Who Owns the Truth in Commercial Real Estate?

On a February morning in 2020, the lawyers went to war. CREXi, the Los Angeles marketplace that lets brokers list commercial property without a paywall, sued CoStar Group in federal court, alleging that the incumbent used its subscription contracts to lock brokers out of competing platforms — and sent investigators after brokers who shared their CoStar passwords. CoStar, for its part, had already built a decades-long reputation for treating its database like protected territory: it had sued rival Xceligent for alleged data theft in a legal fight that ended with Xceligent in bankruptcy in December 2017.

Strip away the pleadings and the fight is about one question: who owns the truth about what things rent for? Not the truth itself — a rent is a fact — but the organized, verified, searchable version of it. In most industries that question resolves to whoever built the database first. In commercial real estate, where the input data belongs to thousands of competing brokers who also happen to be the customers, it's genuinely contested. CREXi's antitrust case survived motions and reached the Ninth Circuit, which let the core claims proceed — a courtroom signal that even the law is unsure where one company's database ends and an industry's commons begins.

The incumbent's case is a cathedral. CoStar, founded by Andrew Florance in 1987, spent nearly four decades and billions of dollars building the canonical record: the properties, the tenants, the leases, the sales. Nearly $2.7 billion in revenue in 2024 says the industry pays for that canon gladly. CoStar's defenders make a reasonable point: the data is expensive to produce. Verification isn't scraped; it's worked — researchers call buildings, confirm square footages, chase lease expirations. If anyone could take the finished product for free, nobody would pay to produce it, and the commons everyone depends on would rot.

The challenger's case is a bazaar. Crexi, founded in Los Angeles in 2015 by Mike DeGiorgio, bet on the opposite model: let brokers list and market property in the open, give them tools to run deals, and the data generates itself as exhaust. Crexi says its marketplace has listed more than $1 trillion of property since launch — a claim that, whatever its precise accounting, underlines the scale a free-at-the-edge model can reach in a market built on inventory. Its argument is equally reasonable: brokers created these numbers. A model that rents their own output back to them at five figures a year is a toll, not a service.

Moody's chose a third lane. When it bought REIS for about $278 million in 2018, the ratings giant wasn't entering the listing business or the brokerage business. It was securing the feed: rental and vacancy data as an input to credit models, bank stress tests, and CMBS surveillance. Moody's doesn't need to win the marketplace. It needs the numbers, defensible numbers, because a rating built on bad rent data is a lawsuit. That tells you something about where the durable value sits in this industry — not in the transaction, but in the measurement.

There's a serious counterargument to all three combatants: the moat might be dissolving under them alike. County assessor records are online. SEC filings disclose REIT rent rolls in detail. Satellite imagery counts cars in parking lots. Large language models now read offering memoranda and extract lease terms at industrial scale, turning a week of analyst time into an afternoon of compute. The raw material of rent truth was never fully ownable, and the cost of assembling a rival version keeps falling. In five years, maybe nobody owns the truth because everybody has it, and value migrates entirely to workflow — the underwriting tools, the CRM, the deal room — that sits on top of commodity data.

Maybe. But notice that this is precisely what the incumbents keep saying while they keep buying data companies. CoStar bought LoopNet, Apartments.com, STR, Ten-X. The pattern isn't workflow acquisition; it's vault acquisition. When a data incumbent pays a control premium for a rival dataset in the age of AI extraction, it's telling you the extraction story is harder than it looks — that verified history, cleaned and reconciled over decades, is not something a model can hallucinate into existence in a weekend.

Which leaves the odd question of what a name is worth in this fight. Data businesses are trust businesses. A REIT analyst choosing a new rent-benchmark vendor, a lender testing an underwriting feed, a broker deciding where to spend her subscription budget — each is choosing a counterparty they will have to defend to a committee. The category's names do real work: CoStar means canonical, REIS means institutional, Crexi means open. The slot that remains unclaimed is the one that describes the product itself. The history of internet categories is written in such names: Voice.com sold for $30 million in 2019 on the strength of two syllables, and Cars.com was valued at $872 million in the 2017 Gannett deal mostly because the name was the category. When a market's truth is contested, the party that owns the category word owns the default answer to the first question every buyer asks: where do I look?

The legal war will produce a winner eventually, or a settlement, or a slow stalemate. The commercial one is already legible. Truth in commercial real estate is a subscription product, a marketplace exhaust stream, and a credit-model input — three businesses that all, at bottom, sell the same thing: certainty about what a building earns. The next company to sell it will need data, capital, and credibility. One of the three can be bought outright, tonight, for $18,888.