Overview
Introduction
Antitrust liability can extend to companies that exchange competitively sensitive information through an intermediary even if there is no direct communication between the companies, as algorithmic methods of transmission do not necessarily confer immunity. Companies exchanging information through an intermediary or an algorithm should be wary of what information is being used, whether that information is public, how frequently the information is collected, and the granularity of the data being used as an input. In the United States, guidance on these issues can be found in case law and consent decrees with antitrust enforcement agencies, and in Korea, guidance can be found in legislation, with a 2020 statutory amendment expressly covering information exchange and a first enforcement decision issued in early 2026. The emerging lesson in both jurisdictions is that the method of transmission itself is not the central antitrust issue. The more important questions are whether competitors contribute or receive non-public competitively sensitive information, whether the arrangement reduces uncertainty about rivals’ conduct, and whether it displaces independent competitive decision-making. As information-sharing mechanisms evolve from direct communications, to benchmarking platforms, to algorithmic pricing tools, those traditional antitrust principles remain the analytical anchor.
Traditional Competitor Information Exchange
Direct competitor communications often face antitrust scrutiny across jurisdictions. In the United States, no statute expressly prohibits competitor communications, but direct competitors have nonetheless been found liable under the antitrust laws when exchanging competitively sensitive information with each other.[1] Additionally, exchanges of competitively sensitive information between direct competitors can be used as evidence of an illegal conspiracy to fix prices.[2] Antitrust law is particularly concerned with the exchange of price information between competitors.[3] Communications between horizontal competitors are generally viewed with greater antitrust concern than communications between parties in a vertical relationship.[4]
While the US has addressed information-sharing through case law, Korea reached the same concern by a different route: legislation. Korea's Monopoly Regulation and Fair Trade Act (the MRFTA) prohibits cartels through an enumerated list of the subjects of agreement (price, output, and so on), and before 2020 an agreement merely to exchange information was not itself a listed type. Korean courts required strict proof of an agreement on an enumerated subject, and a series of information-exchange cases failed on that ground, a history the Korea Fair Trade Commission (the KFTC) itself recounts as the direct background of reforms to the MRFTA.[5] The 2020 wholesale amendment of the MRFTA (in force from December 30, 2021) resolved the issue by statute: it added an agreement to exchange price, output and other key competitive information as a distinct enumerated type of "unfair collaborative act," the MRFTA’s statutory term for cartel conduct (Article 40(1)(9)), and separately introduced a statutory presumption of agreement where parallel conduct coincides with information exchange (Article 40(5)(2)).[6]
The first decision under the new provision came in early 2026, when the KFTC imposed corrective orders and fines totaling approximately KRW 272 billion (US $192 million) on four major commercial banks for agreeing to exchange granular loan-to-value data on real-estate collateral.[7] The exchange itself was entirely traditional in form: staff met in person and handed over printouts covering thousands of individual data points, on request and without a single recorded refusal, from which the KFTC found a meeting of minds to exchange. The decision reasons in terms familiar to US readers: information-exchange agreements, unlike direct price-fixing, restrict competition by reducing uncertainty about competitors’ conduct and strategies and weakening competitive pressure; on that basis, the KFTC held that no separate showing of parallel pricing is required under Article 40(1)(9) once the exchange agreement and its restrictive effect are established.[8] The banks have challenged the decision in court, and it is not yet final. These traditional information-exchange cases establish the baseline for analyzing newer technologies. The antitrust concern does not necessarily disappear merely because competitors replace direct communications with a common intermediary. Instead, the intermediary changes the mechanism through which information flows – and can make the exchange more systematic, granular, and frequent.
Benchmarking Platforms and the Rise of the Information Intermediary
The rise of businesses that collect and distribute data in reports to participants in particular industries has made it easier for companies to share information through an intermediary without ever communicating directly. Companies that share competitively sensitive information through an intermediary still face liability even if they do not share information directly with their competitors. For example, in the United States, the US Department of Justice (DOJ) and several States sued meat industry data analysis firm Agri Stats, alleging that Agri Stats operated as an intermediary that facilitated coordination of competitors' prices in violation of the antitrust laws.[9] The Complaint alleged that Agri Stats collected competitively sensitive information related to price, cost, and output between meat processors and distributed comprehensive, weekly reports based on that information between the processors, effectively providing them with pricing data from their competitors.[10]Particularly problematic was providing non-public price and sales data at a facility level on a weekly basis, enabling competing processors to adjust output, pricing, or both based on near-total visibility into their rivals’ operations.[11]The DOJ and States alleged that the sharing of this information facilitated coordination between direct competitors in violation of the Sherman Act even without direct competitor contact. Though the DOJ and States settled with Agri Stats in May 2026, requiring Agri Stats to stop providing any sales reports or non-public pricing information and to stop reporting production, cost, and labor data pending final approval from the court,[12]enforcers and plaintiffs continue to litigate cases using an intermediary with a focus on data granularity, frequency, and the intermediary's role in reducing market uncertainty.
Korea has no enforcement decision or court precedent yet on benchmarking-platform-type exchange, but the framework already reaches intermediaries. Alongside the amended MRFTA's entry into force, the KFTC adopted review guidelines that expressly treat transmission of information through an intermediary, including a trade association or a third-party undertaking, as information exchange.[13] More recently, the KFTC's "Data and Competition" policy report, published in December 2025, discusses data pooling and data sharing through third-parties as emerging competition issues, while noting, based on a written survey conducted for the KFTC, that data pooling and similar new arrangements are not yet widespread in Korea.[14] For companies operating in Korea, in other words, the absence of a Korean Agri Stats is a matter of enforcement history, not of legal immunity: the statutory and guideline basis for pursuing intermediary-based exchange is already in place.
AI Algorithms and the RealPage Litigation
Recent advancements in technology allow third party data companies to more effectively recommend prices, and often in real-time. Where companies like Agri Stats compiled data and distributed reports to companies for companies to set their own prices, some third-party data companies now deploy algorithmic software, sometimes built on top of artificial intelligence. Algorithmic software, particularly software that relies on artificial intelligence, goes a step further than simply producing reports: it will recommend prices, often in real-time, and occasionally based on non-public competitor data. While using algorithms and artificial intelligence for setting prices typically does not run afoul of antitrust laws in and of itself, there is a risk that the use of such software could be considered collusive conduct if it allows competitors to set prices based on competitively sensitive pricing information.
For example, in the United States, the DOJ investigated and sued RealPage, a technology platform that provides data analytics for the real estate industry, alleging that the company’s algorithm that recommends apartment rental prices based on competitors’ pricing information amounted to a conspiracy by competing firms to coordinate prices through a shared intermediary. The DOJ reached a consent decree with RealPage in November 2025 requiring that the company use only historical rental data at least twelve months old – taking the data out of the “real-time” dynamic pricing realm – and that it not report rental pricing information more narrowly than at the statewide level.[15]
Most recently, a federal appellate court reversed the dismissal and remanded a suit brought against hotels in the resort area of Atlantic City, New Jersey that fed sensitive data to algorithmic pricing software called Rainmaker, which then suggested prices that the hotels accepted 90% of the time.[16] The court held that the use of the software in tandem with the hotels acting against their economic interest through upward pricing and downward room occupancy, was sufficient to show conscious parallel conduct.[17] Notably, this was a split from another federal appellate court, which affirmed dismissal of a suit alleging similar facts against hotels that used the same Rainmaker software in Las Vegas, Nevada.[18]
Because of this split between appellate courts and because the RealPage suit ended in a settlement, there remains an unresolved question of when reliance on a shared algorithm becomes unlawful coordination in the United States. When there are differing federal appellate court decisions, the law is often later clarified by the US Supreme Court, but that has not yet happened. The DOJ recently provided guidance that it would consider criminal probes of companies that use algorithmic pricing software if the companies know that nonpublic data are used by the software to set prices. Acting Deputy Assistant Attorney General Daniel Glad said that "the agreement among competitors[] is the element that determines whether this is a vertical arrangement . . . or a horizontal conspiracy subject [opening] the door to . . . to criminal enforcement."[19] Korea again has no precedent on algorithmic pricing coordination, but the question is squarely on the regulator's desk. The KFTC’s December 2025 policy report flags the gap directly: under the current guidelines, purely one-way transmission of data to an intermediary falls outside "exchange," and the report, citing the DOJ's RealPage suit itself, notes that coordination through a common intermediary that collects data and returns pricing recommendations requires further review.[20] The report records that algorithm-based collusion is the subject of a separately commissioned KFTC study, and the KFTC's 2025 annual work plan designates AI-enabled collusion among the new types of collusion slated for in-depth analysis and the development of response measures.[21]
A Common Theme Across All Three Models
Across direct exchanges, benchmarking platforms, and algorithmic pricing tools, the central antitrust concern is remarkably consistent: arrangements that give competitors access to non-public competitively sensitive information can reduce uncertainty about rivals' behavior and facilitate coordination. Technology changes the mechanism, but not necessarily the underlying competition concern. A conversation between competitors, a weekly benchmarking report, and an algorithmic pricing recommendation may look very different operationally, yet each can raise antitrust concerns when it allows rivals to coordinate rather than make genuinely independent competitive decisions.
Regardless of jurisdiction, companies should focus on how they share and receive competitively sensitive information: train employees not to discuss prices, output, new product lines, and other commercially sensitive issues with competitors; be cautious when receiving reports or using software that relies on proprietary, non-public pricing information from competitors; when using pricing algorithms, understand where the underlying inputs come from and how they are used. At a minimum, a company should determine what information is being used, what information the company is sharing, whether that information is public, how frequently the information is collected, and the granularity of the data being used as an input. Put differently, if the information used as an input would result in antitrust liability if received directly from a company, then using these reports or software could result in antitrust liability as well. That checklist travels well: because both US case law and Korean law and guidelines already cover exchange carried out through an intermediary, the same questions apply to companies receiving such reports or using such software in both countries.
[1] See, e.g., United States v. United States Gypsum Co., 438 US 422, 441 n.16 (1978); United States v. Citizens & S. National Bank, 422 US 86, 113 (1975).
[2] In re Flat Glass Antitrust Litig., 385 F.3d 350, 368-69 (3d Cir. 2004) (jury could infer that exchange of pricing information between upper hierarchy of glass producers was concerted action designed to fix prices); In re Petroleum Prods. Antitrust Litig., 906 F.2d 432, 445-50 (9th Cir. 1990) (exchanges of price information treated as plus factor from which jury could infer agreement among rivals to fix prices)
[3] United States v. Container Corp. of America, 393 US 333, 337 (1969).
[4] See Toys "R" Us, Inc. v. FTC, 221 F.3d 928, 930 (7th Cir. 2000).
[5] Korea Fair Trade Commission (Full Committee), Decision No. 2026-019, Case No. 2023Kachong0733, In re Unfair Collaborative Acts of Four Commercial Banks, Feb. 13, 2026 (announced Jan. 21, 2026) [hereinafter Four Banks Decision], paras. 121-123 & nn.55-56 (recounting the pre-amendment enforcement record as the direct background of the reform). The failed cases include Supreme Court of Korea, Decision 2013Du16401, July 24, 2014 (life-insurance interest rates), and Decisions 2013Du25924, Dec. 24, 2015, and 2013Du26309, Jan. 14, 2016 (instant noodles). The Four Banks Decision is subject to pending annulment litigation and is not yet final.
[6] MRFTA, wholly amended by Act No. 17799, Dec. 29, 2020 (effective Dec. 30, 2021) (S. Kor.), arts. 40(1)(9), 40(5)(2). Article 40(1)(9) prohibits, as a distinct enumerated type of unfair collaborative act, an agreement to exchange price, output, cost or transaction-term information prescribed by the Enforcement Decree, where the agreement unduly restricts competition; Article 40(5)(2) provides that an agreement may be presumed where undertakings engage in outwardly consistent conduct and have exchanged the information necessary for that conduct. The categories of covered information are set out in Article 44(2) of the Enforcement Decree of the MRFTA, wholly amended by Presidential Decree No. 32274, Dec. 28, 2021 (effective Dec. 30, 2021).
[7] Four Banks Decision, supra note 5.
[8] Four Banks Decision, supra note 5, paras. 249-251 (finding a meeting of minds from the parties' sustained course of on-request exchange), paras. 259-261 (setting out the theory of harm and holding that outward conformity of prices need not be separately established under Article 40(1)(9)).
[9] Justice Department Sues Agri Stats for Operating Extensive Information Exchanges Among Meat Processors, Dep't of Justice (Sept. 28, 2023), https://www.justice.gov/archives/opa/pr/justice-department-sues-agri-stats-operating-extensive-information-exchanges-among-meat.
[10] Justice Department Sues Agri Stats for Operating Extensive Information Exchanges Among Meat Processors, Dep’t of Justice (Sept. 28, 2023), https://www.justice.gov/archives/opa/pr/justice-department-sues-agri-stats-operating-extensive-information-exchanges-among-meat.
[11] Justice Department Requires Agri Stats to End Exchange of Competitively Sensitive Information Among Nation's Largest Meat Processors that Suppressed Competition and Increased Prices for Decades, Dep't of Justice (May 7, 2026), https://www.justice.gov/opa/pr/justice-department-requires-agri-stats-end-exchange-competitively-sensitive-information; Justice Department Sues Agri Stats for Operating Extensive Information Exchanges Among Meat Processors, Dep't of Justice (Sept. 28, 2023), https://www.justice.gov/archives/opa/pr/justice-department-sues-agri-stats-operating-extensive-information-exchanges-among-meat.
[12] Nation's Largest Meat Processors that Suppressed Competition and Increased Prices for Decades, Dep’t of Justice (May 7, 2026).
[13] KFTC, Review Guidelines on Unfair Collaborative Acts Involving Information Exchange Between Enterprises, KFTC Public Notification No. 2021-33, enacted Dec. 28, 2021 (effective Dec. 30, 2021), § III (providing that information exchange includes indirect transmission through an intermediary, such as a trade association or a third-party undertaking).
[14] KFTC, Data and Competition Policy Report, Dec. 2025 [hereinafter Data and Competition Report], § 3.5 (Data-Related Unfair Collaborative Act Issues), at 76-79; see esp. paras. 188-190, at 77-78 (discussing data pooling as an emerging issue and reporting the results of a written market survey)
[15] United States v. RealPage, Inc., No. 1:24-cv-00710-WO-JGM (M.D.N.C. May 19, 2026) (final judgment) (Doc. 194).
[16] Cornish-Adebiyi v. Caesars Ent., Inc., No. 24-3006, 2026 WL 2182291, at *1, 5, 7 (3d. Cir. July 29, 2026).
[17] Id. at *8.
[18] Gibson v. Cendyn Grp. LLC, 148 F.4th 1069, 1076-77 (9th Cir. 2025).
[19] Acting Deputy Assistant Attorney General for Criminal Enforcement Daniel Glad Delivers Remarks at the Antitrust West Coast Conference, Dep’t of Justice (May 14, 2026), https://www.justice.gov/opa/speech/acting-deputy-assistant-attorney-general-criminal-enforcement-daniel-gladd-delivers.
[20] Data and Competition Report, supra note 14, para. 192 & n.176, at 79 (noting that purely one-way transmission of data to an intermediary falls outside the current guidelines and citing the US Department of Justice's RealPage litigation); see also id. § 3.5 n.169, at 76 (recording that algorithm-based collusion is addressed in a separately commissioned KFTC study).
[21] KFTC, 2025 Annual Work Plan, Jan. 2025, pt. III.2.A (Improving Collusion and Unfair Trading Practices), at 6 (designating "new types of collusion," including collusion using AI, for in-depth analysis of cases, legal doctrine and overseas developments, and for the preparation of response measures).
