A software company called RealPage built an algorithm that told landlords what to charge for rent. The algorithm ingested nonpublic pricing data from competing landlords — occupancy rates, lease terms, renewal timing — and generated recommendations that nudged rents in the same direction at the same time. One landlord told RealPage it raised rents more than 25% within 11 months of adopting the software.
The Department of Justice called it a cartel. Then it settled for zero dollars.
The Machine That Learned to Collude
RealPage's revenue management software priced more than 4 million rental units across the United States — roughly 1 in 4 apartments nationwide, according to census data cross-referenced with the company's own disclosures. The algorithm didn't just recommend prices. It created a feedback loop: landlords shared their real-time data with RealPage, RealPage fed that data into a model, and the model told each landlord to charge more. The system worked because everyone followed it. An auto-accept feature made it easier to obey the algorithm than to override it.
Researchers calculated the damage. Renters in units priced by algorithmic tools paid $70 more per month on average — a 4% premium over market rate. Total overcharge in 2023 alone: $3.8 billion, spread across millions of households that never knew a machine was setting their rent.
ProPublica exposed the scheme in 2022. Senators demanded answers. RealPage's responses were, in the words of a Senate subcommittee, "alarming." The DOJ sued in August 2024 and followed up by suing six of the nation's largest landlords — companies managing 1.3 million apartments across 43 states. Greystar, which manages nearly 950,000 apartments, settled separately.
The Settlement That Settled Nothing
On November 24, 2025, the DOJ announced its proposed settlement with RealPage. The terms: RealPage must stop using nonpublic data less than 12 months old. It must disable auto-accept features. It must stop geographic modeling below the state level. A court-appointed monitor will watch for seven years.
No fines. No admission of wrongdoing. No money returned to renters.
A separate class-action yielded $141 million from landlords — roughly $35 per affected apartment, or less than one month of the algorithmic premium each tenant paid.
RealPage didn't invent rent increases. But it automated the coordination mechanism that antitrust law exists to prevent. The algorithm replaced the smoke-filled room with a SaaS dashboard. The DOJ's settlement acknowledged this by restricting how the software works — while carefully avoiding the precedent that algorithmic pricing is inherently illegal.
The Algorithm Spreads
RealPage was first. It won't be last. The FTC claims Amazon's pricing algorithm generated over $1 billion in additional revenue by influencing market-wide price dynamics. That case goes to trial in October 2026. Germany's Federal Cartel Office sent formal charges to Amazon in June 2025 over algorithmic price caps that disadvantage third-party sellers. The European Commission confirmed in July 2025 that it's investigating algorithmic pricing collusion in airlines, banking, and pharmaceuticals. The Netherlands is probing airline pricing separately. Poland opened investigations into banking and pharma.
California enacted an algorithmic pricing law effective January 1, 2026, targeting algorithms that enable competitors to align prices by sharing sensitive data — with criminal penalties. New York signed its own ban into law in December 2025 — and RealPage immediately sued, calling the ban "a sweeping and unconstitutional ban on lawful speech." The New York Attorney General responded that agreements not to compete don't become protected speech just because they're carried out through software. Spokane banned algorithmic rent-setting outright. DOJ Assistant Attorney General Gail Slater warned that algorithmic pricing probes will increase.
Wendy's learned the consumer-facing version of this lesson when it floated "surge pricing" for hamburgers. The backlash was immediate and severe. But Wendy's made the mistake of telling customers what the algorithm was doing. RealPage's insight was subtler: let the algorithm coordinate pricing between businesses, not between the business and its customers. The renters never saw the dashboard.
Why This Matters More Than Rent
The RealPage case is a template. Any industry where competitors share data through a common platform — hotels, insurance, freight, healthcare — can replicate the same dynamic. The algorithm doesn't need to be told to collude. It learns that coordinated price increases maximize revenue for all participants. The collusion emerges from the optimization function, not from a conspiracy.
Legal scholars at Berkeley found the core problem: traditional antitrust law requires proof that competitors agreed to fix prices. An algorithm that independently discovers that convergent pricing maximizes returns doesn't fit neatly into that framework. The DOJ's RealPage settlement avoided testing this question. By settling instead of going to trial, the government preserved the ambiguity — and with it, the loophole.
Four million apartments. $3.8 billion extracted. Zero fines. The algorithm worked exactly as designed. The legal system is still trying to figure out whether that's a crime.
Top comments (0)