RealPage Paid No Fine and Became the Government's Witness. Forty Landlords Are Paying $359,925,000.

RealPage got no fine, no admission of wrongdoing, and a seven-year consent decree. Its customers got a $359,925,000 bill and a state attorneys general case that is still running. The judge signs off on October 15.

The Justice Department sued RealPage in August 2024 and settled with the company on November 24, 2025. RealPage paid nothing. It admitted nothing. In exchange it agreed to stop training its models on competitors' nonpublic lease data, to accept a monitor, and, per the government's own competitive impact statement, to cooperate in the United States' lawsuit against the property management companies that used its software.

Read that sequence again.

The vendor that sold the pricing recommendations is now a witness against the operators who took them.

Meanwhile the private class action consolidated in the Middle District of Tennessee has produced a $359,925,000 settlement fund from roughly forty property managers, including Greystar, Equity Residential, Camden, Cortland, Bozzuto and Lincoln Property Company. The class covers anyone who paid rent at a property using RealPage revenue management, LRO, YieldStar or AIRM between October 18, 2018 and November 21, 2025. Final approval is set for October 15, 2026.

Claims are due January 29, 2027.

The consent decree, the government's competitive impact statement and the class notice run to a few hundred pages between them. Here is what they say about your exposure, your pricing and your claim, in that order.

The algorithm's actual trade was rate for occupancy, and operators took it anyway

The popular version of this story is that software secretly inflated rents across America. The best academic work on it says something more useful, and more damning for the operators who signed up.

Sophie Calder-Wang and Gi Heung Kim at Wharton studied adoption decisions against market-rate apartment data from 2005 to 2019. Their finding on the 2013 adoption cohort is the one every owner should have read before signing: those buildings showed significant price growth versus non-adopters after 2014, and experienced "almost 1.5 percentage points lower in occupancy." Higher rent, emptier building.

The earlier 2007 cohort did the reverse in the recession, cutting price aggressively and gaining occupancy. The software wasn't hiding what it did. It was doing exactly what it was sold to do.

That isn't a conspiracy. Call it the trade: a revenue strategy with a known cost, which management companies bought on purpose because in a tightening market the rate side wins. The authors are careful to say the cross-market pattern is consistent with either coordination through the algorithm or widespread pricing error among the people who did not adopt it. Both readings are bad for the operator. In the first you have a legal problem. In the second you paid a vendor to tell you what your own lease trade-out report already knew.

The strategy stopped paying before the lawyers showed up

The Census Bureau put the national rental vacancy rate at 7.3 percent in the second quarter of 2026, up from 7.0 percent a year earlier and the highest level since 2017. Apartment List's September 2026 national report has rents down 0.1 percent for the month and down 0.4 percent year over year, with a vacancy index of 7.0 percent on its narrower market-rate sample. The 7.3 percent is the one to underwrite, because Census measures the whole rental stock.

So the sector is sitting on negative rent growth and the loosest market in nine years, while writing a nine-figure check for rent increases it took between 2018 and 2025. Divide $359,925,000 across forty settling companies and you get roughly $9 million each before anyone counts legal fees, and that is only the private class. The states' case is separate and alive: a federal judge refused to throw out the state attorneys general claims, and attorneys general from Kentucky, Washington, New Jersey, Maryland and the District of Columbia have taken an appeal to the Sixth Circuit over an order limiting their damages, with twenty-five states filing in support.

If vacancy is the loosest in nine years, why would anyone still pay for rate optimization?

Because the rules are being written now, for the next tight market. Five states and eight cities have already legislated what no court has yet ruled on. California, New York and New Jersey have enacted statutes restricting algorithmic rent-setting, Connecticut and Maryland have their own versions, and San Francisco, Berkeley, Philadelphia, Minneapolis, San Diego, Seattle, Jersey City and Hoboken have passed local ordinances. Twenty-one states have had bills introduced or advanced.

None of those ordinances ask how many units you own.

The eleven-door landlord is inside this too

So this is a Greystar problem, not an eleven-door problem?

Here's where that reading breaks.

• Assuming this is a big-operator problem. The city ordinances and the state statutes regulate the conduct, not the portfolio size. If you run eight units in Philadelphia through a pricing tool that ingests other landlords' nonpublic lease data, you are inside the same rule as Greystar. Read the ordinance where your doors actually are.
• Confusing public comps with pooled private data. Pricing off listed asking rents, your own trade-outs, and published market reports is ordinary business judgment. The thing under attack is a vendor pooling nonpublic, forward-looking lease terms from competitors and feeding the output back to all of them. That distinction is the whole case.
• Treating the DOJ decree as the end. It runs seven years with a three-year monitor, it binds RealPage and not its customers, and it converted the vendor into a cooperating party. The exposure moved downstream, it did not disappear.
• Underwriting 2021-era rate growth into a 7.3 percent vacancy market. The algorithm's whole premise was that you can push rate and absorb the vacancy. At the loosest national vacancy in nine years, with rents negative year over year, that premise is wrong, and the pro forma built on it breaks at the vacancy line, not the rent line.

Where I land

The software didn't invent rent increases. It sold the settling operators the trade, higher rate for lower occupancy, and then handed the government their file. I wouldn't run my pricing through any third party that pools my competitors' nonpublic lease terms, at any price, because rate lift nobody has yet sized is never worth an open-ended contingent liability plus whatever a Sixth Circuit panel decides about state damages. The consensus take, that the algorithm is why rent is expensive, lets the actual culprit off the hook: the industry delivered over 600,000 multifamily units in 2024, the most in a single year since 1986, and concentrated them in a handful of Sun Belt metros. Those metros are exactly where rents are falling fastest right now, with the software still switched on. San Antonio is down 4.5 percent year over year and Phoenix down 3.5 percent.

My call, dated: by the end of Q2 2027 at least one of these forty settling companies will publicly disclose a second nine-figure rent-pricing liability from the state cases, and the five states with algorithmic pricing statutes on the books today will be at least ten.

What I can't see is the allocation. The settlement doesn't break out what each company pays, so whether roughly $9 million a head reads as a deterrent or as a cost of doing business is something I'm guessing at like everyone else. Price your units off your own comps and your own lease files. It is the cheapest legal defense in real estate and it is also, in a 7.3 percent vacancy market, the better pricing.

How to actually claim, and who to contact

This is reporting, not legal advice. If you paid rent at a property using RealPage revenue management, LRO, YieldStar or AIRM between October 18, 2018 and November 21, 2025, you are in the class. File at RealPageRentalSettlement.com by January 29, 2027. The administrator's line is 1-888-995-4213 and the mailing address is P.O. Box 59, East Brunswick, NJ 08816-9998. You will want your lease dates, the property name and the rent you paid, since the fund is distributed pro rata by rent paid against estimated overcharge. The final approval hearing is October 15, 2026 in the Middle District of Tennessee, Nashville Division, case 3:23-md-03071, and no money moves until after it.

How to Use PropGPT for This

The pricing edge now belongs to whoever can rebuild a rent comp set from public ownership and transaction data instead of renting one from a vendor.

That's a PropGPT query.

"Show me all multifamily properties with 50 or more units within 3 miles of 2100 Charlotte Ave, Nashville, TN, with current owner name, mailing address, last sale date and last sale price."

This is your competitive set, built from public record rather than a data co-op. Owner name plus last sale tells you who bought at peak pricing and therefore who has to defend rate hardest, which is exactly who will blink first on concessions.

"List apartment properties of 20 to 100 units in Charlotte, NC owned by out-of-state LLCs, held more than seven years, with no mortgage recorded, including owner mailing addresses."

Long-held, free-and-clear, out-of-state owners are the operators least likely to have ever licensed revenue management software and the most likely to be underpricing. They are both your acquisition list and your honest read on where market rent actually sits.

"Find 5-to-20 unit multifamily in Philadelphia, PA where the owner also owns at least three other multifamily properties in the metro, with all addresses grouped by owner."

Portfolio clustering tells you which small operators move rents together in a submarket. If you are competing with them, you want to know the cluster. If you are buying from them, you want to know it is a portfolio trade and not a one-off.

"Run the numbers on a 24-unit building in Columbus, OH at $2.1 million, $1,150 average rent, 7.3% vacancy, 42% operating expense ratio. Show NOI and cap rate, then rerun at $1,085 rent."

Underwrite the market vacancy, not the vendor's. The second number is what the building does if you give back the last three years of algorithmic rate, which is the realistic downside case in a loosening market.

"Show me multifamily properties in Seattle, WA sold in the last 24 months to buyers who own more than 1,000 units nationally, with sale price and price per unit."

Large national buyers are the ones now operating under consent decrees and city ordinances. Their price per unit tells you what the constrained-pricing cohort is willing to pay, which is a floor you can underwrite against.

The bottom line

The settling operators are about to pay $359,925,000 for a pricing strategy whose only measured result is 1.5 percentage points less occupancy and price growth nobody has yet put a number on, while the vendor that sold it walks away with a consent decree and a cooperation agreement. The lesson is not that technology is the enemy. It is that outsourcing your pricing decision to a party holding your competitors' confidential data means outsourcing your legal exposure too, and the price of that convenience just got published to four significant figures.

Build your own comp set. Price your own units. The ownership and transaction data is already public, and it is a better comp set than the one your competitors licensed.

That's the part I keep turning over. The data I'd have paid a vendor for was sitting in the public record the whole time.