Yelp Sold OpenAI Its Review Moat. The Money Is Still a Claim.

On July 23, an Axios report that Yelp had licensed its reviews, ratings, and business data to OpenAI sent the stock up 8% in a day. ChatGPT would now surface Yelp's content in local recommendations, with Yelp branding, back-links, and a "Request a Quote" button that lets a user contact a service business without leaving the chat. On its face that inverts the threat that has hung over Yelp for years: instead of AI assistants steering people past it, the company becomes the trusted data layer inside them.

That is real distribution, and the strategic logic deserves three seconds of respect. But a distribution deal is not revenue, and the numbers Yelp has reported since the announcement don't yet show it turning into economics. The pop is a narrative read. The delivered proof is still ahead, and the deal's structure leaves a genuine chance Yelp gets paid for the discovery less generously than the market assumed.

What Yelp gave up, and on what terms

The agreement, announced in late July and rolled into ChatGPT through August, licenses Yelp's roughly 330 million reviews along with its business listings to OpenAI. It is non-exclusive, so Yelp can sell the same content to other AI partners the way it already licenses data to Apple Maps and Alexa. That breadth is the hedge: Yelp is treating its archive of human-written reviews as the one asset AI models can't fabricate, and spreading it across every surface where local discovery now happens.

The load-bearing detail is that the money and the customer relationship flow through a feature, not a fee. A quote request submitted from inside ChatGPT counts as a billable Yelp lead, the same economic event as a lead generated in the Yelp app. But the per-lead price is unpublished, the leads carry no source tag so a local business can't tell a ChatGPT lead from an app lead, and OpenAI controls how the content is presented. Management has attached no specific revenue number to the partnership.

Where the money actually sits today

Yelp's second-quarter report, a few days after the deal went live, lays out the gap between the story and the operating results. Revenue rose 1% year over year to $376 million. Services advertising was flat at $241 million; Restaurant and Retail advertising fell 10% to $102 million. Those two lines still make up roughly 91% of revenue, and they are not growing.

The growth is confined to "Other revenue," which nearly doubled year over year to $33 million, about 9% of the total. Crucially, most of that jump is not the OpenAI license. It comes from the Hatch acquisition, Yelp Host — the AI phone-answering service — and data licensing broadly. In other words, the AI transformation the market paid up for lives in a small line item, while the traditional ad engine that still carries the company shrinks. Net income fell 28% year over year to $32 million, full-year adjusted EBITDA guidance was pulled into a $315–325 million range, and Yelp paused its buybacks to pay down its credit line, planning to resume returning capital in 2027.

The market has noticed the tension. Near $21 a share and down roughly 30% year to date, the stock trades at about 4.8 times trailing EBITDA and under 10 times trailing earnings — cheap, but cheap partly because the core is declining and the AI upside is still a promise. Management has a target: a $250 million annual run rate in "other revenue" by the end of 2028. That is a roadmap, and a roadmap creates value only if the products and partners actually deliver it on schedule.

Who owns the local customer now

The strategic bet underneath all of this is a value-migration play. Discovery is moving from the search results page — where Yelp competed for clicks and traffic it mostly didn't control — into the AI chat window, where a user asks a question and gets an answer rather than a page. Yelp is trying to become the "actions" layer in that window: making the reservation, joining the waitlist, sending the quote. That is the part of the local economy people are willing to pay for, and it is smarter than renting out reviews for a flat licensing fee alone.

But the same structural problem follows it there. OpenAI controls the interface and the relationship with the end user. The deal is non-exclusive, so Yelp's single best asset is simultaneously offered to every competitor that wants it. If AI assistants become the front door for local discovery, Yelp risk being paid as a content supplier — rent for the data that fuels someone else's moat — rather than as the owner of the transaction.

This is what separates the two readings of the July pop. If Yelp converts ChatGPT's reach into billable quotes, reservations, and bookings at its own economics, then there is a real growth story hiding underneath a single-digit multiple. If the licensing fees are modest, the leads are unattributable and hard to price, and OpenAI captures the relationship, then the deal is a hedge that slows the decline without reversing it, and the market's skepticism is earned.

The company itself says the impact is too early to measure. That is the honest state of the evidence: real distribution, real reach, sound strategic logic — and no delivered revenue yet. The stock's move is a price outcome, and price outcomes are not operating results. The question isn't whether the ChatGPT deal is good news; it is whether Yelp gets paid for the discovery or just rents the data. The deal proves distribution. The economics are still a claim.