Key figures: RealPage settlement: November 2025, seven-year consent decree, no fine, court-appointed monitor · DOJ criminal warning: May 14, 2026 · California AB 325: effective January 1, 2026 · New York S.7882: effective December 15, 2025 · Municipal bans: San Francisco, Philadelphia, Seattle, Minneapolis, others

On May 14, 2026, Acting Deputy Assistant Attorney General Daniel Glad stepped to the podium at the Antitrust West Coast Conference in San Francisco and delivered a message that the technology and real estate industries had been expecting for months. "Software cannot launder collusion," he said. "When competitors exchange competitive intentions in a hotel suite or through a trade association, it is well settled that that raises antitrust concerns. So too with a text thread or a common algorithm."1

The speech, titled "Old Crime, New Code," made two things explicit. First, the DOJ considers algorithmic pricing tools capable of facilitating the same kind of price coordination that the Sherman Act has prohibited since 1890. Second, criminal charges, not just civil settlements, are now on the table for companies and individuals involved.1

This is not a hypothetical enforcement posture. It follows the DOJ's November 2025 settlement with RealPage, a real estate software company at the centre of the most significant algorithmic pricing case in US history, and a wave of state legislation that has rewritten the rules for how competitors can use shared data.2

The RealPage Case: How the Algorithm Worked

The case against RealPage began with a ProPublica investigation in October 2022, which revealed that the company's revenue management software was being used by competing landlords in a way that legal experts said could amount to cartel-like behaviour.3

Hub-and-spoke collusion: A form of price coordination in which competitors do not communicate directly with each other. Instead, they share information through a common intermediary (the "hub"), which aggregates the data and issues pricing recommendations back to each competitor (the "spokes"). The competitors never need to speak, but the outcome can resemble an explicit price-fixing agreement.

The mechanism was straightforward. Competing landlords subscribed to RealPage's software. Each landlord fed the system their own non-public data: vacancy rates, lease renewal dates, current rents, and planned pricing changes. RealPage's algorithm pooled this data across competitors and generated rental pricing recommendations that each landlord could then adopt. The DOJ alleged that the software included features designed to restrain price decreases, effectively embedding a one-way ratchet against lower rents, and that RealPage hosted meetings where competing property managers discussed pricing strategies using non-public data.4

The economics of the arrangement map precisely onto a classic coordination problem. In a competitive rental market, individual landlords have an incentive to undercut each other on price to fill vacancies. This is the prisoner's dilemma: each firm would prefer that all competitors keep prices high, but each individual firm benefits from cutting its own price. Cartels solve this problem through explicit agreement. RealPage's software solved it through shared data and algorithmic recommendations, without anyone needing to pick up a phone.

A corporate fine is paid by the shareholders. A sentence of imprisonment is not. That distinction has driven the Antitrust Division's deterrence strategy for decades, and nothing about the move from yellow pads to language models changes the math.

Daniel Glad, Acting Deputy Assistant Attorney General for Criminal Enforcement, DOJ Antitrust Division, May 14, 2026.1

The Settlement and Its Limits

The DOJ and RealPage reached a proposed settlement on November 24, 2025. The terms were detailed but carried no financial penalty. RealPage did not admit liability. Under the seven-year consent decree, RealPage must stop using non-public, competitively sensitive data from competing landlords when generating unit-level rent recommendations. Runtime pricing may only draw on a landlord's own data and publicly available information. RealPage must restrict any data it does use for model training to information at least 12 months old and aggregated to no finer than the state level. A court-appointed monitor will oversee compliance for three years.2

Apartment units priced using RealPage revenue management software (millions), 2017 to 2024. Source: ProPublica; RealPage marketing materials; US Senate letter.5
5

The scale of the network matters. By 2021, RealPage's marketing materials claimed the software was "trusted by over 4 million units," representing roughly 8 percent of all rental units nationwide according to a US Senate letter citing the data.5 In certain metro areas, concentration was far higher: the DC Attorney General's lawsuit contends that more than 90 percent of large apartment buildings in the Washington, D.C., area use RealPage to set rents.5 The more landlords that fed data into the system, the more comprehensive its view of the market became, and the more effective its pricing recommendations were at steering the market upward. This is the network effect applied not to consumers, but to competitors.

The settlement also required RealPage to eliminate any software feature that uses competing landlords' data to prevent prices from falling below algorithmically set floors, unless the same feature equally permits prices to exceed ceilings. This addressed a core concern: that the software was not merely aggregating information but actively steering prices upward.4

Critics argue the settlement is weak. There is no fine. No admission of wrongdoing. And the restrictions still permit RealPage to operate its revenue management business, provided it uses older, more aggregated data. The DOJ has also separately sued six major landlords, including Greystar, as co-defendants for their role in using the software. Some have reached individual settlements.3

The Legal Frontier: Can an Algorithm Constitute an Agreement?

The central question in algorithmic pricing cases is whether the use of a common algorithm, fed by competitors' data, constitutes the kind of "agreement" that Section 1 of the Sherman Act requires. Traditional antitrust law draws a clear line: conscious parallelism, where firms independently arrive at similar prices by observing the same market conditions, is legal. Explicit agreements to fix prices are not. The difficulty is that algorithmic pricing sits in the grey zone between the two.

Glad's May 2026 speech addressed this directly. He argued that when competitors "understood that their sensitive non-public data will be used to set prices for competitors" and participated in the system "on that understanding," such evidence could support allegations that use of a common algorithm amounted to an agreement to fix prices. Applied to large language models, he warned: "If your pricing system depends on your competitors' confidential inputs to function, you should expect us to ask why that is not anticompetitive coordination."1

This is a significant doctrinal step. It suggests the DOJ views the knowing, mutual submission of non-public data to a common algorithm as functionally equivalent to a hub-and-spoke conspiracy, even without direct communication between the spokes.

The State Response

While federal enforcement has relied on existing antitrust statutes, states have moved to write algorithmic pricing into law explicitly.

California's AB 325, signed by Governor Newsom on October 6, 2025 and effective January 1, 2026, amends the Cartwright Act to make it unlawful to use or distribute a "common pricing algorithm" as part of an agreement in restraint of trade. It defines a common pricing algorithm as any technology used by two or more persons that ingests competitor data to recommend, align, or influence a price. Critically, it also lowers the pleading standard: plaintiffs need only allege that collusion is "plausible," removing the federal requirement to plead facts tending to exclude the possibility of independent action.6

New York's S.7882, signed by Governor Hochul on October 16, 2025 and effective December 15, 2025, is narrower. It amends the Donnelly Act to prohibit residential landlords from using algorithmic software that performs a "coordinating function" between competing property owners. The law specifically targets tools that collect price, supply, or lease data from two or more landlords and use it to generate rental recommendations.7

Cumulative state and municipal algorithmic pricing laws enacted, October 2024 to June 2026. Source: Arnold & Porter; Wilson Sonsini; Perkins Coie.6
6

Chart note: Includes municipal bans (San Francisco, Philadelphia, Seattle, Minneapolis, San Diego, Jersey City, Providence, and others) and state laws (California AB 325, New York S.7882, New York Algorithmic Pricing Disclosure Act). Count is approximate.

At the municipal level, at least eight cities have enacted bans on algorithmic rent-setting tools, including San Francisco, Philadelphia, Seattle, Minneapolis, San Diego, Jersey City, and Providence.7 Several additional state legislatures have introduced bills, including New Jersey and Pennsylvania.6

When Machines Learn to Coordinate

The algorithmic pricing debate tests a foundational assumption in antitrust: that collusion requires conscious agreement between human beings. If an algorithm can replicate the outcome of a price-fixing cartel without any direct communication between competitors, the legal framework built around proving "agreement" may be inadequate.

The economic concern is not that algorithms are inherently anticompetitive. A firm using its own proprietary data to optimise pricing is simply competing more efficiently. The problem arises when the algorithm draws on competitors' non-public data, because at that point it is no longer optimising in isolation. It is incorporating information about competitors' intentions, costs, and strategies, which is precisely the information exchange that antitrust law has always regarded with suspicion.

Glad's speech acknowledged a further horizon: autonomous algorithms that learn to coordinate without being explicitly programmed to do so. When independent pricing algorithms interact repeatedly in the same market, they can converge on supra-competitive prices through a process analogous to tacit collusion. Current law and the current wave of enforcement do not yet reach this scenario. But the DOJ has signalled that it is watching.1

The acceleration from a single ProPublica investigation in 2022 to criminal enforcement warnings, a federal settlement, two state laws, and eight municipal bans in under four years is remarkable. It reflects a recognition, across branches of government and levels of jurisdiction, that the tools of competition have changed faster than the rules governing them. The question is no longer whether algorithmic pricing can facilitate collusion. The RealPage case answered that. The question now is whether the legal system can keep up.

Footnotes

  1. DOJ, Acting Deputy Assistant Attorney General for Criminal Enforcement Daniel Glad Delivers Remarks at the Antitrust West Coast Conference (opens in a new tab), 14 May 2026Axinn, "Old Crime, New Code" – DOJ Outlines its Views on When Software and AI Can Facilitate Collusion (opens in a new tab), 18 May 2026. ↩ ↩2 ↩3 ↩4 ↩5

  2. DOJ, Justice Department Requires RealPage to End the Sharing of Competitively Sensitive Information and Alignment of Pricing Among Competitors (opens in a new tab), 24 November 2025Fenwick, DOJ's RealPage Settlement: A Blueprint for 'Safer' Algorithmic Pricing? (opens in a new tab), 5 December 2025. ↩ ↩2

  3. ProPublica, DOJ and RealPage Agree to Settle Rental Price-Fixing Case (opens in a new tab), 26 November 2025. ↩ ↩2

  4. Wilson Sonsini, DOJ Settles Its Algorithmic Price-Fixing Case Against RealPage (opens in a new tab), 1 December 2025ReedSmith, Algorithmic pricing under pressure: DOJ's RealPage settlement changes the rules for rental markets (opens in a new tab), 3 December 2025. ↩ ↩2

  5. ProPublica, Department of Justice Opens Investigation Into Real Estate Tech Company Accused of Collusion with Landlords (opens in a new tab), 23 November 2022 (1.5M and 4M unit figures from RealPage CEO and marketing materials)ProPublica, Senators Had Questions for the Maker of a Rent-Setting Algorithm. The Answers Were "Alarming." (opens in a new tab), 21 March 2023 (8% of all rental units figure)ProPublica, DOJ Backs Tenants in Case Alleging Price-Fixing by Big Landlords and a Real Estate Tech Company (opens in a new tab), 16 November 2023 (DC metro 90% figure). ↩ ↩2 ↩3

  6. Davis Polk, New laws regulating algorithmic pricing enacted in New York and California (opens in a new tab), 27 October 2025Arnold & Porter, Algorithmic Pricing Bans Go Coast to Coast (opens in a new tab), 28 October 2025. ↩ ↩2 ↩3

  7. Wilson Sonsini, New York Prohibits Use of Pricing Algorithms for Rent-Setting (opens in a new tab), 22 October 2025Perkins Coie, Algorithmic Price-Fixing: US States Hit Control-Alt-Delete on Digital Collusion (opens in a new tab), 30 December 2025. ↩ ↩2