We know. It's a holiday.

You probably shouldn't be thinking about property taxes today.

But for all the real estate nerds out there who can't quite turn it off — even on Labor Day — we've got you covered.

Here are five weird real estate factoids you can casually drop at the barbecue this weekend.

California has a grizzly bear on its flag even though there hasn't been a wild grizzly documented in the state in more than 100 years.

Oregon gave us Nike, world-class Pinot Noir…and one of the more unusual property tax systems in the country.

South Carolina's famous palmetto tree once helped a fort withstand British cannon fire.

Michigan has a left turn that sometimes requires you to turn right first.

And underneath a small Pennsylvania town, a fire has been burning since 1962.

Fun facts on their own.

But because we're RealEstateAPI.com, we couldn't leave it there.

Each one gives us an excuse to talk about an equally weird quirk in that state's real estate tax system.

And those quirks aren't just trivia. Miss one when underwriting a property and it can affect your cash flow, your NOI and ultimately what the property is worth.

So grab whatever you're drinking, find a comfortable spot, and nerd out with us for a few minutes.

Here are five of our favorites.

1. California: The Bear That Isn't There

Look at the California state flag and you'll see a grizzly bear.

Go looking for one in California and you'll have considerably less luck.

The last documented sighting of a wild California grizzly was in 1924. A century later, the animal remains one of California's most recognizable symbols despite no longer living there.[1]

California property taxes have their own way of preserving the past.

In 1978, voters approved Proposition 13.

Instead of continuously resetting a property's taxable value to its current market value, California generally establishes a base-year value when a property changes ownership. From there, increases in that base value are generally capped at 2% per year until another change in ownership occurs.[2]

That produces a strange result.

Imagine two nearly identical houses sitting next door to each other.

One owner bought decades ago.

The neighbor just bought an identical house for $1.2 million.

Their homes may have roughly the same market value.

Their property tax bills can be dramatically different.

Same street. Same neighborhood. Similar house.

Completely different tax burden.

And when the long-time owner eventually sells, the change in ownership will generally trigger reassessment of the property to current fair market value.[3]

The lesson for investors: Never infer California property taxes from the house next door — or assume the seller's current bill will be yours.

2. Oregon: Nike, Pinot Noir…and a Tax Value That Doesn't Behave Like You'd Expect

Oregon has produced two exports that seem unlikely to come from the same place.

One is Nike, which grew from its Oregon roots into one of the world's most recognizable athletic brands.

The other is Pinot Noir.

Beginning in the 1960s, pioneering winemakers began planting Pinot Noir in Oregon despite skeptics who doubted it would work. Today, Pinot Noir is Oregon's flagship wine grape and represents the majority of the state's planted wine acreage.[4]

But Oregon has another unusual creation:

Measure 50.

The system uses a concept called Maximum Assessed Value, or MAV.

Generally, that value can increase by no more than 3% per year, absent certain changes to the property such as new construction, improvements or subdivision.[5]

Here's what makes that interesting.

The value used for taxation can become significantly disconnected from the property's current market value.

So you might buy a property for $900,000 and find that the value governing its tax assessment looks nothing like the number you just wired to the seller.

That distinction is easy to miss if you're accustomed to thinking:

Purchase price = new taxable value.

In Oregon, it isn't that simple.

The lesson for investors: Don't assume what you paid for a property tells you what value the tax system is using.

3. South Carolina: The Tree That Could Take a Cannonball

South Carolina's nickname is the Palmetto State.

There's a good reason.

During the Revolutionary War, American forces constructed a fort on Sullivan's Island using walls made from palmetto logs and sand.

When British warships attacked in 1776, something unexpected happened.

Palmetto wood is soft and spongy. Combined with the sand packed between the walls, it absorbed the impact of British shot and shell remarkably well.[6]

The fort held.

The palmetto eventually became one of South Carolina's most recognizable symbols.

South Carolina's property tax system contains its own lesson about absorbing shocks — particularly depending on how a property is used.

A qualifying primary residence generally receives a 4% assessment ratio.

Other real estate, including many investment and second-home properties, is generally assessed at 6% of fair market value.[7]

At first glance, the difference between 4% and 6% doesn't sound enormous.

It is.

That's a 50% increase in the assessment ratio before you even get to the applicable millage and other differences in how the property is taxed.

For a real estate investor evaluating a property based on the seller's current tax bill, that distinction can dramatically change the economics.

A house that looks inexpensive to carry as someone's primary residence can become considerably more expensive when treated as investment property.

The lesson for investors: In South Carolina, don't just ask what the property taxes are. Ask why they are what they are.

Classification matters.

4. Michigan: The Tax Version of a Michigan Left

Michigan drivers have something outsiders occasionally find baffling.

The Michigan Left.

At certain intersections, you can't simply make a traditional left turn. Instead, you continue through the intersection or turn right and then use a median crossover to make a U-turn.[8]

To go left, sometimes you first go right.

Michigan property taxes can feel similarly indirect.

Under Michigan's Proposal A system, a property's taxable value is generally limited in how quickly it can increase while the same owner holds it.

Then the property sells.

You might assume the taxable value immediately jumps.

It doesn't.

A qualifying transfer of ownership generally causes the property's taxable value to “uncap” in the calendar year following the transfer. At that point, the taxable value is generally set to the property's State Equalized Value.[9]

So imagine buying an investment property in July.

You review the existing tax bill.

You close.

You operate the property for the rest of the year.

Then the next calendar year arrives.

Now the property's taxable value can uncap, potentially changing the tax burden materially.

One important nuance: Michigan assessors aren't supposed to simply take your purchase price and divide it by two. The assessor must determine true cash value using the same valuation principles applied to other properties.[10]

In other words, the tax bill you see when underwriting the property may not be the tax bill you inherit over the long term.

Like the Michigan Left, the destination isn't necessarily where the first turn takes you.

The lesson for investors: In Michigan, the tax consequence of buying a property can arrive in the next calendar year.

That's a potentially expensive detail to miss in an acquisition model.

5. Pennsylvania: The Tax Assessment From Another Era

There's a fire burning underneath Pennsylvania.

In Centralia, an underground coal-seam fire has been burning since May 1962. Efforts to control it over the years met with limited success, and the dangers from gases and ground subsidence ultimately contributed to the relocation of much of the town.[11]

Pennsylvania has another strange way the past can remain very much alive.

Property assessments.

Pennsylvania's property tax system is highly localized, and the relationship between assessed value and today's market value can vary substantially from county to county.

That means the assessed value printed on a tax record may look nothing like today's market value.

That isn't necessarily an error.

Pennsylvania actually maintains something called the Common Level Ratio, or CLR, which measures the relationship between a county's base-year assessments and current real estate market values.[12]

And the differences can be striking.

In published state valuation factors, the relationship between assessment and market value varies dramatically across Pennsylvania's 67 counties.[13]

So a house worth hundreds of thousands of dollars today might carry an assessed value that looks absurdly low to someone unfamiliar with the system.

That doesn't necessarily mean the owner discovered some magical tax loophole.

It may simply mean the assessment system is speaking in the dollars of another era.

Just like Centralia's fire, history has a way of sticking around underground.

The lesson for investors: In Pennsylvania, an assessed value without context can be almost meaningless.

You need to understand the county's valuation framework before the number tells you what you think it tells you.

Why These Weird Rules Actually Matter

It's easy to read all of this as interesting real estate trivia.

It isn't.

Property taxes are an operating expense. And for an income-producing property, an unexpected increase in operating expenses doesn't just reduce the cash you put in your pocket.

It reduces Net Operating Income, or NOI.

And NOI is one of the fundamental inputs investors use to value income-producing real estate.

Consider a simple example.

You underwrite a property expecting annual property taxes of $10,000.

But you miss one of the quirks we've described above — an uncapping in Michigan, a classification difference in South Carolina, or a reassessment triggered by a California sale.

Your actual tax bill turns out to be $15,000.

That's $5,000 less cash flow.

But it's also $5,000 less NOI.

Now assume comparable properties trade at a 6% capitalization rate.

Using the basic income approach:

Property Value = NOI ÷ Cap Rate

So that $5,000 annual mistake isn't necessarily just a $5,000 problem.

At a 6% cap rate:

$5,000 ÷ 6% = $83,333

That means an overlooked $5,000 recurring expense can translate into roughly $83,000 of implied property value.

At a 5% cap rate, it's $100,000.

The mistake compounds

Suppose you buy a property believing it produces $40,000 of NOI.

At a 6% cap rate, that income supports a valuation of roughly:

$667,000

But if the property's true tax burden reduces NOI to $35,000, the same 6% cap rate supports a valuation of only:

$583,000

You didn't just overestimate your annual cash flow.

You may have overestimated what the property was worth.

And if you used leverage to acquire it, the impact on your actual cash-on-cash return can feel even larger because the unexpected expense is being absorbed by the equity you invested in the deal.

This is why seemingly obscure property-tax rules deserve attention during underwriting.

A reassessment date.

A homestead classification.

An assessment ratio.

A taxable-value cap.

A decades-old county valuation framework.

They sound like footnotes.

But sometimes the footnote changes the deal.

The Bigger Point: Property Data Needs Context

Real estate investors spend enormous amounts of time looking at numbers:

Purchase price.

Taxes.

Insurance.

Rent.

NOI.

Cap rate.

Comps.

Assessed value.

The temptation is to treat those numbers as facts.

But a number without the rules that produced it can be misleading.

A California tax bill doesn't necessarily tell you what the next owner will pay.

An Oregon purchase price doesn't necessarily tell you the property's taxable value.

A South Carolina tax bill doesn't tell you what the property will cost to carry if its classification changes.

A Michigan tax bill may tell you what the seller pays today — not what you'll pay next year.

And a Pennsylvania assessment may reflect a valuation framework with little obvious relationship to today's market price.

That's why good real estate data isn't simply about collecting more fields.

It's about understanding the logic underneath the fields.

Because in real estate, the weird little local rule is sometimes the thing that changes the entire deal.

Happy Labor Day!

Sources

[1] California Department of Fish and Wildlife, Human-Wildlife Conflicts: Black Bears. CDFW notes that the grizzly has been extirpated from California and that the last documented sighting occurred in 1924.

[2] California State Board of Equalization, History & Milestones of the State Board of Equalization. Proposition 13 established the base-year-value system and generally limits annual increases in that value to 2%.

[3] California State Board of Equalization, Frequently Asked Questions: Change in Ownership. BOE explains that a qualifying change in ownership generally causes the assessor to reassess the transferred property to current fair market value.

[4] Oregon Wine Board, Oregon Wine History and Oregon Wine Varieties & Grapes. The Wine Board dates the Willamette Valley's Pinot Noir era to 1965 and identifies Pinot Noir as Oregon's flagship variety.

[5] Oregon Department of Revenue, Property Assessment and Taxation. Oregon's Maximum Assessed Value generally cannot increase by more than 3% annually, subject to exceptions for certain property changes.

[6] National Park Service, Fort Sumter and Fort Moultrie National Historical Park. NPS describes the original fort's palmetto-log-and-sand construction and how the materials absorbed British shot and shell during the 1776 attack.

[7] South Carolina Department of Revenue, Local Government Services — Property Assessment Ratios. Primary residences are generally assessed at 4% of fair market value, while other real estate is generally assessed at 6%.

[8] Michigan Department of Transportation, Michigan Lefts. MDOT explains the indirect-left-turn design and notes that Michigan Lefts have been used in the state since the late 1960s.

[9] Michigan Department of Treasury, Changes in Ownership and Uncapping of Property. A qualifying transfer causes taxable value to uncap in the calendar year following the transfer.

[10] Michigan Department of Treasury, Transfer of Ownership Guidelines. The state explains that assessors may not automatically establish assessed or taxable value at half of a property's sale price.

[11] Pennsylvania Department of Environmental Protection, Centralia Mine Fire Resources and Chronology. DEP dates the start of the underground fire to May 1962 and documents subsequent efforts to control the fire and relocate affected residents.

[12] Pennsylvania Department of Community & Economic Development, State Tax Equalization Board, Frequently Asked Questions. The state defines the Common Level Ratio as the relationship between a county's base-year assessments and current real estate market valuations.

[13] Pennsylvania Department of Revenue, Common Level Ratio Real Estate Valuation Factors. The state's published county-by-county factors illustrate the substantial differences between assessed and market values across Pennsylvania.

A year ago, the software industry was worried that AI agents would kill SaaS.

We may have gotten the direction of the threat exactly backwards. The competitive risk may not be that your customers stop using your software because of AI. It may be that your customers want their AI agents to use your software — and you won't let them.

A LinkedIn post caught my attention this week. An executive in commercial real estate, logged into CoStar while running Claude's browser extension. Instead of his search results, he got an interstitial:

"AI Agent Activity Detected."

Disable the extension. Restart the browser. Try again.

The comments were predictable. Dinosaur. Monopolist. Ripe for disruption. That was roughly my first reaction too.

Then I read CoStar's contract, and I think something more interesting is going on.

CoStar Knows Exactly What's Coming

On July 6, 2026 — just six weeks ago — CoStar updated its License Agreement Terms and Conditions. The changes around AI are remarkably specific.

The agreement now says passcodes may not be shared with "any third-party AI services," explicitly including AI agents. It adds a broader prohibition against exposing CoStar's product to AI tools outside of CoStar's own product, including using the data as input to generative AI or retrieval-augmented generation.

This isn't dusty boilerplate written before ChatGPT existed. CoStar went into its agreement six weeks ago and deliberately closed the agent door.

But the language itself isn't the interesting part. The remedy is.

Contract drafting tells you what the drafter is actually worried about, because the remedy has to match the injury. If your fear is data theft, your remedy is termination, injunctive relief, deletion, audit. CoStar has all of those elsewhere in the agreement and doesn't hesitate to use them.

The remedy attached to the new AI-agent sentence is none of those. It's the right to increase license fees "to account for any additional users detected."

Read that again. If your agent touches the data, the consequence isn't that you get cut off. It's that you get billed — because the agent has been counted as an additional user.

That's not a data protection clause. That's a pricing clause.

And it lines up with something that's been in CoStar's agreement for years: everyone at a licensed location who benefits from the product must be a named Authorized User, regardless of how much they actually use it. Not per query. Not per session. Per beneficiary.

CoStar isn't confused about what agents are. They've already decided what an agent is worth, and the answer is: one more seat.

Agents Don't Break SaaS Economics. They Break the Proxy.

That distinction matters, because seat pricing was never sacred. It was convenient.

Human attention is finite. People work a relatively predictable number of hours. Counting the number of humans using software was a reasonably good way to approximate how much software an organization consumed. One analyst can only do so much work; add another analyst, add another seat.

Then somebody gives the first analyst an agent. Now she researches more properties, runs more analyses, monitors more markets and generates more reports without adding another human being. The agent severs the relationship between headcount and consumption — permanently, and in the direction that costs the vendor money.

From CoStar's perspective, you can see the problem clearly. One analyst with a capable agent might eventually do the work that previously required three analysts and three licenses. The better the agent gets, the worse the seat performs as a proxy for the value being consumed. Every efficiency gain the customer captures comes directly out of the seat count the vendor bills against.

So perhaps CoStar isn't being technologically naive at all. Perhaps it's just protecting the meter.

The timing supports that reading. CoStar's Q2 revenue grew 18% to $925 million — the 61st consecutive quarter of double-digit growth — and adjusted EBITDA more than doubled to $184 million. Good numbers. But net new bookings came in at $69 million, down roughly 26% year over year, and the company cut full-year revenue guidance.

Revenue growth is the past selling itself: multi-year contracts, automatic renewal, CPI escalators. Bookings are the future. When bookings decay while the contracted base still compounds, there's exactly one place left to find growth — price and seat count on the customers you already have. An agent attacks both at once.

Of course they closed the door. But there's another way to look at exactly the same phenomenon.

What if one customer with ten agents isn't a platform cannibalizer? What if it's an explosion in consumption?

The problem may not be the agent. The problem may be what you're measuring.

We Were Worried About the Opposite Problem

This question feels familiar because a year ago I was writing about whether SaaS would survive AI at all.

The argument, which Satya Nadella had put into circulation on the BG2 podcast in late 2024, made intuitive sense. A lot of SaaS is ultimately CRUD — Create, Read, Update, Delete — wrapped in a user interface and some business logic. If an AI agent can interact directly with databases and execute that logic itself, why does the user need the application?

We took it seriously. RealEstateAPI provides data infrastructure to hundreds of PropTech and FinTech companies, many of them SaaS businesses. If SaaS was facing an extinction-level event, we had a front-row seat. So we analyzed our customer base and built a framework for measuring AI resilience.

Our conclusion was different. The CRUD framework wasn't really the problem — the question was who controlled the logic. As I wrote at the time:

"If a product's intelligence is hardcoded by its developers, AI can likely replicate it. But if users can define and evolve their own logic through custom workflows, tools and services, or conditional automations, then the platform becomes more defensible."

There was a second reason we were skeptical of the extinction thesis. Good vertical SaaS contains an enormous amount of accumulated domain knowledge. A generic AI might know how to query a database. That doesn't mean it understands the trade-offs a loan officer makes when underwriting a borrower, the signals an investor uses when evaluating a distressed property, or the sequence of decisions a transaction coordinator makes to get a deal across the finish line.

Agents wouldn't eliminate those applications. They would need them.

A year later, I think we underestimated the implications of our own conclusion. Because if agents need specialized software, the next question becomes obvious:

Why wouldn't we want them using as much of it as possible?

From Blocking Agents to Recruiting Them

Software companies have three basic choices.

The first is Block — detect agents and shut them down. The second is Tolerate — let customers bring Claude or ChatGPT, authenticate it, establish permissions, and let it interact with the product.

The third interests me much more: Recruit.

Design the product from the beginning on the assumption that some of your most valuable users won't be human. Then go one step further. Don't merely let customers bring agents — give them agents.

Imagine buying a real estate investment platform and, instead of one username and a link to the documentation, receiving a team:

• Research Agent — find potential opportunities
• Underwriting Agent — analyze the economics
• Monitoring Agent — watch properties and markets for changes
• Reporting Agent — prepare investment committee materials
• Workflow Agent — move information between systems and initiate the next action

The customer can still bring Claude; open ecosystems will matter. But the software company has something Claude doesn't have on day one. It knows exactly how its own product should be used.

Think about how much money SaaS companies spend transferring that knowledge. Documentation, webinars, certification programs, customer success teams, implementation consultants, tutorial videos. All of it exists because the vendor possesses knowledge the customer doesn't: how to use the product optimally.

In an agentic world, some of that knowledge becomes executable. Instead of here is our software and 47 pages explaining how to use it, you can say: here is our software, and here is a pod of digital employees already trained to use it.

Your customer's first day no longer begins with an empty dashboard. It begins with a workforce.

And the objective changes. For twenty years we tried to maximize human adoption. In the next era, we may be trying to maximize authorized machine consumption.

CoStar Already Knows This. They're Just Doing It in One Direction.

Here's the part that convinced me the "dinosaur" reading is wrong.

CoStar is not a company that fears AI. They shipped Homes AI. They shipped Apartments.com AI in June and logged more than 500,000 AI sessions within weeks, with users averaging roughly 20 minutes per session — several times longer than non-AI users. Their CEO has told analysts they're capturing more in AI-driven cost savings than they're paying in token costs, and that they're running under budget on inference. Goldman has them tagged as an AI productivity winner.

This is one of the more aggressive AI adopters in real estate.

So look at the prohibition again. It bars exposing the product to AI tools outside of the CoStar Product. That carve-out is the entire clause.

CoStar recruits agents when it owns them and blocks them when you do.

That's not technological naivety. That's a strategic decision that AI is a product they sell rather than an interface they serve. Every argument I just made about recruiting agents, CoStar is already executing — on the consumer side, where the agent drives engagement they monetize. They simply haven't extended it to the professional side, where the agent would compress the seat count they bill.

Which means the capability isn't the constraint. The business model is.

Where I Part Company

This is where I disagree with CoStar's current approach — not because they should let Claude scrape their website indiscriminately. They shouldn't.

Data owners have legitimate interests to protect. We certainly do. We rate-limit access, restrict redistribution, monitor usage, and revoke credentials when necessary. Anyone building valuable proprietary data and letting autonomous agents run wild against it won't have valuable proprietary data for very long.

The answer isn't unrestricted agent access. It's managed agent access.

Agents should authenticate. They should carry explicit permissions. Their activity should be attributable to a customer and logged. Vendors should be able to meter it, rate-limit it, price it, and revoke it.

In other words: don't build a better bot detector. Build a bot-shaped front door — and charge appropriately for what comes through it.

The blocking approach fails on its own terms anyway. Detection at the browser session gives you exactly one bit of information — is something automated touching this page — and one bit in means one bit out. Allow or deny. There's no dial. Detection is probabilistic, so you generate false positives against paying customers doing legitimate work, which is precisely the screenshot that irritated that LinkedIn user. Enforcement is binary, so sophisticated users route around it and compliant users just eat the friction. And you get no telemetry from either. You've traded a measurable problem for an invisible one.

The Part I Have More Trouble With

One provision in the new agreement is harder to reconcile with the idea that this is simply sophisticated protection of proprietary data: the specific prohibition on using CoStar data as input to retrieval-augmented generation.

RAG isn't model training. The model doesn't absorb CoStar's database into its weights. It retrieves information the user is already authorized to access, at the moment it's needed, and reasons over it. That's much closer to reading with assistance.

Prohibiting AI-assisted reading is a different thing from prohibiting data theft. Today that distinction may seem academic. I don't think it will for long. We're approaching a world in which asking an AI to read something, compare it to something else, extract what's relevant and help you reason about it isn't a specialized activity. It's just how we read.

A data company can reasonably say 'you can't copy our database, or you can't train your model on our intellectual property, or you can't resell our data. But you may read this yourself, and the AI working on your behalf may not help you understand it' is going to be an increasingly difficult line to hold.

The Browser Is the Wrong Battlefield

There's a larger implication here for data businesses. For twenty years we competed partly on interfaces. Who has the better search, the better map, the better filters, the prettier dashboard.

Now imagine telling an agent: find every industrial property in Northern New Jersey over 100,000 square feet where rents have increased more than 15% in three years, ownership has held the asset for at least a decade, and the mortgage matures within 24 months.

If the agent can execute that request, I care considerably less about your filter panel.

The interface hasn't disappeared. It has moved. The agent is becoming the interface, and that shifts competitive advantage toward what sits underneath it — proprietary data, domain expertise, unique workflows, user-defined logic, reliable infrastructure, and the ability to expose all of it safely to machines.

Which means companies that embrace agents don't have to give away their moats. They may make those moats considerably more valuable.

SaaS Doesn't Die. But the Seat Might.

A year ago the industry was asking whether agents would kill SaaS. I don't think they will. Agents still need specialized data, domain-specific workflows, business logic, and the accumulated intelligence embedded in good vertical software.

But they will force us to rethink how we package and price access to all of it. The winners won't merely possess things agents need — they'll make those things easy for agents to consume. Eventually the smartest SaaS companies will go further: actively recruiting agents as users, optimizing for machine consumption, letting customers bring their own agents, and handing them preconfigured agents designed to extract more value from the software than a human could alone.

For decades, software companies have asked: who is authorized to access our product?

The more important question for the next decade may be: on whose behalf is the agent acting?

Answer that correctly — technically, contractually, and economically — and the economics start to look very different.

Agents may not kill SaaS. They may kill the seat.

And if software companies learn to meter what replaces it, that might be very good news for SaaS.

Disclosure: I'm the CEO of RealEstateAPI, a property data platform. We operate in adjacent territory to CoStar and have an obvious stake in how this question gets answered. Contract language is quoted from CoStar's publicly posted License Agreement Terms and Conditions effective July 6, 2026, compared against the archived February 2025 version; both are available on costar.com. Financial figures are from CoStar's Q2 2026 earnings call and public market data as of mid-August 2026.

Last August, at Inman Connect, Clelia Warburg Peters said something that stopped me cold: it doesn't make sense to run a brokerage without ancillary services anymore.

I wrote it up the same week. The argument was simple. Commission margin was collapsing under volume pressure, agent splits, and NAR settlement fallout. Rocket's CEO had just put a number on the endgame — cut the average cost of a transaction in half, from $40,000 to $20,000.

And that number raised a question nobody at the conference seemed eager to answer out loud:

Where does that lost revenue get recaptured?

My answer was that brokerages would become holding companies. Title, mortgage, insurance, home services, data. Multiple profit centers, because the math on a $20,000 transaction doesn't work if commission is your only one.

Own more of the consumer relationship, or risk becoming a commodity onboarding service for agents.

Q2 2026 earnings came out last week. The recapture question has an answer now.

It's in the segment tables.

And it's more aggressive than what I described.

Zillow lost audience and grew revenue 18%

Start with the number that shouldn't be possible in a flat market.

Zillow's Q2 2026 revenue hit $772 million, up 18% year over year. Average monthly unique users fell 2%, to 239 million. Total visits fell 2%.

Fewer people showed up.

On a rough revenue-per-visitor basis, each one was worth about 20% more.

The composition is where the story lives.

For Sale revenue grew 14% — and the Residential piece of it, the core listings-and-leads business everyone in this industry means when they say "Zillow," grew 7%.

Mortgages grew 75%, to $84 million. Rentals grew 31%, to $209 million, with multifamily up 42%.

Seven percent in the legacy business. Seventy-five in mortgages. Same company, same quarter.

Mortgages and rentals together are now about 38% of company revenue. A year ago they were about 32%.

Six points of revenue mix in twelve months.

The housing market may be flat. But that almost misses the point.

Zillow isn't just outperforming the housing market. It's outperforming its own dependence on the housing market.

That is the ancillary thesis executed at scale — except Zillow arrived there without starting as a brokerage at all.

Rocket bought the demand instead of building the stack

Rocket ran the same play from the other end, and it's the cleaner proof.

It spent $1.75 billion on Redfin.

Not simply for the brand.

For the shoppers.

A year in, the mortgage attach rate on Redfin-originated purchase business is 47%, against a 50% synergy target management set at close. Redfin-driven mortgage leads doubled year over year in June. Rocket posted $2.78 billion in net revenue and its most profitable quarter in four years.

Read that attach rate again.

A buyer who comes in through the portal is now close to a coin flip to finance through the company that owns the portal.

That's the $20,000 transaction being funded.

Not by defending every dollar of brokerage economics, but by owning enough of the surrounding stack that you can afford to compress one part of the transaction because you have more places to make money across the customer relationship.

I said the math required multiple profit centers.

It does.

What I didn't say is who would be standing in the middle of them.

What I got wrong

I put the vertical integration at the brokerage layer.

It's happening one layer up.

The holdco isn't the brokerage.

It's the platform that owns demand.

And that is a materially harder problem for everyone downstream, because a brokerage that vertically integrates still has to go buy its customers — increasingly from a company that has already integrated vertically, owns the demand, and decides how that demand gets routed.

Vertical integration is powerful. Vertical integration plus distribution is something else entirely.

If I own title but you own the consumer relationship, I'm downstream from you.

If I offer mortgages but you decide which lender the buyer sees first, I'm downstream from you.

If I run a brokerage but my buyers increasingly originate inside somebody else's portal, I'm renting my distribution.

And whoever owns that distribution can keep adding ways to monetize the relationship.

The companies controlling the funnel are optimizing revenue per visitor, not visitor count.

Zillow's numbers show it working.

And nothing about a 47% mortgage attach rate suggests portal-sourced demand stays neutral.

The second signal nobody's reading

Everyone is going to write about the mortgage number.

The one I'd stare at longer is rentals.

$209 million in revenue, up 31%. Multifamily up 42%.

That makes rentals more than twice the size of Zillow's mortgage segment.

Why does that matter?

Mortgage still needs someone to buy a house.

Rentals don't.

Rental marketplaces, property-management workflows, resident services, and the data generated around them create recurring economic relationships that don't depend on transaction volume.

When a $209 million quarterly business growing 31% is also the one least dependent on home-sale transaction volume, that tells you where the stack goes after title and mortgage saturate.

A year ago, I listed data and software as the underrated pieces of the ancillary stack.

I'd revise that upward. As these platforms move from one-time transactions into recurring relationships, the data and workflow layer underneath stops being ancillary and starts being the business.

That deserves its own post, and it'll get one.

Where I land, a year later

The core call holds:

You either own the relationship or you rent access to it.

What changed is the scale of who's doing the owning.

I thought brokerages would respond to deteriorating transaction economics by becoming vertically integrated holding companies.

They are.

But the more important race is happening above them.

Zillow is moving downstream from search into financial services.

Rocket moved upstream from financial services into search.

They are approaching the same destination from opposite directions:

Own the demand. Own more of the stack. Monetize the relationship several times.

And if you're building downstream from those platforms, your strategic response can't simply be to add more features.

You need to think carefully about what you actually own.

Your customer relationship.

Your workflow.

Your data.

Because those are the things somebody upstream can't take from you — unless you architect your business in a way that gives them away.

That's obviously not a neutral observation coming from me. We built RealEstateAPI around the idea that companies should control the property-data layer underneath their products. But the case was on the record a year before these earnings releases confirmed it, which is the only reason I'm comfortable restating it now.

One more called shot, so this is checkable next August:

By Zillow's Q2 2027 release, mortgages plus rentals will represent at least 45% of total revenue.

If I'm wrong, it'll be right here, with the numbers.

Where does the $20,000 go?

It goes to whoever owns the customer after the click.

A year ago, that was a question.

It isn't anymore.

Sources

• Zillow Group Reports Second-Quarter 2026 Financial Results — official release (Aug. 5, 2026)www.prnewswire.com
• Rocket Companies Announces Second Quarter 2026 Results — official release (Aug. 5, 2026)www.prnewswire.com
• Rocket gains purchase and refinance market share in Q2 2026 — HousingWirewww.housingwire.com
• Zillow to lay off 500 staff in restructuring — HousingWirewww.housingwire.com
• Zillow Q2 2026: Revenues Up 18% As 500 Layoffs Announced — Online Marketplaceswww.onlinemarketplaces.com
• Z Stock Layoffs: What to Know About the Latest Zillow Job Cuts — Yahoo Financefinance.yahoo.com
• Zillow Lays Off 500 Employees Ahead of Q2 Earnings Announcement — RISMediawww.rismedia.com
• Zillow cuts 500+ roles ahead of Q2 earnings release — Real Estate Newswww.realestatenews.com

Realtor.com's 2025 purchase data shows institutional buyers retreating to their smallest share since 2011, while small investors quietly took over.

Everyone's Still Furious at "Wall Street Landlords." The 2025 Purchase Data Says They Already Left.

Search "corporations buying up all the homes" and you'll find a permanent, boiling internet argument — Reddit threads, cable news segments, state legislatures passing purchase caps to fight it. Meanwhile the actual 2025 purchase data just landed, and it tells a different story: the mega-institutional investors everyone pictures when they say that just posted their smallest share of the investor market since 2011, down nearly 70% from their 2021 peak.

The buyers who took their place weren't other institutions. They were people who look exactly like the person reading this article. Small investors — landlords buying ten properties a year or fewer — made up roughly 63% of every home an investor bought in 2025, the highest share in more than fifteen years, according to Realtor.com's 2025 Investor Report, published June 23, 2026.

This isn't a footnote. It's a full changing of the guard, and almost nobody arguing about institutional buyers online has caught up to it. If you've been sitting on the sidelines because you assumed you'd be outbid by a hedge fund with a bottomless line of credit, the market you're picturing stopped existing a while ago.

The Investor Market Didn't Shrink. It Changed Hands.

Total investor purchases actually held up well in an ugly year — 534,000 homes bought in 2025, up 0.7% from 2024, even as overall (non-investor) home sales fell 2.1% to a multi-decade low. Investors as a group grew their share of all purchases to 11.3%, from 11.0% the year before. So the "investors are buying more of the market" headline is technically true. What it hides is which investors.

Break the 534,000 purchases down by size and the picture flips. Mega investors — the 351-purchases-a-year-and-up tier, the actual definition of an institutional landlord — made up just 7.5% of investor purchases in 2025, down from a 2021 peak north of 16%. Large investors (51–350 purchases) are down roughly 30% from their peak and have flipped to net sellers. Meanwhile small investors (1–10 purchases) bought about 53,000 more properties than they sold, and did it at a median price of $330,000 — roughly 25% below the overall market median of $440,000, meaning they're concentrated exactly where first-time buyers and the "starter home" panic live.

Realtor.com senior economist Hannah Jones put it plainly in the report: "With small investors now comprising nearly two-thirds of all investor purchases and large institutional players continuing to pull back, the dynamics shaping competition in entry-level housing are shifting." RealEstateNews.com's coverage of the same release called mega-investor activity a 14-year low outright.

Here's the mechanism behind the reversal, and it's not mysterious: mega investors buy with leverage and exit fast when the math stops working. Rates near 6.5-7% for three straight years crushed the spread between cap rates and financing costs that institutional single-family portfolios were built on in 2021. Small investors mostly don't lever the same way, don't answer to a fund's quarterly return targets, and were never priced out of the entry-level tier the way a 2021-vintage $60,000/door acquisition target was. When the trade stopped working for Wall Street, Wall Street left. Main Street didn't have a trade to stop.

The Numbers

The clearest way to see this is metro by metro. Investor buyer share still runs hot in the traditional cash-flow Midwest/South markets: Memphis leads the country at 23.7%, followed by Kansas City (21.2%), St. Louis (21.1%), Birmingham (21.0%), and Oklahoma City (17.9%). If you're buying in any of those five metros, you are competing directly with the deepest-pocketed players left standing — that's real information, not trivia.

At the other end, investor buyer share is lowest in Portland (5.8%), Sacramento (6.1%), Hartford (6.1%), Seattle-Tacoma-Bellevue (6.4%), and San Jose (7.1%). Portland and Sacramento make sense — high price points scared everyone off. Hartford doesn't, and that's the data point worth sitting with: Realtor.com's Top Housing Markets forecast ranks Hartford-West Hartford-East Hartford as the #1 market in the country for combined 2026 sales-and-price growth at 17.1%, ahead of Rochester (15.5%) and Worcester (15.0%). RISMedia's mid-year tracking, published July 15, 2026, confirms Hartford has actually held the #1 spot through June — this isn't a forecast that fizzled, it's tracking live.

Put those two facts next to each other: the market Realtor.com projects to grow the fastest in the country in 2026 also has close to the least investor competition of any major metro measured. That combination — high forecasted growth, rock-bottom institutional and even small-investor density — doesn't show up on its own in either report. It only shows up when you cross-reference them, and as far as I can tell nobody else has.

Common Mistakes Investors Make Here

• Fighting a market that's already gone. Passing on a deal because "the institutions will outbid me" made sense in 2021. In 2025, that outbidder had a 70% smaller checkbook than three years earlier — verify current investor density before you assume you're outgunned.
• Chasing high-investor-share metros out of habit. Memphis, Kansas City, and St. Louis are proven cash-flow markets, and there's nothing wrong with them — but "proven" now means "crowded." A 23.7% investor share means nearly one in four transactions is against another investor, often one who's done this longer than you.
• Ignoring refuge markets because they're not the usual suspects. Hartford, Rochester, and Worcester don't show up on the average BiggerPockets top-ten list. That's exactly why the competition is thinner.
• Confusing "small investor" with "unsophisticated." The data shows small investors buying at a 25% discount to market median with more purchases than sales for three straight years. That's not luck — that's a segment that's been quietly out-executing the big guys since rates rose.

Where I Land

The "corporations are buying up America" narrative was mostly true in 2021 and is mostly false in 2025 — and I'd bet real money that most people still repeating it haven't looked at a single purchase-tier chart since then. If I had capital to deploy this quarter, I'm not spending another minute worrying about getting outbid by a fund in Memphis or Kansas City — I'm running the numbers on Hartford, Rochester, and Worcester, where the growth forecast is the best in the country and the investor competition is close to the thinnest. The market didn't get harder for small investors over the last three years. It got easier, and almost nobody noticed because the headlines never caught up to the purchase data. Bet on what the transactions are actually doing, not on what the comment section is still angry about.

How to Use PropGPT for This

"Rank the top 20 U.S. metros by projected 2026 home price and sales growth, then cross-reference with current investor purchase share by metro, and show me only markets in the top half for growth and bottom half for investor density." This is the exact screen behind this article's thesis — it surfaces high-growth, low-competition markets before you spend a weekend building the spreadsheet by hand.

"Find active listings in [metro] priced at or below $330,000 with at least 3 comparable rentals showing positive cash flow at a 6.5% rate." Small investors' median purchase price nationally was $330K — this prompt puts you in the same buy box that's actually been winning.

"Pull the last 12 months of sold comps in [zip code] and flag what share of buyers appear to be repeat/LLC purchasers versus individual owner-occupants." A rough, deal-level way to sanity-check investor density in a specific neighborhood, not just a metro-wide average.

"Compare cap rate and days-on-market trends for Hartford CT, Rochester NY, and Worcester MA over the last 6 months." Turns the refuge-market thesis in this article into a side-by-side you can actually underwrite against.

"Alert me weekly to new listings in [metro] under $350,000 that have been on market 45+ days with no price cut." Aged, un-discounted listings in a low-competition metro are often the clearest sign a seller hasn't adjusted expectations yet — worth a lowball before anyone else notices.

The Bottom Line

The purchase data for 2025 shows an investor market that quietly rebuilt itself around people buying one to ten properties a year, while the institutions everyone's still angry at retreated to their smallest share since 2011. The opportunity isn't in outbidding a hedge fund — that fight barely exists anymore. It's in finding the handful of high-growth metros, like Hartford, that the crowd hasn't priced up yet because the crowd is still fighting yesterday's war. Pull the screen, run the numbers, and go compete in a market where you're actually the deepest pocket in the room.

A few months ago, homeowners in Delaware opened property tax notices that looked less like routine municipal paperwork and more like a margin call.

For decades, Delaware had been taxing homes using old values. New Castle County’s last reassessment dated back to 1983. Sussex County’s went back to 1974. The market moved. Neighborhoods changed. Homes appreciated. But the tax roll stayed frozen in time.

Then the courts forced the state to mark the books.

In New Castle County, average residential assessments rose by roughly 477%. Non-residential assessments rose by much less. The tax burden shifted toward homeowners. Some bills doubled.

It would be easy to describe that as a tax shock. But the shock was not the reassessment. The shock was the 40-year fiction that came before it.

For decades, the paper record said one thing. The market said another. The court simply forced the government to reconcile the two.

That is what property tax reassessment really is: not administrative housekeeping, but a forced mark-to-market of real estate’s most ignored carrying cost.

The quiet fiction inside the tax bill

Most people assume a property tax assessment is a current estimate of value. In many places, it is not close.

Pennsylvania is the clearest example.

In a base-year assessment system, a county picks a year, values every property as of that year, and freezes those values. Counties, municipalities, and school districts can still change millage rates, but those rates apply across the jurisdiction. The individual property value does not automatically update just because the market changes.

This is not California’s Proposition 13, where assessment growth is capped and resets on sale. In Pennsylvania, a new buyer generally inherits the old assessment. A home may have sold yesterday, but its taxable value may still be anchored to an old market.

That sounds taxpayer-friendly. Sometimes it is. But only if your property appreciated faster than the county average.

Neighborhoods do not move in lockstep. A frozen tax base preserves the distribution of value from the year the county stopped looking. Over time, properties that appreciated least become over-taxed relative to market value. Properties that appreciated most become under-taxed. In many places, lower-value homes in weaker markets carry a heavier effective tax burden than higher-value homes in stronger markets.

That is the heart of the lawsuits now working through Pennsylvania.

Pennsylvania is the tell

Pennsylvania is now the state to watch because it is the last major holdout for indefinite base-year assessments.

Roughly 50 of its 67 counties are using assessment years at least 15 years old. Franklin County’s base year traces to 1961. Lackawanna County’s was 1968 before its recent reassessment.

The state publishes one number that reveals the drift: the Common Level Ratio, or CLR. The CLR is the median ratio of assessed value to market value across valid sales in a county. If a county’s CLR is 100%, assessments are roughly aligned with current market values. If it is 50%, properties are assessed at about half of market value. If it is 10%, the tax roll is not just stale. It is an archaeological artifact.

Pennsylvania’s 2025 CLR table is remarkable. Bucks County is at 5.6%. Butler is at 6.0%. Franklin is at 7.2%. Westmoreland is at 8.7%. Pike is at 8.9%. At the other end, counties that just reassessed reset to 100%.

So within one state, some counties are marking property close to market, while others are taxing homes at a nickel or dime on the dollar. That spread tells you where the paper record has drifted farthest from economic reality, where lawsuits are likely to keep coming, and where the next tax bill may already be hiding.

The mark-to-market tax

In capital markets, we understand the difference between book value and market value. If you hold an asset on the books at an old price, and the market moves, the mark eventually matters. The exposure exists whether or not the statement reflects it yet.

Real estate has its own version of that problem.

Call it the mark-to-market tax.

A property’s current tax bill is based on assessed value and millage. But its future tax exposure depends on how that assessed value compares to true market value and to the county’s common level ratio.

If a property is assessed below the county norm, it may be carrying a hidden liability. If it is assessed above the county norm, it may be carrying a hidden asset, including a possible appeal opportunity today.

Take a recently sold house in Point Breeze, one of Pittsburgh’s stronger residential neighborhoods. Round numbers: it sold for about $335,000. Its assessed value is roughly $104,000. That means it is assessed at about 31% of market value. Allegheny County’s current CLR is 49.3%.

So this house is under-assessed relative to the county’s own median relationship between assessment and market value. Its current tax bill is about $2,700, or roughly 0.8% of market value. If the parcel were marked closer to the county norm, the annual tax bill would move toward roughly $4,300.

That is a $1,600 annual gap. For a homeowner, that is household cash flow. For an investor, it is NOI. Capitalized at a 6% cap rate, that hidden tax burden represents something like $27,000 of value.

Nobody can tell you exactly when the mark lands. But you can underwrite it anyway. The risk exists before the reassessment notice arrives.

The mark cuts both ways

Now look at the other side of the ledger.

In McKeesport, another Allegheny County property recently sold for roughly $30,000. Its assessed value is about $26,000. That means it is assessed at close to 88% of market value. Again, the county CLR is 49.3%.

This house is over-assessed relative to the county norm. Its current tax bill is about $1,000 a year, or roughly 3.4% of market value. If it were marked closer to the county relationship between assessment and value, the bill would fall toward roughly $570.

Put the two houses next to each other and the inversion is hard to unsee. The Point Breeze owner is paying roughly 0.8% of market value. The McKeesport owner is paying roughly 3.4%.

That is about a four-times difference in effective tax burden, produced not by a special assessment, not by a local surcharge, not by anything the owner did, but by a frozen snapshot from 2012.

The cheaper house subsidizes the pricier one. That is the system working as designed. And that is exactly what the lawsuits are attacking.

Sometimes the stale mark is an asset

I had the same reaction many owners will have when they first understand this system.

I own a few rentals on the Northampton side of Bethlehem, Pennsylvania. When I first learned how far some counties’ tax rolls had drifted from market reality, my instinct was defensive. Maybe I had a big tax bill waiting for me. Maybe the frozen assessment was quietly flattering my returns.

It turned out to be the opposite.

My property appears to be modestly over-assessed relative to Northampton County’s common level ratio. In other words, the stale mark may be an asset, not a liability. I am now mulling an appeal.

That matters because I had been viewing the property as a marginal hold. The town has not allowed me to add the units to the parcel that I first envisioned when I underwrote the deal, and I had even considered selling. But the prospect of an assessment appeal changes the keep-versus-sell math, at least at the margin.

Any recovered tax drops straight to NOI. Not once. Every year.

A stale assessment is not automatically good or bad. It is a mispriced position. Sometimes the mark is coming for you. Sometimes it is sitting there waiting to be claimed.

Why almost nobody sees it

The data needed to see this problem is public. The parcel record exists. The sale exists. The assessment exists. The tax amount exists. The county ratio exists. But the truth is not sitting in one neat field.

In Delaware, after reassessment hit, Spotlight Delaware wanted to show residents how the new values landed by neighborhood. To build a usable map, it had to file public-records requests across three counties, work with a data lab, and spend roughly $10,000. The public record was technically public. But it wasn't usable.

Even a clean parcel record can mislead you. A field labeled “market value” may simply echo the old assessed value in a base-year county. An “assessment year” may tell you the current roll year, not the actual base year. A tax amount may capture only one layer of the bill, while the school district carries most of the burden.

How to see the mark before it lands

None of this requires a black-box “reassessment risk” score.

Most of what you need is already sitting in a property-detail response. Pull the parcel’s total assessed value, current tax, and an independent estimate of market value. Then compute the ratio that matters: assessed value divided by market value. That ratio tells you how far the book has drifted from reality.

But the parcel record alone will not tell you everything. The assessment year usually means the current roll year, not the year the value was actually set. To understand how stale the basis may be, you need the county’s base year — an administrative fact published by the county assessor, and one we compiled for Pennsylvania alongside this piece.

Then comes the number that turns “this looks under-assessed” into “here is the dollar gap”: the county’s common level ratio. In Pennsylvania, the state publishes that ratio each year. You can also compute a proxy yourself from assessed-value-to-sale-price pairs, if you have the data depth to do it.

The logic is straightforward. If a property is assessed below the county norm, it may be carrying a latent tax increase. If it is assessed above the county norm, it may be carrying a latent tax benefit, including a possible appeal opportunity.

One caution: the tax field is the one people trust too quickly. Tax divided by assessed value gives you an effective rate, which is what the exposure math needs. But do not mistake that derived rate for statutory millage. And do not assume the tax amount always represents the full bill. In some jurisdictions, the clean-looking figure may capture only one levy layer.

Once those pieces are connected, a portfolio’s worth of latent tax exposure becomes something you can rank, filter, and underwrite.

The assessment roll may be the oldest, least-examined dataset in American real estate. That ledger will not stay hidden forever. The gap between assessed value and market value has become too large to ignore. The regressivity is too visible. The litigation template is now established.

For owners, the current tax bill is not enough. For investors, taxes cannot be underwritten as a static expense line. For lenders, debt service coverage may be flattered by an artificially low public burden. For software builders, the parcel record is not the answer. It is the beginning of the question.

The market already knows what the property is worth. The tax bill may not.

The opportunity is in seeing the difference before the government does.

Most distressed properties don't look stuck anymore.

They look finished.

That's the first thing that changed in this market.

Five years ago, a distressed deal often advertised itself. Obvious equity. Obvious motivation. Deferred maintenance. Tired landlord. Inherited property. A seller who simply wanted out.

Today, every obvious deal has fifty buyers attached to it. The easy spread has been arbitraged away.

What remains are the properties that appear unworkable on the surface — reverse mortgages, unresolved estates, tax issues, old liens, missing heirs, probate confusion, assignment chains nobody bothered to untangle.

The hair is now the moat.

I spend a lot of my time on the worst-conditioned end of the residential market — estate-owned homes, reverse-mortgage encumbrances, properties where the owner of record died years ago and nobody ever cleaned up the title. Almost none of these are as unsalvageable as they appear. They look closed because the easy data sources are shallow, and shallow data is confidently wrong in a very specific way: it shows you the last thing that was indexed, not the thing that's actually controlling the property today.

Here’s the secret: Most deals with this profile end up in the market’s discard pile. Weekend warriors throw these deals away because the surface record looks impossible. But to a serious operator, the question is different: is the deal actually dead, or has no one done the curative work required to bring it back to life?

That gap is the whole game. In this market, the money has moved from being a deal finder to becoming a deal doctor. And triaging these properties is a data problem before it’s anything else.

The Property Wasn't Dead. It Was Unresolved.

Recently I worked through a pair of distressed properties in northern New Jersey.

On the surface, both looked dead. Every aggregated data source agreed on roughly the same conclusion: reverse mortgage complications, estate ownership, clouded title, unknown heirs, questionable economics. In other words: skip it.

But the recorded layer told a different story. And the gap between what the market thinks the property is and what the documents actually say is where the opportunity lived.

One of the properties carried a HECM — a Home Equity Conversion Mortgage, better known as a reverse mortgage. For many investors, that's an automatic disqualifier. The assumption is simple: the balance is too high, the lender will take it back. Nothing to see here…

But reverse mortgages are heavily engineered instruments. The mechanics matter. In many HECM situations, heirs are not personally liable for the deficiency balance, and the estate may satisfy the debt through HUD-governed payoff mechanisms tied to appraised value rather than the raw accrued balance. That distinction matters enormously when interest, servicing advances, and fees have ballooned the note over time.

If you don't understand the rules governing the instrument, you're not underwriting the deal. You're underwriting rumors about the deal.

That realization alone changed the economics materially. But it got more interesting from there.

On one of the properties, the original lender named on the mortgage wasn't the actual decision-maker anymore. The loan had been assigned multiple times, and walking the assignment chain forward eventually revealed the current holder. Obvious in hindsight, invisible in most surface-level property tools.

This is where modern distressed investing breaks down. People search for properties. But difficult deals are often really searches for the current lien holder, the actual decision-maker, the authorized heir, the probate status, the payoff rule that actually governs the debt, and whether the recorded documents prove something different from what the database says.

That's a different skillset entirely. Less lead generation. More forensic accounting mixed with title work.

Mortgages and Discharges Should Reconcile One-to-One

The single highest-yield habit I have on one of these is almost embarrassingly simple. Every mortgage that gets recorded should eventually have a corresponding discharge or satisfaction when it's paid off. So when I build the document chain, I literally line them up: mortgage → discharge, mortgage → discharge.

When the count doesn't match, one of two things is true. Either the lien is genuinely still live, or there's a discharge that exists but was never properly indexed against the property. Both are worth real money to know.

On this deal, a lien looked active in several systems. But when we reconstructed the recorded chain manually, we found that the discharge had been filed years earlier — under a slightly different instrument classification. The mortgage wasn't alive. The indexing was imperfect. To the rest of the market, the property still looked encumbered. That's the kind of asymmetry that makes a deal.

A surprising amount of distressed underwriting is reconciliation work. Was the lien truly released? Was it assigned first? Was the discharge indexed correctly? Was the instrument misclassified? Is the title problem real, or just a retrieval problem?

Experienced title professionals understand this intuitively. But many investors increasingly operate several layers removed from the recorded documents. They consume summarized data products and mistake aggregation confidence for ground truth. Those are not the same thing.

The deeper I get into this market, the more I believe distressed investing is becoming an exercise in archaeological thinking. You are reconstructing institutional history from fragments. The players that become exceptional at curative title work — not just lead generation — will own the next phase of this market.

"Estate Of" Does Not Tell You Who Can Sign

Another major issue involved ownership. The deed reflected an estate.

But "Estate of John Doe" is not a person. It does not tell you whether probate was opened, who the heirs are, who has authority to convey, or whether an administrator exists.

That layer lives somewhere else entirely — in probate filings, surrogate court records, heir identification, intestacy analysis. Eventually we identified the appropriate heir structure and began coordinating the procedural steps necessary to move the matter forward legally.

This is the part of the market many people don't want to hear. The bottleneck in distressed real estate increasingly isn't finding a motivated seller. It's determining who actually possesses the legal authority to act.

That's a very different game.

The Best Operators Solve Problems the Market Avoids

There's another side to this worth naming.

This level of investigative work isn't just an advantage for the investor. It can also create better outcomes for families and estates trapped inside administrative chaos.

A lot of distressed situations aren't simply motivated sellers. They're unresolved institutional messes — heirs who don't understand probate, families overwhelmed by reverse-mortgage paperwork, title defects nobody has explained clearly, properties frozen between servicers, courts, and municipalities.

Many wholesalers market themselves as "problem solvers" while operating almost entirely as marketing organizations. The next generation of distressed operators will look different. The edge comes from the ability to untangle title, coordinate stakeholders, reconcile records, navigate probate, and create a path to legal conveyance where one previously did not exist.

That's good for investors because it moves them away from an overcrowded commodity trade. And it's good for homeowners and heirs because someone is finally willing to do the hard administrative work required to create clarity.

Try This With ChatGPT. Then Try It With an AI That Can Actually See the Property.

This is the part worth sitting with, because it's where the last two years of AI hype meet the ground.

Ask a general-purpose model — Claude, ChatGPT, whatever you like — to help you work a reverse-mortgage estate deal, and it will impress you right up until the moment it can't. It will explain what a HECM is. It will walk you through how assignments work, why an open probate matters, how to think about a tax certificate's redemption clock. As a tutor, it is genuinely excellent. It knows the concepts cold.

Then you ask the only questions that matter for an actual deal:

• Who holds this specific lien right now?
• Is there a discharge on the books for this specific mortgage, or not?
• Who is the living, authorized human who can sign a deed on this specific estate?
  

And the model goes quiet; or worse, it doesn't. A general model has no live connection to the recorded layer for a particular parcel. So it either tells you, correctly, that it can't see that data, or it confidently invents a plausible-sounding answer. In a business where the difference between the originator and the current assignee is the difference between a closed deal and a wasted month, a confident hallucination isn't a minor flaw. It's a landmine.

The concepts are public knowledge. The ground truth about a specific address is not, and it isn't sitting on the open web for a model to crawl. Recorded instruments, current lien holders, the lien stack, tax certificates, ownership graphs — none of that is reliably indexed anywhere a general model can reach.

There is a massive difference between an AI that can lecture about distressed real estate and an AI that can actually investigate a distressed property. The second one requires eyes. Not just brains.

Depth Is the Moat Now, Not Breadth

For years, the real estate data industry optimized around breadth — more records, more coverage, more nationwide scale. That mattered. But the edge has moved.

The question now isn't whether you have the data. It's whether the system can retrieve the actual document, reconcile assignments and satisfactions, identify indexing inconsistencies, surface the current economic stakeholder, and combine property, court, lien, and tax layers coherently.

That's what we've been building toward at RealEstateAPI. Not just broad access to property records, but the infrastructure to let humans — and increasingly AI agents — reason across the messy institutional layers that actually determine whether a deal is real. Our MCP server lets an AI agent call live property intelligence mid-reasoning, and the next layer of work is specifically about current-assignee tracking, discharge/satisfaction reconciliation, and recorded-document retrieval — closing the seam between what's indexed and what's actually controlling the parcel.

Because reasoning without visibility eventually collapses into speculation. The model needs eyes on the parcel. And increasingly, the best operators will too.

The strange thing about these deals is that once you fully untangle them, they often stop looking complicated.

But that's only visible at the end. At the beginning, they look dead.

That's the opportunity.

A look at what's behind their claim of a 630,000-seller gap — and why builders, lenders, and investors need both views.

TL;DR

• Redfin published a striking finding last month: 630,000 more home sellers than buyers nationally — the largest gap in their records dating back to 2013..
• Their model measures who's actively competing in the market right now — not who could enter next.
• I spent an hour pulling REAPI data on the supply that's not on the MLS yet: vacant properties, pre-foreclosures, and absentee high-equity owners.
• In Florida alone: 4.78 million absentee high-equity owners. If just 5% list, that's 239,000 new sellers from one state — nearly 40% the size of Redfin's entire national gap.
• These two views aren't competing. They're complementary. Redfin shows you the stage. We show you backstage.
• Buyers' Market vs Sellers Market is highly regional.
• Anyone with capital exposed to Florida residential should be modeling this pipeline now — not after it shows up in next quarter's MLS data.

A community I'm in (mostly options traders, mostly not real estate people) shared a Redfin chart last week showing 630,000 more home sellers than buyers nationally. It's a striking number. Largest gap on record, dating back to 2013.

The conversation in the chat went where you'd expect: "Is this priced in yet? Is this the start of something?"

I went and read Redfin's actual methodology. Then I ran some queries. Here's what I learned.

Redfin's model is good economics.

Most "market reports" you read are vibes wrapped in confidence. Redfin's isn't.

They use a Cobb-Douglas matching function with constant returns to scale — standard methodology in labor economics, applied to housing. The framework comes from peer-reviewed work by Genesove & Han, Anenberg & Ringo, Piazzesi et al., and others.

The clever part is how they count buyers. They can't count buyers directly — there's no MLS for "people actively house-hunting." So they back the number out using Little's Law. If you know how long the median buyer takes from first tour to closing, and you know how many sales happened, you can infer how many buyers must have been in the market. Like estimating how many people are inside a coffee shop by watching how long each customer takes and how many leave per hour.

So when Redfin says there are 1.36 million buyers and 1.99 million sellers in the market, they're measuring something specific and measuring it well: who's actively competing in the matching process right now.

That's a real, important question. It's just not the only one.

What the model isn't designed to capture

Redfin's "sellers" count is — by their own definition — the number of active MLS listings. That's the right call for measuring market tightness. A vacant property isn't competing for buyers. A pre-foreclosure isn't on anyone's tour schedule. An absentee owner who hasn't listed isn't part of the matching dynamic.

But those properties exist. And under the right conditions, they enter the market.

This is where REAPI lives. We track 159 million+ U.S. properties whether they're listed or not. We can answer a different question: what's the supply pipeline behind the active market?

So I ran the numbers.

Florida, where the buyer's market is loudest

Redfin's data shows Miami as the strongest buyer's market in America — 162.6% more sellers than buyers. The whole state of Florida is a buyer's market.

Here's what REAPI surfaces in Florida right now:

Active MLS listings: 409,209 Of those, sitting 60+ days: 277,419 (67.8% of the market) Of those, with a price cut since March 1: 127,390 (31.1% of the market)

For comparison, Massachusetts (a more balanced market): only 1.0% of active listings have cut price in the same window. Florida sellers are cutting price 31x more often than Massachusetts sellers. This is the cleanest real-time signal of buyer leverage you can pull.

So far, this confirms Redfin's directional finding. Buyer's markets are real, and you can see the leverage in the data.

Now the part Redfin doesn't measure.

The supply pipeline behind the active market

In Florida, REAPI shows:

• 193,067 vacant properties — homes nobody lives in, owners haven't listed
• 44,360 pre-foreclosures — supply that's coming whether buyers are ready or not
• 4,782,572 absentee high-equity owners — owners with significant equity who don't live in the property

That last number is the one to sit with. Almost five million Florida properties are owned by people who don't live in them, with meaningful equity, who could exit anytime. They aren't "sellers" today in Redfin's matching-model sense. They're a latent supply pool.

If just 5% of those owners decided to list — say, in response to another insurance premium hike, a soft hurricane season scare, a tax law change, or rates staying elevated through 2027 — that's 239,000 new Florida sellers.

For context: Redfin's entire national gap between buyers and sellers is 630,000.

A 5% conversion of one state's absentee high-equity pool would be nearly 40% of that gap.

This isn't a prediction. It's a sensitivity. The point is that the supply curve behind Redfin's snapshot is enormous — and any builder, lender, or investor with capital exposed to Florida residential should be modeling it.

What this means in practice

For builders: Florida's already at 5.4 months of supply at current sales velocity (910,338 sales in the trailing 12 months, 409K active). Adding even half the vacant + pre-foreclosure pipeline pushes that to 8.5 months. That's not buyer's-market territory. That's oversupply territory.

For lenders: Pre-foreclosure pipelines are leading indicators. When Florida shows 44K pre-foreclosures while seller's-market states like Massachusetts show 2K, you're looking at a regional risk profile that bulk-data flat files won't surface fast enough to act on.

For investors: The absentee high-equity pool is a behavioral cohort. It moves on insurance costs, on tax events, on inheritance, on retirement timelines. If you're modeling Florida residential exposure without that variable, you're modeling the stage and ignoring backstage.

For proptech operators building on top of this: This is exactly why we built REAPI the way we did. Redfin's approach (matching model + proprietary tour data) is brilliant for what it measures. It's not designed to surface vacant counts or absentee pools or distress pipelines at the city, county, or ZIP level in sub-second response times. That's our job.

The two views, side by side

Redfin asks: "Of people actively competing in the market right now, who has leverage?"REAPI asks: "What's the supply pipeline that could enter the market — and under what conditions?"

Both questions matter. Neither is more important than the other.

If you only watch the active market, you'll be late to every regime shift. If you only watch the pipeline, you'll miss the leverage signals telling you where prices move next.

The teams that win — builders timing land buys, lenders sizing risk, investors rotating capital — watch both.

How long it took to pull this

I ran 14 API calls. Total query time: ~2.5 seconds.

The longest single call returned 4.78 million absentee high-equity records in 276 milliseconds.

That's not a flex. It's the table stakes for the next generation of proptech. If your data infrastructure can't keep up with conversational AI agents asking "what's the supply pipeline behind the buyer's market?", you're already behind.

That's why we built REAPI. And that's why we ship MCP servers. The future of property intelligence is conversational, composable, and grounded in real data — not flat files mailed to you on Mondays.

Want to run these queries yourself? Start a free trial — no sales call. Just the API and the docs. Let us know what you find.

Do more. Build faster. Spend less.— Vince

January was a month about depth, not breadth.

Less “shipping for the sake of shipping.” More getting closer to how real decisions actually get made in the field—and letting that shape what we build.

Below is a snapshot of what we shipped, why it matters, and a couple of operator voices that helped sharpen our thinking along the way.…

Speaking of making things seamless: we also embedded a Context7-trained documentation assistant directly in our docs. Ask questions in natural language, get code examples, troubleshoot errors. Access it here.

More than 2B API calls later, the word is out. RealEstateAPI is the go-to platform for devs building the future of proptech, fintech and home services. In recognition of the great work of our product team, we received stellar marks on G2, Source Forge and other industry publications and nearly doubled our customer count this year.

Recognition is great. But, what matters most is what operators tell us directly: getting access to property data should be easy. That's always been our focus. It’s why we got in this game in the first place, to break the chokehold that legacy players have had on the data you need to build great software. We work every day to deliver products that make your work easier. And that is why we were ranked #1 on G2 for 'Easiest Admin' across the entire Property Intelligence category.

Here’s how we spent the last twelve months…

Building the Property Intelligence Layer for Humans and AI

For Developers: Your IDE Now Has a Property Intelligence Layer

Realtime MLS Data

Community: New Podcast + New Resources

Ground Truth: Conversations with PropTech Operators

Vince

CSV (comma-separated values) is one of the simplest and most durable data formats in use today. Despite the rise of APIs, JSON, and modern data warehouses, teams still routinely convert data to CSV format—and for good reason.

Below are the most practical reasons organizations convert data to CSV, along with examples that reflect how CSV is actually used in modern software, analytics, and business workflows.

Why Convert Data to CSV?

People convert data to CSV format because:

1. CSV works across nearly all tools and platforms
2. It’s easy for humans to read, edit, and validate
3. It simplifies data sharing across teams
4. It’s ideal for importing and exporting data
5. It avoids vendor lock-in and proprietary formats
6. It’s easier to parse than XML or JSON for flat data
7. It supports analytics, reporting, and machine learning workflows

Each of these reasons reflects a real operational need—not nostalgia for an old format.

1. CSV Works Across Nearly All Tools and Platforms

CSV is a universal interchange format.

It can be opened, imported, or processed by:

• Spreadsheet tools like Excel and Google Sheets
• BI and analytics platforms
• Databases and data warehouses
• Programming languages such as Python, R, and Java

Because CSV is not tied to any specific vendor or software ecosystem, it remains the lowest-friction way to move data between systems.

Example: Exporting application data so customers can analyze it in their own tools without additional setup.

2. CSV Is Human-Readable and Easy to Inspect

Unlike binary formats, CSV files are plain text. That makes them easy to:

• Open in a text editor
• Spot-check values
• Debug formatting issues
• Validate data before importing it elsewhere

This is especially valuable when technical and non-technical users collaborate on the same dataset.

Example: Quickly validating an export before uploading it to a third-party system.

3. CSV Is Simpler Than JSON or XML for Flat Data

While JSON and XML are excellent for nested or hierarchical data, CSV is often better suited for tabular data.

Reasons teams prefer CSV for flat datasets:

• No nesting or structural ambiguity
• Predictable columns
• Minimal parsing logic
• Lower cognitive overhead

For many use cases, CSV provides exactly the structure needed—no more, no less.

Example: Exporting rows of transactions, properties, users, or events.

4. CSV Is Ideal for Import and Export Workflows

CSV has become the default import format across many platforms, including:

• CRMs
• Marketing tools
• Accounting systems
• Data enrichment platforms

Its consistent structure makes it easy to validate and map fields during imports, reducing errors and rework.

Example: Migrating data from one SaaS platform to another during onboarding or system changes.

5. CSV Simplifies Cross-Team Data Sharing

CSV lowers the barrier for sharing data across roles and departments.

It allows:

• Engineering teams to export data without custom tooling
• Operations and finance teams to analyze data independently
• Product teams to review datasets without writing code

This makes CSV especially valuable in self-serve and product-led environments.

Example: Providing customer-accessible exports that don’t require developer involvement.

6. CSV Supports Analytics, Reporting, and Machine Learning

CSV is widely used as a staging format in analytics and ML workflows.

Common uses include:

• Loading data into Python or R notebooks
• Feeding datasets into visualization tools
• Preparing training data for models

Because it’s lightweight and well-understood, CSV often serves as the bridge between raw data and analysis.

Example: Downloading historical data for exploratory analysis or model prototyping.

7. CSV Avoids Vendor Lock-In

One of CSV’s most underappreciated advantages is durability.

Because it’s an open, text-based format:

• Data remains accessible long-term
• Switching vendors is easier
• Exports don’t require specialized readers

This makes CSV a safe choice when long-term access and portability matter.

Example: Retaining historical records independent of any single software provider.

When You Shouldn’t Convert Data to CSV

CSV isn’t always the right choice.

You may want to avoid CSV when:

• Working with extremely large datasets
• Handling deeply nested or hierarchical data
• Streaming real-time data
• Preserving strict data types or schemas

In those cases, formats like Parquet, JSON, or database-native storage may be more appropriate.

How Teams Typically Generate CSV Files

Most teams generate CSV files in one of three ways:

1. Direct database exports
2. Scripted generation using programming libraries
3. API-driven CSV generation

While basic CSV creation is straightforward, complexity increases when users need:

• Field selection
• Custom formatting
• Consistent schemas across exports
• On-demand generation at scale

At that point, CSV generation becomes a product feature rather than a one-off task.

Final Thoughts

CSV remains popular not because it’s old—but because it solves real problems simply and reliably.

For teams building data-driven products, providing clean, well-structured CSV exports is often the fastest way to make data usable, portable, and trustworthy for customers.

If you’re delivering data via APIs or applications, investing in a robust CSV generation approach can significantly improve the end-user experience—especially for analytics, reporting, and downstream integrations.

If you want, next we can:

• Tighten this for featured snippet optimization
• Write meta titles/descriptions for CTR testing
• Add internal linking anchors to your existing CSV article

We’re proud to share that RealEstateAPI has been rated #1 for Ease of Admin on G2, earning the highest score in our category for onboarding simplicity, setup clarity, and day-to-day management.

More than anything, we’re gratified.

This distinction comes directly from the people who use our platform every day—developers, data engineers, real estate platforms, and operational teams who rely on us to power mission-critical workflows. Their feedback makes this recognition meaningful.

Why Ease of Admin Matters in Property Intelligence

Property data platforms have historically earned a reputation for being difficult to set up: complex schemas, slow onboarding, restrictive licensing, and heavy admin overhead. These barriers don’t just frustrate teams—they slow down product launches, complicate compliance, and delay revenue.

So when users rate a product #1 for Ease of Admin, it signals something important:

1. Faster Time to Value

Customers report that RealEstateAPI is simple to provision and integrate, allowing teams to get into production quickly without multi-week onboarding cycles.

2. Lower Engineering Costs

A clean data structure, straightforward authentication, and predictable endpoints mean fewer engineering hours spent deciphering documentation or reworking integrations.

3. Confidence in Reliability

Ease of Admin isn’t only about setup—it reflects clarity in how the system behaves, how errors surface, how usage is monitored, and how developers stay in control of their implementation.

4. Better Experiences for End Users

The smoother the admin experience, the more efficiently engineering teams can ship features, serve end customers, and unlock the power of real-time property intelligence.

What Our #1 Rating Says About Our Approach

From the beginning, our goal has been to create a property data platform that removes friction—not adds to it. This recognition affirms that commitment.

Clean, consistent API design

Transparent documentation

Fast provisioning and onboarding

Modern developer tooling

Responsive support

A platform built by people who understand real estate data at depth

Every part of the experience is designed so users can get set up, get confident, and get moving—without the busywork.

Thank You to Our Users

Awards are appreciated. But what matters most is knowing that our customers feel supported, empowered, and able to build without friction. Gratitude is the real headline here.

We’ll continue investing in the things that make RealEstateAPI the easiest way to bring property intelligence into any application.

If you’re evaluating platforms or starting a new project, we’d love for you to experience why users rated us #1 for Ease of Admin on G2.

Last week at Inman Connect, I heard something that stopped me cold.

Clelia Warburg Peters—longtime proptech investor and one of the most perceptive voices in the industry—said:

“It doesn’t make sense to run a brokerage without ancillary services anymore.”

That one sentence perfectly captured the tectonic shift happening in real estate.

It’s not a prediction. It’s a statement of fact.

The old model is gone. What replaces it is already taking shape.

The Margins Are Gone

Transaction volume is down. Commission splits are razor-thin. Regulatory and legal pressures—especially the fallout from the NAR settlement—are shaking the foundations of the traditional model.  

Brokerages that once thrived on agent headcount are finding that it's no longer enough. First membership has cratered (down from 1.6M licensed Realtors in 2022 to just 1.2M as of this writing). Second the agents that remain now demand more support, better tech, and keep a larger share of the commission. Meanwhile, consumers expect seamless digital experiences, integrated services, and transparent pricing.

It is in the midst of all of this that the CEO of Rocket—who recently acquired Redfin—shared a staggering goal:

Cut the average cost of a real estate transaction in half, from $40,000 to $20,000.

That’s not just a margin squeeze. That’s a total overhaul of the economics. And it raises the fundamental question:
Where does that lost revenue get recaptured?

Brokerages Are Becoming Holding Companies

The brokerages that are still thriving understand something fundamental: you can’t rely on commissions alone. You need to own more of the transaction.

Modern brokerage companies aren’t just places where agents hang their license. They’re holdcos—holding companies that deliver services around the transaction, not just through it.

Think title. Think mortgage. Think insurance. Think home services, warranty upsells, moving concierge. These are no longer "side hustles." They’re core to the revenue model.

And here’s where the Rocket stat becomes especially relevant:
The only way to sustainably offer a $20k transaction experience is to control multiple profit centers. The math doesn’t work otherwise. Brokerages that can vertically integrate and earn on every layer—rather than relying solely on commission splits—are the ones that will survive and thrive.

Increasingly, these services are bundled into the consumer experience—not just tacked on by agents.

That’s the big shift.

The Ancillary Stack: A Modern Brokerage’s Profit Engine

Let’s break it down:

Title & Escrow
Still one of the highest-margin pieces of the stack. If you’re not capturing title revenue, you’re leaving money on the table—often thousands of dollars per transaction.

Mortgage
Cross-sell mortgage and you can earn origination fees or referral fees. The top brokerages aren’t just recommending lenders—they’re integrating them directly into the buyer journey.

Insurance
Homeowners insurance is a natural upsell. Smart firms are embedding it into the closing process with licensed partners or in-house teams.

Home Services
From moving services to cleaning, landscaping, storage, and repairs—there’s real money in coordinating post-close logistics. Especially if you can earn a cut on every referral.

Data & Software
This one’s underrated. Brokers who build their own tech stack can not only improve agent efficiency—they can sell those tools externally, or monetize the underlying data.

You don’t need to launch all of these at once. But you do need to be thinking like a vertically integrated business.

The Stakes: Commodity vs. Platform

Here’s the hard truth:

If you're running a traditional brokerage and not building toward vertical integration, you're probably becoming a commodity onboarding service for agents—one that can be replaced when a better offer comes along.

But if you're building the stack? You become a platform. You own the consumer relationship. You drive recurring margin. You create defensibility.

What This Means for Proptech and Data Providers

For those of us in the proptech space—especially companies like ours that provide the infrastructure for real estate innovation—this shift is huge.

Brokerages are no longer just customers for CRMs or agent tools. They're building full-fledged software ecosystems. They need:

Flexible APIs

High-fidelity property data

Ownership verification

Valuation models

Comps and investor logic

Consumer-facing search and decision tools

The new breed of brokerage isn’t using software. They’re becoming software companies. If you’re building tools for this space, build for that reality.

Closing Thought: The Moment to Evolve Is Now

The market doesn’t care what a brokerage used to be. It only cares what it can become.

So the question isn’t whether to expand beyond commissions.

It’s whether you’re building a brand agents and consumers return to once—or a platform they return to forever.

Written by Vincent Harris, CEO @ RealEstateAPI

“The reports of my death are greatly exaggerated.”
— The SaaS Industry (or maybe Mark Twain)

TL;DR

• Microsoft CEO Satya Nadella declared that many SaaS apps — especially CRUD+UI systems — are destined for collapse in the AI agent era.
  
• We undertook an audit of our customer base to assess who’s most at risk — and why.
  
• We created a template for gauging resilience against AI. Ten categories covering everything from workflow ownership to real-time interaction.
  
• We found that smaller SaaS firms outperform their larger peers in one crucial area: supporting real decision-making.
  
• Meanwhile, larger firms excel at building data moats, as expected.
  
• Surprisingly, latency tolerance may favor traditional SaaS over AI agents in certain contexts (at least in the near term)
  
• Our Customer Success team has already reached out to low-resilience customers with support and ideas.
  
• We’re now sharing our template openly with the broader SaaS community — no email gate, no fluff.
  
  

The Comments That Lit the Match

It began with a now-famous exchange on the BG2 Podcast between legendary VC Bill Gurley and Microsoft CEO, Satya Nadella.

Gurley posed a deceptively simple question:

“Do you worry that some of the infrastructure for Excel isn’t necessary if you built an AI-first product? The same could be said about CRM—that there’s a bunch of fields and tasks that can be obfuscated for the user?”

Nadella didn’t flinch. His reply made waves:

“I think the notion that business applications exist—that’s probably where they’ll all collapse: in the agent era. Because if you think about it, they are essentially CRUD databases with a bunch of business logic. The business logic is all going to these agents, and these agents are going to be multi-repo CRUD, right?”

Translation: If your product is mostly a database with a user interface, the future is coming for you.

What Is CRUD + UI?

Let’s unpack this. CRUD stands for Create, Read, Update, Delete — the basic functions of most applications. Combine it with a standard User Interface and you’ve got the typical SaaS tool: forms, dashboards, filters, reports.

If AI agents can now orchestrate multi-database actions across many tools, why would users log into your CRM to “filter contacts” when they can just ask an agent to find prospects who opened an email and booked a meeting in the last 10 days?

To Nadella and others, that kind of UX becomes obsolete.

Why We Took This Personally

At RealEstateAPI.com, we’re the data backbone for leading proptech and fintech SaaS innovators.

So yes, Satya’s comment hit close to home.

We asked ourselves: Which of our customers are most at risk in the AI agent era? Which are most resilient? And what can we learn from the outliers?

To answer those questions, we developed a 10-vector rubric to measure business resilience in the AI era.

The 10 Vectors of Resilience

We rated each SaaS customer (excluding our large Enterprise clients and non-SaaS customers) across the following vectors:

1. CRUD + UI: How much of the product is just a database with a visual interface?
   
   
2. Data Moat / Proprietary Data: Does the product own or generate data that no one else has?
   
   
3. Regulatory / Compliance Function: Does the product help navigate legal or regulatory complexity?
   
   
4. Decision-Making Role: Does it help users make important decisions?
   
   
5. Human Network Component: Does it rely on or enhance human relationships?
   
   
6. Workflow Ownership: Is it embedded into the user’s day-to-day process?
   
   
7. Level of Integration: Does it connect to lots of other tools?
   
   
8. Real-Time Interaction: Is speed essential to how the product delivers value?
   
   
9. Configurability: Can users tailor it to specific use cases?
   
   
10. Latency Tolerance: How forgiving are users when the product takes a moment to respond?
    
    

Each vector was scored from 1 to 5, then weighted based on its strategic impact. Decision-making, data moats, and workflow ownership were determined to be the most impactful.

What We Found

• Across the hundreds of SaaS platforms we support the average resilience score was 31.7 out of 50.
  
• Among our Pro Tier customers (highest standard plan) averaged 36.1
  
• Our lowest tier plans (Starter) came in at 29.3
  

This aligned with our expectations — but the real surprises came when we broke things down by vector.

🎯 Small Companies Excel at Decision-Making Support

Starter-tier SaaS companies — the smallest in our portfolio — scored higher on the “Decision-Making Role” vector than their larger peers.

This shocked us.

Intuitively, we expected small teams to build lean tools that simply store and retrieve data. Instead, we found a cohort of products purpose-built to help users make critical decisions - like whether to bid on a property, pursue a lead, or extend credit.

Niche SaaS products often succeed by being laser-focused on a specific problem or persona. Their users rely on the app to guide them to smart outcomes, not just to store information. These apps are often developed by founders with deep domain knowledge and a close connection to users.

That intimacy shows up in product design. A narrow focus can lead to better alignment with the actual decision-making process. AI agents may be powerful, but if they don’t understand the nuanced trade-offs a loan officer, landlord, or realtor faces day-to-day, they’ll miss the mark.

In many ways, this is the counterpoint to Satya’s warning. Even if a product looks like CRUD+UI on the surface, if it’s instrumental in guiding decisions, users won’t let it go easily.

🧱 Big Companies Dominate on Data Moats

Pro-tier customers performed significantly better on the “Data Moat / Proprietary Data” vector.

No surprises here.

Larger, more mature companies have the time and capital to build feedback loops, buy exclusive datasets, or cultivate proprietary user-generated content. That’s exactly what makes them more resilient — because AI agents need source data, and moats protect margins.

The CRUD+UI Myth (Sort of)

Satya's comments implied that a CRUD app with embedded logic is doomed. But we think that's only part of the story.

We found that it's not the CRUD framework itself that's the death knell—it's about who controls the logic.

If a product's intelligence is hardcoded by its developers, AI can likely replicate it. But if users can define and evolve their own logic through custom workflows, tools and services, or conditional automations, then the platform becomes more defensible.

User-Defined Logic Complexity is the key. The harder it is for AI to anticipate edge cases or user-specific workflows, the more defensible the app.

Platforms that aim to generalize and shield users from complexity will be more vulnerable to extinction. Whereas, platforms that invite users into defining and refining the intricacies of their use cases will retain a strategic edge—because they turn the user into a co-author of the product’s intelligence, not just a consumer of it.

🕒 The Latency Paradox

One of our biggest “aha” moments came from analyzing the Latency Tolerance vector.

At first, we assumed latency was a liability — that is, products where users demand real-time responses would be at greater risk of being outperformed by AI.

But on closer inspection, the opposite may be true.

Why?

Because AI is inherently high-latency due to the need to reason, search and synthesize across mountains of data.

But when a user expects a sub-500ms response (e.g., while performing a desktop underwrite, skip  tracing a lead, or generating a comp), any perceptible delay breaks the experience.

In short: AI may be great at thinking, but it still struggles with instant doing.

This suggests that some seemingly fragile apps — if mission-critical and speed-dependent — might remain defensible longer than expected.

We Took Action

Once we had scored every SaaS customer, our Customer Success team:

1. Reached out to below-average scoring performers to offer support and discuss ways to harden their product against disruption.
   
2. Created this post — not to gloat, not to gate, but to share our thinking with others navigating the same waters.
   

You can get our full scoring rubric — no opt-in required — just message us or visit our website.

What This Doesn’t Cover

A few disclaimers:

• This analysis only includes pure SaaS customers. We also serve institutional investors, home service platforms, and AI firms — but they fall outside the rubric’s scope.
  
• We excluded large enterprise clients. Their implementations are so layered and customized that comparing them to off-the-shelf SaaS would skew the results.
  
• This is a point-in-time snapshot. As AI evolves, so will the scoring framework.
  
  

What Satya Said Next

A few weeks after the viral clip above, Nadella made another appearance — this time with a more balanced message.

He advised founders to:

“Simultaneously make something that is impossible today more possible because of what’s coming… while optimizing the thing you built yesterday.”

That’s been our ethos all along.

And it’s why we’re investing in tools to help our customers evolve — not just survive — in the AI agent era.

If you’re a SaaS founder and you’re curious how resilient your product is, message us. We’ll send you a tidy pdf of the rubric. No pitch. No drip campaign.

Just the truth — and maybe a few ideas to stay on the rocket ship.

🚀

Written by Lukas Curtis, Sales Engineer @ RealEstateAPI

Here's the TL;DR

• We doubled our business last year. We're on pace to do it again.
• We are growing our team to keep up with the increased demand
• Despite economic woes, the real estate industry and its participants is teeming with innovation and chutzpah—and we are helping people convert that hunger for progress into action

By solving real-world problems with cutting-edge technology, we’re helping the Real Estate industry move forward—faster than ever before!

Here's a quick snapshot of the major stories driving residential real estate in 2025...

The resolution of the NAR (National Association of Realtors) lawsuit (Aug 2024) surrounding buyer-side agent compensation sent ripples through the industry, challenging long-held practices and sparking innovation in how transactions are approached.

Companies continue to innovate by making the home buying process faster and more flexible. This has taken the form of new approaches (e.g. Private listings on Compass) and Mergers & Acquisitions, the largest of which you’ve certainly heard in the news—Rocket agreeing to buy Redfin & Mr. Cooper for a combined $11B.

Even title insurance is getting a fresh look, as legaltech and proptech join forces to cut the complexity (and costs) associated with ensuring a clear chain of ownership.

In our practice, we see companies working to tackle these and other challenges everyday. It's inspiring to watch the ingenuity they're bringing to the problems. Truly novel approaches to deeply entrenched processes. In fact, I believe that the next series of proptech unicorns are in our client portfolio right now. No spoilers in this post. But perhaps in another one, I'll drop the names of a few that I'd be willing to bet on.

For better or worse, there are plenty of problems on the horizon that will demand even more creativity. The foreboding fear of defaults looming over both Commercial & Residential Real Estate is still present as the Fed stalls on bringing rates down. But even there, the proptech and fintech shops are preparing solutions to mitigate losses and help homeowners avoid squandering their equity.

In order to stay competitive in this ever-evolving market, real estate professionals need access to robust tools. Property and MLS Data APIs like ours fuel their businesses and give them the transparency they need to protect against risk, cut costs, and seize more opportunities in the multi-trillion dollar Housing Market, no matter their niche.

Going from being a Full-Stack Engineer at RealEstateAPI 3 years ago to now sitting in the role of a Sales Engineer, I’ve seen firsthand the intersection (and innovation) of the underlying tech and business models that are emerging to bring that tech to market—from simple homebuyer-facing apps to complex mortgage applications and many more use cases across the industry.

Just how much demand is there for “PropTech” tools? Let’s look at our booking numbers as we’ve grown and helped hundreds of teams.

💡 From 2023 to 2024, our year-over-year bookings growth was nothing short of phenomenal as we sought to crystallize our Product Market Fit:

📈 May: +641.18%
📈 June: +241.03%
📈 July: +412.90%
📈 August: +717.65%
📈 September: +379.41%
📈 October: +246.81%

And in 2025, we haven’t slowed down at all. We are continuing to compound our progress. With over 2500 bookings in the last year (May 2024 - May 2025), we essentially doubled our total bookings from the end of 2024 (+96%). And thankfully, I have @Alex to help field the calls and extend our capabilities to help more and more companies integrate our APIs quickly and at scale.

These numbers aren't just statistics—they're proof that through constant refinement of our sales motion and listening to our customers, we can continue improving our product and positively impact the workflows of more teams across the industry.

Our early success has been driven by our focus on understanding what success looks like for our customers.

So, as proud as we are of having the fastest most expressive APIs in the industry, we’re even more proud to see that our customers see us as collaborators in the big audacious goals they’ve set for themselves. And we’re honored that both startups and publicly traded companies alike trust us to partner with them on such an important part of their business.

In this next phase of growth, we're taking all the lessons we've learned from 3 years of working closely with dev teams across proptech, fintech, solar, home services and creating a rich, self-service customer portal that puts our expertise directly in developers' hands. This isn't just another dashboard – it's the culmination of thousands of customer conversations, feature requests, and integration challenges we've solved together. Our new portal will feature an AI-powered code assistant that generates custom examples on demand, comprehensive usage analytics that help teams optimize their API consumption, and streamlined account management tools that let developers focus on building rather than waiting for support tickets.

We've proven that white-glove service scales relationships, but we also know that great developers want the freedom to experiment, iterate, and deploy at their own pace. This portal represents the best of both worlds—all the institutional knowledge from our sales-led approach, now packaged into an intelligent, always-available platform that grows with our customers' ambitions. Whether you're a scrappy startup burning the midnight oil or a Fortune 500 company with complex integration requirements, you'll have enterprise-grade tools with startup-speed execution.

On a personal note, getting to interact with our awesome customer base and our talented engineering team—@Jose, @Elvan, @Justin, and @David to deliver the best product possible has been one of most fun (and hard!) experiences I’ve had.

And we aren’t done yet, so stay tuned for more!

In today's real estate market, the difference between a 7% mortgage rate and a 3-4% rate isn't just a few percentage points—it's the difference between closing a deal or losing it. With the national average for a 30-year fixed mortgage hovering around 6.85%, savvy real estate professionals are turning to an underutilized strategy that's changing the game: assumable mortgages.

These financial instruments—where the loan "tags along" with the house during a sale transaction—create a rare opportunity in our high-interest environment. It's like offering your clients a time machine back to the historically low rates of 2020-2021. And the best part? As of May 2025, RealEstateAPI (REAPI) data shows there are over 10 million assumable mortgages nationwide, representing roughly 20% of all outstanding home loans, with a significant portion at rates of 4% or below.

In this guide, you'll discover how to leverage assumable mortgages to close more deals, help frustrated buyers, and create new revenue streams for your business—all while positioning yourself as an expert in creative financing solutions.

The Assumable Mortgage Renaissance

We're experiencing a market phenomenon that hasn't occurred at this scale in over 40 years. Take a look at this historical rate chart:

The last time we saw a gap this significant between existing and new mortgage rates was in the early 1980s. Back then, homebuyers were happy to find properties with assumable loans at 12% while banks were offering new mortgages at 15-18%.

Today's market mirrors those conditions, but with even greater potential benefits. A borrower assuming a 3% mortgage instead of originating a new 7% loan on a $400,000 property saves approximately $1,000 per month—unlocking substantial purchasing power and affordability.

What is an Assumable Loan?

Assumable mortgages allow a buyer to take over the seller's existing loan with its original terms, including the interest rate, remaining principal balance, and repayment period. Unlike traditional financing, where the seller pays off their mortgage and the buyer obtains a new loan, an assumption preserves the favorable terms of the existing mortgage.

What Home Loans are Assumable?

Not all mortgages are assumable. Here's a breakdown of the main types:

FHA Loans

• Fully assumable with lender approval
• Requires buyer credit qualification
• Assumption fee: typically 0.5% of remaining loan balance
• Process handled through the current loan servicer

VA Loans

• Fully assumable with lender approval
• Pre-March 1988 loans: may be assumed by anyone
• Post-March 1988 loans: typically requires VA eligibility or approval
• May involve funding fee and qualification requirements

USDA Loans

• Assumable with lender approval
• Must meet rural property requirements
• Income qualification and property eligibility verification required

Conventional Loans

• Generally not assumable, with limited exceptions
• Some adjustable-rate mortgages (ARMs) may be assumable after the fixed period
• Always check loan documents for a specific assumption clause or rider

The Assumption Process

A typical mortgage assumption follows these steps:

1. Verification of assumability: Review loan documents for an assumption rider or clause
2. Buyer prequalification: Ensure the buyer meets the lender's credit requirements
3. Application submission: Complete assumption application with the current loan servicer
4. Underwriting: Lender reviews buyer credentials and property details
5. Closing: Execute assumption agreement and transfer documents

The process usually takes 30-45 days—often shorter than a new loan origination—and costs between $800-$1,500 in fees, significantly less than new mortgage closing costs.

Strategic Benefits by Audience

For Mortgage Professionals

The assumption market represents an untapped opportunity in a challenging lending environment:

• Client retention strategy: Offer assumption expertise when new originations aren't viable
• Lead generation: Position yourself as the go-to expert for this specialized financing option
• Revenue diversification: Earn fees for facilitating assumptions and financing the equity gap
• Cross-selling opportunities: Provide second mortgages or home equity financing for the difference between purchase price and loan balance

For Real Estate Professionals

Agents who master assumable mortgage strategies gain a significant competitive advantage:

• Listing differentiation: Highlight assumable financing as a property's unique selling point
• Buyer prospecting: Attract inventory-starved buyers with the promise of below-market financing
• Transaction creation: Match assumable property inventory with rate-sensitive buyers
• Expert positioning: Become the market leader in creative financing solutions

When showing properties with assumable mortgages, use a simple script. For example:

"This home comes with a financial benefit most properties don't have—the ability to take over the seller's 3.25% mortgage from 2021. With today's rates near 7%, this could save you approximately $950 monthly on your payment and make this home significantly more affordable than comparably priced properties."

For Investors

Assumable mortgages can dramatically improve investment returns:

• Enhanced cash flow: Lower financing costs translate directly to improved monthly cash flow
• Improved ROI: Reduced carrying costs increase overall investment returns
• Competitive bidding advantage: Offer higher purchase prices while maintaining profitability
• Portfolio acceleration: Leverage savings to acquire additional properties faster

How Do I Find Assumable Mortgage Listings?

Traditional Methods and Limitations

Historically, locating assumable mortgage properties has been challenging:

• MLS searches: Limited and inconsistent "assumable" tagging in listings
• Manual research: Time-consuming property-by-property verification
• Direct mail campaigns: Broad targeting with low response rates
• Cold calling: Labor-intensive with minimal efficiency

These approaches yield limited results and don't scale effectively for serious professionals.

The REAPI Solution

REAPI provides the most comprehensive and efficient way to identify assumable mortgage properties:

• Boolean property search: Our simple assumable mortgage filter instantly identifies relevant properties
• Comprehensive coverage: Access to all 10+ million assumable mortgages nationwide
• Detailed loan data: View loan type, origination date, estimated interest rate, and more
• Integration capabilities: Seamlessly connect with your existing CRM and analysis tools

Practical Implementation

To maximize your assumable mortgage strategy, focus on these search parameters:

• Target recent loans: Focus on loans originated between 2020-2022 for the greatest rate differential
• Look for the equity sweet spot: Filter for properties with moderate equity accumulation (typically 15-25%) to minimize the buyer's cash requirement
• Geographic targeting: Combine with neighborhood and market filters to find opportunities in your area
• Loan type identification: Prioritize FHA and VA loans for highest assumption probability

A smart strategy focuses on properties with assumable loans originated within the last 3-5 years. These mortgages have the powerful combination of low rates with manageable equity positions.

Overcoming Common Challenges

Addressing the Equity Gap

The primary challenge with assumable mortgages is bridging the gap between the loan balance and purchase price. For example:

Property market value: $400,000

Assumable loan balance: $330,000

Equity gap: $70,000

Solutions include:

• Seller financing: Negotiate a second mortgage from the seller
• Secondary financing: Arrange a home equity loan or second mortgage
• Cash contribution: Structure deals for buyers with available funds
• Gift funds: Utilize family contributions for qualified buyers

Navigating Lender Requirements

Successfully manage the lender approval process by:

• Early communication: Contact the loan servicer at the beginning of the transaction
• Complete documentation: Prepare comprehensive borrower qualification packages
• Timeline management: Build adequate assumption processing time into contracts
• Professional coordination: Partner with assumption-experienced title companies

Managing Seller Concerns

Address potential seller hesitations about:

• Release of liability: Confirm the assumption includes full release from the original obligation
• Assumption denials: Structure contingencies protecting sellers if the assumption is rejected
• Timeline alignment: Coordinate assumption approval with seller's purchase of their next home
• Transaction security: Use earnest money to protect against assumption issues

Future Outlook and Market Predictions

The value of assumable mortgages directly correlates with the spread between existing and current market rates:

• Rate projections: Most analysts expect elevated rates through at least mid-2026
• Inventory impacts: Assumable properties will command premium pricing in low-inventory markets
• Regional variations: Areas with high concentration of government-backed loans will see the most assumption activity
• Regulatory considerations: Monitor for potential changes to assumption policies on government loans

As long as the rate differential remains above 2%, assumable mortgages will continue to be a powerful transaction tool.

Putting Assumable Mortgage Data to Work

Forward-thinking real estate professionals are already using REAPI's assumable mortgage data to:

1. Create targeted marketing campaigns to owners of assumable mortgage properties
2. Build buyer databases specifically matched to assumable inventory
3. Develop specialized transaction teams focused on assumption deals
4. Position their businesses as creative financing specialists

Unlike traditional financing, the assumable mortgage advantage creates a win for everyone involved:

• Sellers can command higher prices
• Buyers secure below-market financing
• Agents close more transactions
• Lenders generate complementary financing opportunities

Conclusion: Unlocking Hidden Value

Assumable mortgages represent one of the most significant opportunities in today's challenging market. While many professionals focus on the obstacles of high rates and tight inventory, the data-driven approach reveals a $2 trillion pool of below-market financing hiding in plain sight.

By leveraging REAPI's powerful assumable mortgage data, you gain immediate access to this untapped market segment. Whether you're a mortgage professional looking for new business opportunities, a real estate agent seeking competitive advantages, or an investor maximizing returns, assumable mortgages offer a proven strategy to thrive in today's high-rate environment.

Start finding assumable mortgage opportunities in your market. Try our real estate API free for 14 days.