On May 7, 2026, the U.S. Department of Justice (DOJ) Antitrust Division and six state attorneys general filed a proposed Final Judgment to resolve antitrust claims against benchmarking and consulting firm Agri Stats, Inc. (Agri Stats). Agri Stats collects, processes, and distributes benchmarking data to broiler chicken, pork, and turkey processing operations. The reports are based on non-public pricing, profitability, and cost data submitted by meat processors. While Agri Stats purported to aggregate and anonymize the data, the DOJ alleged that those receiving the reports could identify who submitted the non-public data, allowing participants to raise prices.

The terms of the Agri Stats settlement shed light on how the agencies may approach industry information hubs and shared pricing algorithms, particularly set against settlements concerning similar conduct in the meat processing industry and the DOJ’s settlement in the RealPage case.

Background and Settlement Terms

Reports from Agri Stats and other meat processing information hubs providing similar benchmarking have been subject to an array of private suits and government enforcement for a decade (described in more detail in our prior client alert). The DOJ brought its case against Agri Stats in 2023.¹ According to the DOJ, Agri Stats reports allowed participating processors to identify competitors and to align their pricing, output, and production decisions accordingly. The DOJ also alleged parties that did not contribute data were denied access to the associated reports by Agri Stats. The DOJ announced and filed a settlement on May 7, 2026.²

Under the terms of the settlement, Agri Stats will be required to (among more standard terms):

• cease offering its Sales Report Books and reporting any sales data;
• refrain from receiving or maintaining any non-public Sales Data³ limited to exceptions;
• for any report of non-sales data covering 50 percent or more of U.S. sales for the relevant segment, ensure all aggregated data consists of at least three meat processors and that no single processor represents more than 70 percent of the data;
• require quartile averages to contain at least three complexes⁴ per quartile;
• ensure reported information averages at least 45 days old generally and 90 days old for production decision data;
• prohibit participant lists, processor rankings, “flags” and other reporting features that could facilitate reverse engineering of competitor identities; and
• make all reporting and manuals available for purchase by any person in the United States on non-discriminatory terms, without conditioning access on a contribution of data.

Comparison to Prior DOJ Case Settlements

The Agri Stats settlement follows a separate suit and settlement that the DOJ brought against several meat processors and a different report provider in 2022: United States v. Cargill.⁵ But the Agri Stats settlement is strikingly different from the settlement in Cargill. In that case, the DOJ broadly defined “competitively sensitive information” as “non-public information that is relevant to, or likely to have an impact on, at least one dimension of competition, including price, cost (including Compensation), output, quality, and innovation.”⁶ The settlement then broadly prohibits the defendants from communicating such information with third parties or using such information received from others.⁷

In Agri Stats, the DOJ took a more rules-based approach. It outright barred collection and distribution of sales data and, for non-sales data, borrowed from a bygone safe harbor for information sharing from agency guidance withdrawn under President Biden.⁸ The report composition and data recency requirements track the structure of the old safe harbor but with more lenient thresholds (e.g., requiring three participants instead of five, allowing a single participant to represent up to 70 percent of the data rather than just 25 percent).

Agri Stats should not be read as a revival of the 1996 Statement safe harbor, particularly given agency leadership’s strong focus on case-by-case assessment. Rather, it reveals updated agency thinking on the potential competitive harms of pooled information, a greater attention to the ability of sophisticated parties to reconstruct competitively sensitive data, and a broader preference for more concrete requirements in behavioral remedies in this Administration.

The Agri Stats settlement should also be read with the DOJ’s settlement of the Biden-era RealPage litigation (discussed in detail in our prior client alert).⁹ Both case concern arrangements whereby a third-party hub collects, processes, and disseminates potentially sensitive participant data. However, in RealPage the information hub undertook analysis of the submitted data itself to produce individualized pricing recommendations to participants; Agri Stats by contrast is a more traditional information sharing case whereby participants received the same processed pooled data and analyzed it themselves.

The difference is reflected in the terms of the two settlements. While both contain terms covering both the input to the third-party data collector and the collector’s output, the heavy lifting is done at the point where data may become actionable business logic. The RealPage settlement therefore concerned itself primarily with the training and runtime inputs to RealPage’s AI models and the scope of analysis those models could provide. In Agri Stats, the data confidentiality provisions on the reports seek to ensure that participating meat processors cannot obtain inappropriately sensitive information for their own analysis.

Takeaways

The Agri Stats settlement provides insight into this DOJ’s view of competitive harms that may arise from information sharing arrangements and a potential framework for addressing those harms.

• Compliance requires consideration of the whole sharing cycle. Firms sharing information with hubs can no longer rely on either formalistic analyses (like the 1996 Statement) or view their own data contribution in isolation. Agencies are focused on the ultimate effects of the arrangement, which requires careful assessment of how that data can be processed and applied by any participant.
• Agencies are sensitive to real-world data analysis capabilities. The data recency requirements in the Agri Stats settlement reflect consideration of how the involved firms could use particular categories of data. Firms should expect close and case-specific examination of how specific data can be analyzed and how it might affect business decision making. For instance, both Agri Stats and RealPage show that it is not enough that data simply be historical; it matters what the data represents and how it is used.
• Agencies are sensitive to where the work is done. In a similar vein, the agencies have shown sensitivity to who enables the transformation from data to a business decision. The more analysis and refinement a party performs, the greater the risk that a remedy will operationally constrain that party.
• Some information may be too sensitive to share. The DOJ outright prohibited the collection or dissemination of sales data, seemingly considering that the risk of reconstruction could not be fully mitigated by confidentiality or aggregation requirements.
• Limited participation can increase risk. The Agri Stats settlement requires that the company make its reports available to all, rather than just to participants. The administration has expressed similar information bottleneck concerns in the Executive Order on AI policy (see our prior client alert on the order).

For more information, please contact any member of the firm’s Antitrust and Competition practice.

---

[1] Complaint, United States v. Agri Stats, Inc., No. 0:23-CV-03009-JRT-JFD (D. Minn. Sept. 28, 2023).

[2] Proposed Final Judgment, United States v. Agri Stats, Inc., No. 0:23-CV-03009-JRT-JFD (D. Minn. May 7, 2026).

[3] Defined as non-public information concerning “amounts charged or paid for broiler chicken, pork, or turkey products sold in the United States,” including aggregate or statistical data based on that information.

[4] Defined as a standalone facility engaged in any aspect of meat processing operations.

[5] Proposed Final Judgement, United States v. Cargill Meat Solutions Corp., No. 1:22-cv-01821-ELH (July 25, 2022).

[6] Id. at ¶ II.F.

[7] Id. § IV.

[8] The FTC and DOJ’s 1996 Statements of Antitrust Enforcement Policy in Health Care (Statement 6) had for nearly three decades described an information-exchange safety zone covering: exchanges for at least five participants, where no participant over 25% of the data, and where collected data were at least three months old and statistically aggregated in reports. U.S. Dep’t of Justice & Fed. Trade Comm’n, Statements of Antitrust Enforcement Policy in Health Care Statement 6, at 49-52 (1996). Both agencies withdrew from the Statement in 2023.

[9] Proposed Final Judgment, United States v. RealPage, Inc., No. 1:24-cv-00710-WO-JLW (M.D.N.C. Filed Nov. 24, 2025).