunction. We review this challenge de novo. See Garcia v. Yonkers Sch. Dist., 561 F.3d 97, 103 (2d Cir. 2009). The injunction reads, in relevant part:

The Court hereby ORDERS ․ that (i) Nielsen (including Nielsen's officers, employees, and agents) is enjoined and restrained from enforcing its Network Policy; and (ii) Nielsen is enjoined and restrained from charging a commercially unreasonable rate for its Nationwide Report as a complete, standalone product. For purposes of this Order, a rate that is equal to or lower than the highest annual 2026 rate Nielsen charges any broadcaster (whether network or local) for Nationwide is presumptively reasonable.

Cumulus, 2026 WL 63294, at *18. Nielsen specifically disputes the latter half of the order, which prohibits it from charging “a commercially unreasonable rate” for standalone Nationwide. Id. It claims that this prohibition is insufficient for two reasons: it both violates the specificity requirement of Federal Rule of Civil Procedure 65(d) and fails to remedy Cumulus's claimed irreparable harm. We disagree with both challenges.

First, Rule 65(d). That Rule requires “[e]very order granting an injunction” to “describe in reasonable detail ․ the act or acts restrained or required.” Fed. R. Civ. P. 65(d)(1). We refer to this mandate as the “specificity requirement.” Sanders v. Air Line Pilots Ass'n, Int'l, 473 F.2d 244, 247 (2d Cir. 1972). “The normal standard of specificity [under the Rule] is that the party enjoined must be able to ascertain from the four corners of the order precisely what acts are forbidden.” Id. Compliance with the specificity requirement is particularly important because “basic fairness requires that those enjoined receive explicit notice of precisely what conduct is outlawed” before they risk incurring contempt. Schmidt v. Lessard, 414 U.S. 473, 476, 94 S.Ct. 713, 38 L.Ed.2d 661 (1974).

Nielsen argues that it lacks explicit notice of what constitutes an “unreasonable rate” for Nationwide. While we note that similar “reasonable rate” injunctions have been upheld before, see, e.g., United States v. Glaxo Grp. Ltd., 410 U.S. 52, 64, 93 S.Ct. 861, 35 L.Ed.2d 104 (1973), we need not decide whether a prohibition on charging an “unreasonable” price may be too ambiguous on its own, because we find that the district court's safe harbor provision lends this injunction sufficient clarity. The safe harbor provision states that “a rate that is equal to or lower than the highest annual 2026 rate Nielsen charges any broadcaster (whether network or local) for Nationwide is presumptively reasonable.” Cumulus, 2026 WL 63294, at *18. Nielsen can thus determine a presumptively reasonable rate to charge Cumulus by surveying the prices it charges other large network clients, either under current contracts or based on contracts negotiated this year.⁹ While Nielsen is concerned that it will not know its highest 2026 rate for Nationwide until after the year is through, counsel for Cumulus noted at oral argument that contracts for services in 2026 are typically negotiated well in advance. Moreover, Nielsen failed to describe in its briefs or at oral argument what a more specific version of the district court's injunction would look like. That inability to articulate a more precise injunction further demonstrates that the district court's remedy does fulfill the specificity requirement of Rule 65(d). Finally, we are not persuaded by Nielsen's argument that relief in this context cannot regulate a product's price, since “[t]o ensure ․ that relief is effectual, otherwise permissible practices connected with the acts found to be illegal must sometimes be enjoined.” Loew's, 371 U.S. at 53, 83 S.Ct. 97.

We conclude that the district court's injunction does not violate the specificity requirement of Rule 65(d). A district court need not “describe all possible, permissible future” prices that Nielsen can charge in order to satisfy that Rule; instead, a district court's injunction must merely “assist [Nielsen] in determining whether a proposed” price is unreasonable. See S.C. Johnson & Son, Inc. v. Clorox Co., 241 F.3d 232, 241 (2d Cir. 2001). The district court's safe harbor provision does just that.

Finally, we disagree with Nielsen that the district court's injunction will not remedy Cumulus's claimed harms. In effect, Nielsen argues that only a forced sale of Nationwide will certainly alleviate the irreparable harm that flows from Cumulus's impending loss of access to that product. But Cumulus alleges precisely that the Network Policy impeded its ability to properly negotiate and thereby obtain that product. In the Sherman Act context, “the precise practices found to have violated the Act should be specifically enjoined.” United States v. Grinnell Corp., 384 U.S. 563, 579–80, 86 S.Ct. 1698, 16 L.Ed.2d 778 (1966) (emphasis added). Here, the district court's injunction is tailored to the conduct it found to be unlawful: Nielsen's express and constructive enforcement of its Network Policy. Prohibiting Nielsen from engaging in “the precise practices found to have violated the Act” will assuage the serious risk of Cumulus's harm coming to pass by barring Nielsen from engaging in anticompetitive conduct during negotiations. Id. We think the district court's injunction is sufficiently tailored to remedy Cumulus's harm.

CONCLUSION

This interlocutory appeal raises a number of complicated and novel issues. We resolve many of them not for all time, but on a preliminary basis and with considerable deference to the district court's factual findings and discretionary decisions. Still, we hold as a threshold matter that we may proceed with this appeal notwithstanding Nielsen's counterclaims against Cumulus below, because Section 362(a)(1) of the Bankruptcy Code automatically stays only discrete claims against debtors, rather than the entire actions in which they are asserted. Further, we hold as a matter of law that “constructive tying”—the pricing of two products that has the effect of conditioning the sale of one product on the other—can in some cases violate Section 2 of the Sherman Act. Finally, we reject Nielsen's several remaining charges of error, and we hold instead that the district court properly (1) found that Cumulus was coerced based on Nielsen's constructive tie, (2) found anticompetitive effects of Nielsen's conduct, (3) discounted Nielsen's procompetitive justification, (4) concluded that Cumulus made a strong showing of irreparable harm absent interim relief, (5) weighed the relative hardships and public interest in Cumulus's favor, and (6) issued an appropriately specific and tailored preliminary injunction.

The December 30, 2025 order of the district court granting Cumulus's motion for a preliminary injunction is AFFIRMED. The February 3, 2026 order of this court granting Nielsen's motion for a stay is VACATED, and the case is REMANDED for further proceedings consistent with this Opinion.

FOOTNOTES

1.   Nielsen's Network Policy applies only to customers who operate both a national broadcasting network and local radio stations. That comprises 12 customers, including “the three largest companies in radio,” which together control about a third of all radio advertising spend. Cumulus, 2026 WL 63294, at *4.

2.   Cumulus also asserted five other causes of action in its complaint, none of which are relevant for the purposes of this interlocutory appeal.

3.   In doing so, we join the majority of circuit courts to have considered the question. See Mar. Elec. Co. v. United Jersey Bank, 959 F.2d 1194, 1205 (3d Cir. 1991); U.S. Abatement Corp. v. Mobil Expl. & Producing U.S. (In re U.S. Abatement Corp.), 39 F.3d 563, 568 (5th Cir. 1994); Parker v. Bain, 68 F.3d 1131, 1137 (9th Cir. 1995); Seiko Epson Corp. v. Nu-Kote Int'l, Inc., 190 F.3d 1360, 1364 (Fed. Cir. 1999); Lehman v. Revolution Portfolio L.L.C., 166 F.3d 389, 392 n.5 (1st Cir. 1999); In re Hall, 304 F.3d 743, 746 (7th Cir. 2002); Thomas v. Blue Cross & Blue Shield Ass'n, 333 F. App'x 414, 420–21 (11th Cir. 2009) (summary order); cf. Dominic's Rest. of Dayton, Inc. v. Mantia, 683 F.3d 757, 760–61 (6th Cir. 2012).

4.   Though we decline to exercise our discretion to stay this appeal, we acknowledge that the district court's contrary decision concerned different considerations, and that the district court is free to revisit that decision at any time on remand.

5.   The district court explained that “[t]hree of the five prongs of the tying test are unchallenged,” and that at issue was coercion and anticompetitive effects in the tied market. Cumulus, 2026 WL 63294, at *12–13. On appeal, Nielsen primarily challenges the viability of a constructive tying theory—that is, whether there can be such a thing, and if so, whether the constructive tie was coercive in this case—thus implicating the first two prongs as well. As to the district court's findings on anticompetitive effects, Nielsen asserts that such analysis was beside the point because it was premised on an express rather than constructive tie. See Appellant's Br. at 27–28.

6.   “For antitrust purposes, the concept of a market has two components: a product market and a geographic market.” Concord Assocs., L.P. v. Ent. Props. Tr., 817 F.3d 46, 52 (2d Cir. 2016). The product markets in this case are for “local radio ratings data” and “national radio ratings data.” Cumulus, 2026 WL 63294, at *10. At times, the district court uses the shorthand “the local market” to describe the separate geographic markets here—i.e., “each local geographic area for which a ratings report is generated,” id. at *11—but it is clear from its analysis that it ultimately concluded each individual local market in which Cumulus did not seek to purchase Nielsen's data had been effectively “tied” by Nielsen's Network Policy.

7.   The district court held that, because Nielsen is an undisputed monopolist in the market for nationwide ratings data, its constructive tie is illegal per se without a showing of anticompetitive effect. See Cumulus, 2026 WL 63294, at *13 (citing Jefferson Par., 466 U.S. at 17, 104 S.Ct. 1551). Because we find no error in the district court's alternative holding—that Nielsen's Network Policy and standalone Nationwide offer had anticompetitive effects in the local ratings data markets, see id. at *13–14—we need not and do not reach the question whether tying can constitute a per se violation of Section 2.

8.   To be clear, the district court's findings of irreparable harm remain relevant to our analysis, Cumulus's bankruptcy filing notwithstanding. While bankruptcy is one basis for finding irreparable harm, see Tucker Anthony Realty Corp. v. Schlesinger, 888 F.2d 969, 975 (2d Cir. 1989), it does not necessarily negate or subsume the other bases asserted here. For example, customer good will continues to exist—and can be a relevant consideration—during bankruptcy proceedings themselves. Cf. In re Windstream Holdings, Inc., 105 F.4th 488, 497 (2d Cir. 2024).

9.   The record demonstrates that Cumulus is not Nielsen's only large network client for Nationwide. While Nielsen claims that “Cumulus is a significantly larger customer than any other that purchases the standalone Nationwide report,” Appellant's Br. at 44, the district court found only that Cumulus is one of “the three largest companies in radio,” Cumulus, 2026 WL 63294, at *4, and further found that at least one of those three large companies (apart from Cumulus) also purchases Nationwide from Nielsen, see id. at *6.

Nathan, Circuit Judge: