Four States Now Ban Algorithmic Rent Pricing: What Small Landlords Must Change Before July 2027

1. What do the algorithmic rent pricing laws actually prohibit?

These laws make it unlawful for residential landlords to set rents using software that pools or coordinates pricing, occupancy, or lease data across unaffiliated competing property owners. Four states have enacted them: New York effective December 15, 2025, California and Connecticut effective January 1, 2026, and New Jersey effective July 1, 2027.

An important qualification belongs next to that sentence. None of these statutes categorically outlaws every algorithmic pricing tool. They reach a narrower category of conduct: setting rents on the recommendation of a system performing a coordinating function across competitors. The popular shorthand calls them bans on algorithmic rent pricing, which is close enough for a headline but too blunt for a compliance decision.

A tool that looks only at your own portfolio, your own vacancy rates, your own lease trade out history, and publicly posted asking rents is not the central target. A tool that reaches into a shared pool of numbers contributed by the competing owner across the street and hands you back a recommended rent is precisely the target. The legislative theory is that when enough competitors outsource pricing to one engine fed by each other’s data, the result can function as a cartel even though no one ever spoke to anyone.

This distinction determines whether your existing software stack is a problem. The answer is not driven by how sophisticated the tool is. It is driven by whose data goes in.

2. Which states have restricted algorithmic rent setting, and when does each take effect?

Four states have enacted restrictions as of August 2026, and each draws its boundary in a different place. New York and New Jersey wrote housing specific statutes. California wrote a general antitrust amendment that captures rental pricing among other markets. Connecticut wrote a housing specific rule narrower than the other three in one respect.

New York moved first among the states. Governor Kathy Hochul signed Senate Bill 7882 on October 16, 2025, adding Section 340-b to the General Business Law, effective December 15, 2025. As Foley and Lardner documented, the statute reaches software that collects past and present price data from two or more competing owners, processes it, and recommends rents or lease terms. The definition does not turn on whether competitor inputs are confidential, which makes New York the broadest of the four. Per Arnold and Porter, it carries fines up to $100,000 for an individual and $1 million for a corporation, plus potential imprisonment of up to four years.

New York’s private enforcement picture deserves care. Violations sit inside the Donnelly Act framework, which provides civil and criminal remedies including private treble damages actions in appropriate cases, and a separate pending bill would expressly authorize class actions. Do not assume every violation automatically produces every remedy. Relief depends on the claim, the plaintiff, and the procedural posture.

California took a different route. Assembly Bill 325, signed October 6, 2025 and effective January 1, 2026, amended the Cartwright Act. Per Morgan Lewis, a common pricing algorithm is any methodology used by two or more persons that uses competitor data to recommend, align, stabilize, set, or otherwise influence a price or commercial term. AB 325 also added Section 16756.1, eliminating the requirement that a complaint allege facts tending to exclude the possibility of independent action, which makes these cases easier to plead and harder to dismiss. Penalty exposure is significant: a corporate criminal fine of the greater of $6 million or twice the pecuniary gain or loss, individual fines up to $1 million, imprisonment up to three years, and a civil penalty up to $1 million.

A landlord using a pricing tool limited to its own data is less likely to fall inside the common pricing algorithm concept, since the definition requires use by two or more persons and competitor data. Treat that as a sound practical reading rather than a categorical safe harbor. California’s provision turns on agreements in restraint of trade and coercion, so the analysis stays fact specific and is not confined to housing.

Connecticut enacted HB 8002 in the November 2025 special session as Public Act No. 25-1, signed by Governor Ned Lamont and effective January 1, 2026. Baker McKenzie flags what makes it distinct: it is the first state statute to expressly limit its prohibition to nonpublic competitor data, leaving room to use public information. It bars using a revenue management device to set rental rates or occupancy levels for residential dwelling units, and carves out aggregated rental data reports and affordable housing compliance tools. Reported penalties run to $100,000 for individuals and $1 million for businesses.

New Jersey gives operators the most runway. Governor Mikie Sherrill signed the Forbidding the Algorithmic Inflation of Rent Act, known as the FAIR Act, on July 20, 2026, effective July 1, 2027. As Crowell and Moring and DLA Piper detail, it bars landlords and third party coordinators from using algorithmic devices to coordinate rents, and from sharing sensitive market data such as rents, occupancy levels, or material lease terms. The Act ties violations to New Jersey antitrust remedies, which may include Attorney General enforcement, injunctive relief, and private claims where the statutory requirements are met. The Attorney General must also stand up an online complaint database, lowering the cost of reporting a landlord from “hire a lawyer” to “fill out a form.”

The FAIR Act contains a useful definitional carve out. An algorithmic device does not include a spreadsheet that operates without artificial intelligence and requires human analysis to perform calculations, or a database that uses an algorithm only to query unprocessed data. If your pricing process is a manager with a spreadsheet, New Jersey says that is not the thing being regulated.

Read that table for what it does not say. There is no revenue threshold, no employee count, no unit minimum. Compare that to the pattern in frontier AI regulation, where thresholds do the sorting: Illinois SB 315 reaches only developers above $500 million in annual revenue. The rent pricing statutes contain no analogous filter, so a 40 unit operator can be covered. That is not the same as saying a small operator and a national REIT face identical risk. Exposure, exemptions, penalty calculation, and enforcement likelihood all vary with conduct, property type, and jurisdiction.

3. What did the DOJ RealPage settlement change for landlords who are not RealPage?

It converted an abstract antitrust theory into a published set of engineering rules, which makes it the most useful enforcement benchmark available even though it binds only RealPage. DOJ and a group of states sued RealPage in the Middle District of North Carolina in August 2024, and the parties filed a proposed settlement on November 24, 2025 in United States v. RealPage, Inc., No. 1:24-cv-00710, subject to court approval.

Analyses by Paul, Weiss and Fenwick converge on the core terms. At runtime, when the software produces a rent number, RealPage may not use nonpublic competitively sensitive data from competing landlords, and recommendations must rest on the landlord’s own data plus public information. For model training, nonpublic data may be used only when at least 12 months old and not tied to active leases, and new models may not use nonpublic data filtered below the nationwide level.

The product design terms are equally telling. Auto accept must be configurable and overrideable rather than defaulted on. Default settings may not favor price increases. The “governor” feature must treat increases and decreases symmetrically. RealPage may not incentivize adoption of recommended rents. Market surveys, meaning calling or emailing competitors for nonpublic numbers, are prohibited going forward.

Two cautions on how to use this. The consent judgment applies to RealPage. It is not a safe harbor for any other vendor or landlord, and it does not displace stricter state requirements, most obviously New York’s coordinating function rule, which does not depend on data being nonpublic. Calibrate to the strictest law reaching your properties. That said, the terms make a serviceable diagnostic: a product engaging in conduct RealPage agreed to stop presents heightened risk and warrants prompt legal and technical review.

The separate civil exposure has grown considerably. In the consolidated private litigation, In re RealPage, Inc., Rental Software Antitrust Litigation (No. II), No. 3:23-md-03071 in the Middle District of Tennessee, property owners and managers have entered 37 settlements totaling approximately $359.925 million: a first wave of 26 preliminarily approved in November 2025 and a second wave of 11 deals covering 14 companies preliminarily approved May 22, 2026, with a final approval hearing set for October 15, 2026. Those settling defendants are landlords and property managers, not the software vendor, and the plaintiffs needed no new state statute to bring the claims.

4. Which cities have their own restrictions, and do they reach small operators?

Yes, and the municipal layer is where most small operators get caught, because city ordinances arrive with less coverage than state statutes and several include a private right of action. City action also came first: local governments moved well ahead of the states.

San Francisco led in 2024, followed by Philadelphia. Arnold and Porter’s survey catalogs a wave through 2025 including Minneapolis, Jersey City, Providence, San Diego, Hoboken, Seattle, and Berkeley, while Santa Monica, Portland, and Spokane have adopted or considered measures. Treat any list as a snapshot and check your specific municipality rather than assuming only the two best publicized ordinances matter.

Seattle is the clearest on penalties. The City Council passed Council Bill 121000 on June 24, 2025, Mayor Bruce Harrell signed it on July 1, 2025, and it is codified as Ordinance 127241 at SMC 7.34. It authorizes penalties of up to $7,500 per violation and gives tenants a private right of action for damages up to $7,500 per violation. Publicly available rent estimates anyone can access without a contract are not violations. How a violation is counted across units and repricing events should be confirmed against the ordinance text before assuming your exposure is a single penalty.

Minneapolis amended Title 12, Chapter 244 of its city code, effective March 1, 2026. Per Winthrop and Weinstine, it bars residential owners and operators from using an algorithmic device that performs calculations of nonpublic competitor data on local or statewide rents or occupancy. Sources that merely publish or aggregate rental information without recommending are outside the definition, as are affordable housing compliance tools. Tenants have a private right of action for compensatory damages and attorney fees, enforcement is complaint based through Regulatory Services, and the ordinance adds a self attestation at rental license renewal, with penalties reaching license revocation.

That license attestation is the detail small operators most often miss, because it converts a passive obligation into an affirmative annual statement you sign. Fee shifting is the other item to take seriously: one tenant claim may be small, but a plaintiff firm aggregating claims across a portfolio is a different proposition.

5. Does this apply to my properties? A coordinating function screening test

Use the New York coordinating function definition as an initial risk screen, then apply the actual definitions and exemptions of each law reaching your properties. New York is the most detailed of the four, which makes it a useful first filter, but it is a screening tool rather than a universal legal test.

Under that definition, software performs a coordinating function when it does all three of the following: collects rent, supply, or lease data from two or more unaffiliated residential property owners or managers; analyzes or processes that data using computational methods such as machine learning; and recommends rental prices, lease renewals, occupancy targets, or other lease terms.

Run each tool through those three prongs. A product satisfying all three is a live compliance question in New York today and warrants review everywhere else. A product failing the first prong because it uses solely your own data sits outside the central prohibition, though California’s agreement and coercion framing and each city’s definitions still deserve a look.

The word “unaffiliated” matters for operators holding multiple entities. Small portfolios are often split into separate LLCs for liability reasons while remaining under common ownership. Data use within a commonly controlled portfolio generally presents lower coordination risk than pooling with independent competitors, but common control, management agreements, and investor overlap can all matter in an antitrust analysis. Document the ownership relationships rather than assuming the structure resolves the question.

Two further points shape your real exposure. First, there is no size floor in any of the four statutes. Individual investors own roughly 70 percent of rental properties and about 38 percent of all rental units, according to Congressional Research Service analysis of Census Rental Housing Finance Survey data. Legislatures did not exempt that population, so small operators are in scope by default even though their practical risk profile differs from an institutional owner’s.

Second, your state of residence is not the test. The property’s location is. An operator headquartered in Pennsylvania with two buildings in Jersey City faces both New Jersey’s statute as of July 1, 2027 and Jersey City’s ordinance today.

6. What can I still legally do with revenue management software?

You can generally continue using software, analytics, and machine learning to price units where the inputs are limited to your own portfolio data and genuinely public information. These statutes do not broadly prohibit a landlord from analyzing its own data. Risk rises sharply when a tool pools or coordinates data from unaffiliated competitors or otherwise functions as a common pricing mechanism. The DOJ settlement reflects the same line by letting RealPage continue offering recommendations built on a landlord’s own data plus public information.

The practical dividing line looks like this.

One caution on the left column. New York’s definition does not distinguish confidential from public information once a tool collects from two or more competing owners and generates recommendations. That makes New York stricter than Connecticut, which limits its prohibition to nonpublic competitor data, and stricter than the DOJ settlement. If you operate in New York, calibrate to New York.

Human decision authority is the other recurring theme, and it is worth being precise about what it buys you. A documented process where a named person reviews recommendations, can reject them, and sometimes does, supports independent judgment and creates useful evidence of governance. It does not cure a prohibited data source. A tool that pools competitor data does not become lawful because someone clicks approve, and a reviewer who rubber stamps every recommendation is not exercising independent judgment.

7. Will the First Amendment challenges overturn these laws?

Possibly, but not on a timeline you can plan around, and the challenges now span multiple jurisdictions. RealPage sued New York Attorney General Letitia James in the Southern District of New York in November 2025 in RealPage, Inc. v. James, arguing that Section 340-b imposes content based, viewpoint based, and speaker based restrictions violating the First Amendment. The company frames a pricing recommendation as analysis and speech a state may not single out and forbid. RealPage has brought a parallel challenge against Berkeley’s ordinance, where the court denied a temporary restraining order after the city agreed not to enforce against RealPage pending the injunction motion.

These matters remain in active litigation. Do not assume a pending constitutional challenge suspends any statute, and confirm current docket status rather than relying on a published summary, including this one. Multifamily Dive maintains a running tracker.

Three reasons not to build a compliance strategy on this litigation. First, the New York case challenges one statute in one state, and a RealPage victory there would leave California, Connecticut, New Jersey, and a long list of city ordinances standing. Second, the underlying conduct was already actionable under antitrust law before any of these statutes passed. Governor Jared Polis made that point vetoing Colorado’s HB 25-1004 on May 30, 2025, noting that landlords colluding to raise rents would already violate the Colorado Antitrust Act. The $359.925 million in Tennessee landlord settlements is the demonstration. Third, a loss would leave operators who paused compliance work with a shorter runway and a record of having waited.

8. How does a small operator build a defensible rent pricing governance file?

Build one folder with seven documents and you will have answered most of what a state Attorney General, a plaintiff’s lawyer, or an insurance underwriter is likely to ask. None of it requires a compliance department.

Document 1: A pricing tool inventory

List every product that touches a rent number: revenue management platforms, property management systems with pricing modules, listing site suggestions, and any advisor who supplies market numbers. Record the vendor, the properties it touches, and the date last reviewed. Most operators find two or three tools they had not thought of as pricing tools.

Document 2: A written vendor data sourcing attestation

Send each vendor the screening test from section 5 and ask them to confirm in writing whether the product collects rent, supply, or lease data from two or more unaffiliated owners and returns price recommendations. Ask separately about runtime and training inputs, since the DOJ settlement treats those differently. A vendor that will not answer in writing has told you something.

Document 3: A jurisdiction map

List each property against its state and city, flagging New York, California, Connecticut, and New Jersey. Check every municipality separately, since city ordinances now reach well beyond the two most publicized ones. Scattered site portfolios routinely include a building inside a city boundary the owner assumed it was outside.

Document 4: A written rent setting policy

Two pages is enough. State which data sources are approved and prohibited, who holds final pricing authority, that recommendations are advisory rather than binding, and that no employee may solicit or receive nonpublic pricing or occupancy data from a competing owner. Date it and have the owner sign it.

Document 5: An override log

Record every instance where a human rejected or adjusted a recommendation, with date and reason. This is the most persuasive artifact you can produce, because it shows independent pricing judgment as lived practice rather than policy language. If the log is empty after six months, your process is automated in substance whatever the policy says.

Document 6: Staff training acknowledgment

A short annual session on what may and may not be shared with competing owners, including at industry events and in local landlord groups. Collect signatures. Informal information exchange among local operators predates any software and remains independently actionable under antitrust law.

Document 7: A contract review note

Check whether your vendor agreement indemnifies you for regulatory exposure from the product’s data practices, and whether it requires notice of material changes to data sourcing. Most standard agreements do neither. New Jersey’s July 1, 2027 date gives operators there a real window to renegotiate at renewal rather than under duress.

This is the governance pattern that works in every regulated AI use case: inventory the systems, document the data lineage, assign human accountability, and keep evidence the controls operate. It is the structure behind the NIST AI Risk Management Framework, and it scales down to a 60 unit portfolio.

9. What other AI compliance exposures do landlords have beyond pricing?

Tenant screening is the second major exposure, and its federal guidance landscape shifted in 2026 in a way that is easy to misread. HUD’s Office of Fair Housing and Equal Opportunity issued guidance on May 2, 2024 on applying the Fair Housing Act to tenant screening, including screening run by third party companies using artificial intelligence. A companion document dated April 29, 2024 covered housing advertising through digital platforms and algorithmic ad targeting.

In 2026 HUD withdrew a set of fair housing guidance documents, finalized in a Federal Register notice published April 6, 2026 and effective as of September 17, 2025. Published summaries of that withdrawal list include the April 29, 2024 advertising guidance and do not list the May 2, 2024 tenant screening guidance. Confirm any specific document’s status against the notice itself before relying on it.

Do not over read either fact. Guidance interprets the law; it is not the law. The Fair Housing Act is unchanged, and withdrawing guidance removes a roadmap rather than the underlying liability. A screening process that produces a discriminatory effect can create Fair Housing Act risk, particularly where the provider cannot articulate a substantial, legitimate, nondiscriminatory interest or where a less discriminatory alternative was available. Liability is not automatic because outcomes differ across groups, but neither is risk eliminated by outsourcing screening to a vendor.

Separately, the Fair Credit Reporting Act continues to govern adverse action obligations when a consumer report contributes to a denial, whether or not the screening process includes automated or AI enabled elements.

Two other exposures deserve a line each. AI generated listing copy carries disclosure and deception risk under Federal Trade Commission authority and a growing set of state advertising rules. And AI assisted lease drafting raises the same accuracy and unauthorized practice problems state bars have addressed for lawyers, a topic we covered in our review of state bar AI ethics rules for small firms.

10. What is coming in the 2026 and 2027 legislative sessions?

Expect the covered map to keep expanding, because introduced bill volume far outruns enactment. Consumer Reports counted 51 algorithmic pricing bills across 24 states in early 2025, up from 10 in all of 2024, most of them targeting rent setting software, as Tech Policy Press documented. By July 2026, Newsweek identified 21 states where lawmakers had introduced, advanced, or enacted such legislation. Treat that as a map of legislative activity, not legal status.

Illinois is among the closest to acting. SB 343, which would amend the Illinois Antitrust Act to make price coordination for residential rental units a violation, passed the Illinois Senate on May 21, 2026 and moved to the House. Operators with Illinois units should confirm its current status before their next repricing cycle rather than relying on any published summary.

Two patterns are worth noting. Effective dates are lengthening, from New York’s roughly six week runway to New Jersey’s eleven months, suggesting legislatures have absorbed the argument that operators need time to unwind vendor contracts. And private enforcement is appearing in more state and local measures, though the availability of a private action, damages, fees, and administrative enforcement still varies materially by jurisdiction.

For a small operator, this is not a wait and see area. The direction of travel is consistent, the compliance work is cheap relative to the exposure, and the artifacts you build now will largely satisfy the next several jurisdictions. Operators who watched the state by state employment AI patchwork develop will recognize the pattern from our analysis of what actually applies to employers after Colorado’s AI law never took effect.

This article is provided for general informational purposes and reflects the state of the law as of August 2026. It is not legal advice. Regulations in this area are changing quickly, and several of the statutes and ordinances discussed here are the subject of active litigation. Confirm current requirements and consult qualified counsel before making decisions for your organization.