Why OpenAI Paid 20 Times Revenue for an Interview Program
There is a number that does not add up at first glance and that, precisely for that reason, is worth examining carefully: more than 100 million dollars for a daily technology program that generates approximately 5 million dollars in annual revenue. That is a valuation multiple of more than 20x on sales for a media asset, in a sector where typical multiples rarely exceed 3x or 4x of revenue. This is not a calculation error. It is a strategic declaration.
OpenAI purchased TBPN in early 2026. Paramount Skydance acquired The Free Press — the newsletter and podcast business built around Bari Weiss — for around 150 million dollars at the end of 2025 and subsequently named Weiss director of CBS News. Joe Rogan renewed with Spotify for a reported amount of 250 million dollars. Alex Cooper moved her Spotify contract to SiriusXM for 125 million dollars, according to reports. Pat McAfee licensed his daily program to ESPN for 85 million dollars over five years. The Kelce brothers signed with Amazon's Wondery for 100 million dollars.
And now, according to Reuters, James Murdoch's investment firm Lupa Systems is reportedly in advanced talks to acquire New York Magazine and the Vox Media podcast network for 300 million dollars or more.
None of these transactions makes sense if analyzed through the conventional criteria for media valuation. All of them make perfect sense if one accepts the underlying thesis: in a market where artificial intelligence can produce generic content at near-zero marginal cost, the scarce asset is not the content itself, but the trust relationship between a recognizable voice and its audience.
---
A Multiple That Does Not Lie
Twenty-one times annual revenue for a technology conversation program. That number deserves to be held for a moment before moving on to analysis, because it compresses the entire argument.
If a company of the scale of OpenAI — with access to the best financial models in the world and the capacity to build any content format from scratch — decides to pay that multiple, there are two possible readings. The first: it made a poorly founded decision. The second: it is paying for something that its own models cannot replicate and that its internal metrics confirm has disproportionate strategic value.
The second reading is the most reasonable, and the reasons are concrete. TBPN is not worth 100 million dollars because of its current advertising revenue. It is worth what it is worth because it represents direct and recurring access to a high-quality technical audience that has chosen to listen to that voice on a regular basis. For OpenAI, which competes in a market where brand perception and technical credibility matter as much as model benchmarks, this is not a media expense: it is positioning infrastructure.
The pattern repeats itself across the other transactions. Spotify did not pay 250 million to Joe Rogan for the present value of his recorded episodes. It paid for the listeners who tune in every week because they want to hear Joe Rogan, not because Spotify recommended similar content to them. Amazon did not pay 100 million to the Kelce brothers for their archive rights. It paid for a community that already exists and that has consumption habits formed around those specific individuals.
The difference between buying a content library and buying a loyal audience is not semantic. It is structural. A library will depreciate in value as soon as the cultural context changes or as soon as AI can generate equivalent content. A loyal audience has real substitution costs for the listener, who has invested time and attention in building a relationship with that voice.
---
The Four Models and the Murdoch Bet
The Fortune article describes with precision four business architectures that different organizations are using to capture the value generated by that audience loyalty.
The first is strategic talent incorporation: bringing the voice inside the organization and protecting what made it work. OpenAI with TBPN and Paramount Skydance with The Free Press follow this pattern. The bet is that a trusted voice can reposition the buyer's brand without corrupting what made that voice trustworthy in the first place. The risk is obvious: if the audience perceives that the voice is becoming institutionalized, the trust that justified the purchase price erodes.
The second is infrastructure as a service: offering production, distribution, and sales to creators who retain editorial independence. Red Seat Ventures built this model for commentators such as Tucker Carlson, Megyn Kelly, and Bill O'Reilly. Following its acquisition in 2025 by the Tubi Media group of Fox, and with the subsequent incorporation of Backtracks for advertising and Supercast for subscriptions, Red Seat offers a complete monetization chain. Fox gains brand building; the talent gains access to distribution channels.
The third is the institutional bundle: integrating creator-led podcasts into a subscription product. The New York Times did this early on with voices such as Andrew Ross Sorkin, Ezra Klein, and Michael Barbaro, and moved its podcast catalog behind a paywall in 2024. Netflix is testing the same logic. As Ted Sarandos noted on a 2025 earnings call: "the lines between podcasts and talk shows are becoming quite blurry."
The fourth model is the one emerging with the greatest force right now, and it is the most interesting from a business economics perspective: building or acquiring journalistic and creator brands as editorial scaffolding for high-value in-person experiences. Revenue comes primarily from sponsorships, ticket sales, brand partnerships, and access to a curated audience that can be converted into subscribers. Journalism and podcasts are the inputs. In-person events are the high-margin output that sustains the model.
Jay Penske has been operating a version of this for years, where Variety, The Hollywood Reporter, Rolling Stone, Dick Clark Productions, and a stake in SXSW are combined into a vertically integrated culture and events business. The Atlantic under Laurene Powell Jobs has developed its AtlanticLIVE business alongside its journalism.
James Murdoch has publicly stated that "live events are the core business" of Lupa, and that his investment thesis centers on "the capacity to bring communities together around strong brands." Tribeca and Art Basel already deliver on that. With New York Magazine and the Vox Media podcast network, Lupa would be positioned to connect with audiences through in-person experiences across media, technology, culture, and art. The Futurific Institute — a large-scale global festival of ideas backed by Kathryn and James Murdoch, planned for 2026 — suggests where that portfolio is heading.
The financial logic of the convening model is sound because it resolves the scalability problem that purely digital models face: in-person events have limited capacity, and that scarcity creates price. An audience that pays for physical access to the community it follows digitally has much higher abandonment costs than a monthly subscription that can be cancelled with a single click.
---
The Durability of the Asset and What Buyers Are Assuming Without Fully Validating
Every investment thesis has its weak point, and this model is no exception. The question that every buyer in this race should have answered with data before signing: how much of audience loyalty is tied to the individual, and how much survives corporate integration?
The history of media acquisitions is full of cases where the value paid evaporated precisely because the voice that attracted the audience lost its distinctive character once it became subordinated to the institutional incentives of the buyer. When the public perceives that the person has changed — that they speak differently, that they avoid certain topics, that the content seems produced to satisfy a corporate owner — the relationship breaks down faster than financial models anticipate.
That does not mean the model is poorly conceived. It means that its durability depends on a variable that is not easy to control from a spreadsheet: the buyer's capacity to preserve the conditions that made that voice authentic in the first place. Red Seat Ventures and the infrastructure model resolve this structurally, by not incorporating the talent but rather serving it. The direct acquisition models — OpenAI with TBPN, Paramount with The Free Press — assume an integration risk that has no clear precedent in comparable transactions.
The other risk that the Fortune article names with precision is audience volatility. Talent can lose credibility, fall out of sync with the moment, or simply become less compelling. When the audience relationship weakens, the asset can erode rapidly. There is no permanent moat in a media personality: permanent moats exist in physical infrastructure — theme parks, stadiums, festivals — which is exactly where Murdoch's convening model points.
What makes the Lupa portfolio comparatively more robust is that it combines the digital audience asset with the physical experience asset. If New York Magazine loses cultural influence, the festival of ideas built on that editorial foundation can continue to bring communities together. If a podcast loses listeners, the in-person event it generated does not automatically disappear. There is structural redundancy in the model that does not exist in a pure talent acquisition.
---
AI Amplifies Exactly the Problem These Buyers Are Trying to Solve
The central argument of this investment thesis becomes stronger, not weaker, as the generative capacity of artificial intelligence advances. AI increases the supply of generic content. A text about the latest trends in technology, a synthesis of financial news, a summary of recent research: all of that can be produced at scale and at near-zero marginal cost. What AI cannot produce is the accumulated personal history of a voice that has spoken to its audience for years, that has made mistakes in public, that has changed its mind in a recognizable way, that has a style its listeners can identify before the first sentence is finished.
That distinction is not philosophical. It has direct operational consequences for monetization models. Advertisers are already measuring the difference between attention paid to generic content and attention paid to a trusted voice. Conversion rates in established creator podcasts are systematically higher than in generic audio formats, not because the format is different, but because the audience has a pre-activated disposition of trust toward the person speaking.
For corporate buyers, that translates into something measurable: lower customer acquisition cost, higher subscription retention, and a higher sustainable price for event tickets. The 20x multiple that OpenAI paid for TBPN must have been justified internally with a projection of how much that conversion differential is worth over several years, applied to the products that OpenAI needs to position in front of a sophisticated technical audience.
If the projection is correct, the purchase will have been cheap. If the integration destroys the authenticity of the voice, it will have been expensive. There is no way to know yet. What we do know is that the decision to buy was made, that the multiple was paid, and that at least half a dozen similar transactions are being executed in parallel with the same underlying hypothesis.
That convergence of sophisticated bettors moving in the same direction does not prove that the thesis is correct. It does prove that there is sufficient accumulated evidence about the value of a loyal audience for organizations with access to the best available financial analysis to be willing to pay extraordinary premiums for it. Humanity — measured as trust, habit, and community — is becoming a line item in the asset sheets of platform companies. And those who arrive late to acquire it will pay even higher multiples.
