OpenAI just bought a voice it hasn't earned

Hey, I’m Isaac 👋 I founded Pistachio, a growth agency working with B2B brands like Atono and Clay to build trust, relationships and loyalty with their current and future customers.

If you want to chat about your own growth strategy, book a free call with me here.

Interested in sponsoring these emails? See our partnership options here.

Last week, OpenAI acquired TBPN, a daily tech talk show hosted by founders John Coogan and Jordi Hays. The show streams on YouTube and X for three hours a day, and counts Mark Zuckerberg and Satya Nadella among its regular guests. It's on track to do $30 million in revenue this year.

The acquisition price was somewhere in the “low hundreds of millions” and they’ve already completely shutdown TBPN's advertising business.

In their official announcement, OpenAI said "One thing that's become clear is that the standard communications playbook just doesn't apply to us". They pledged to protect editorial independence. TBPN will continue to set its own agenda, choose its own guests, make its own editorial decisions.

That promise is the bit I find most interesting about this whole story.

Before we get into the problem, the acquisition logic here is sound and OpenAI are far from the first to see it.

HubSpot bought The Hustle in 2021. Not for the content library itself, but for the 1.5 million subscribers who trusted it enough to open it every morning. They bought instant distribution. Credibility already assembled. Audience already warmed up. When you have product scale but distribution problems and no trust at scale, you go shopping for someone else's credibility.

a16z built out a full media operation to shape the narrative around venture capital. Shopify produces editorial content like a publisher. These companies have understood and acted on an insight traditional marketing still struggles with. The most valuable asset a brand can own is an audience that trusts you before you've asked them for anything.

TBPN fits this model almost perfectly. The show became Silicon Valley's go-to venue for candid conversation specifically because it wasn't owned by anyone with a stake in the outcome. That editorial independence attracted the guests. That trust built the audience. That audience is what made it worth hundreds of millions.

So OpenAI acquiring them makes sense. In theory.

Here's where it falls apart.

We've already seen how this plays out. When Jeff Bezos bought The Washington Post in 2013 for $250 million, he made the same promise. Editorial independence. No interference. For over a decade, things were stable enough. Then in 2024, he personally blocked the paper from endorsing a presidential candidate in the lead-up to the election. Around 200,000 loyal subscribers cancelled. Senior staff resigned. The editorial independence promise survived right up until the moment it was truly tested.

So OpenAI's editorial independence pledge has been met with scepticism, largely because the people making the promise have a specific track record of commitments that don't hold when they become inconvenient.

Over the past year, the share of American adults who dislike Sam Altman has more than doubled.

That's active erosion of trust in the brand, both personal and business. The reasons aren't hard to trace.

Earlier this year, ChatGPT introduced conversation-specific ads to its free tier. Anthropic, makers of Claude, responded at the Super Bowl by running ads with the tagline "There is a time and place for ads, and AI chats aren't it". Sam Altman called them "dishonest" and "deceptive", arguing OpenAI would "obviously never" run ads the way Anthropic depicted.

Then there's the Pentagon. As an existing DoD vendor, Anthropic was asked to remove safety guardrails for use in "autonomous weapons targeting" and "domestic surveillance". They refused. OpenAI stepped in and took the contract.

TBPN's value is built on independence. Zuckerberg, who notoriously doesn’t do many media interviews at all, sat down for three candid hours because the show provided a fair space for genuine conversation. The moment it becomes an OpenAI property, every guest booking, every topic selection, every question asked or avoided gets filtered through that lens by audiences, regardless of what the hosts actually do.

As Digiday’s article noted, "the line between where the marketing ends and entertainment begins isn't yours to draw — it's the audience's".

OpenAI have proved over the past few years that their commitments carry a price tag. When keeping a promise costs them nothing, they keep it. When it costs them something real, they find a way around it to achieve the better commercial outcome. That's the pattern that's making the editorial independence pledge so hard to land, and it's why this deal, structurally and strategically smart as it is, has a huge problem baked in from day one.

The contrast with Anthropic is something I've written about in detail before.

Both companies launched brand campaigns around the same time last year. ChatGPT went with normalisation, showing ordinary people achieving small everyday wins through AI, running the Google playbook of volume and ubiquity. Claude went the other way with identity, the "Keep Thinking" platform, the anti-slop positioning, the thinking caps that 5,000 people queued around the block for at a single New York pop-up. Two completely different bets about what builds a durable brand in a maturing category.

But talk is cheap. Any company can commission a campaign and write a mission statement. What matters is how those companies act when those values are tested.

Anthropic turned down the Pentagon contract, which cost them. Contracts across Treasury, State, HHS, and FHFA were terminated. The DoD designated them a "supply chain risk", a label usually reserved for hostile foreign suppliers. Then Claude shot to number one on the App Store, overtaking ChatGPT for the first time. Free signups surged 88% in a single day. The audience Anthropic had spent two years building responded exactly the way a loyal audience does when their brand holds under genuine pressure.

That's the difference between a brand and a brand campaign.

Menlo Ventures data shows Anthropic has now overtaken OpenAI in enterprise LLM usage, at 32% against 25%, a dramatic reversal from OpenAI's dominant 50% position just two years ago. In code generation, Claude holds 42% against OpenAI's 21%. That kind of turnaround can’t be achieved with just a few clever ads. It’s only pulled off through trust at scale.

This is exactly why the editorial independence promise would land differently coming from a different company. If Anthropic, or Stripe, or PostHog, or any brand that has visibly held its principles under commercial pressure, acquired a media property and pledged “editorial independence”, the scepticism would be more muted. Same words. Very different track record.

Brand strength isn't about what you say when it's cheap to say it. It's about what you do when holding your position costs you something. OpenAI have answered that question consistently over the past two years, and those answers are making this acquisition harder to execute than it should be.

The brands winning right now didn't go out and buy an audience. They built one by consistently doing what they said they’d do, and talking about it in a way that attracted the right people. The media presence followed the trust. Not the other way around.

If you're early-stage, the time to start building an owned audience is before you need it. Not because you're planning to sell it, but because earned trust compounds. The brands investing in it now are creating something that gets more valuable over time. Consistency before scale.

If you're more established, you should invest in owned media. But only if your brand has actually earned the trust that makes owned media work. A podcast built on a shaky brand doesn't fix the brand. It amplifies the shaky-ness.

And if anyone from OpenAI is reading, know that I’m still hopeful. The editorial independence promise might still hold. Coogan and Hays might genuinely be given the freedom to run the show exactly as they always have. But this is harder than it needed to be. Every pointed question about OpenAI, every guest who pauses before accepting an invite, every topic the show covers or quietly avoids will be scrutinised through the lens of ownership.

You can buy distribution. You can't buy the trust that makes distribution valuable in the first place.

If you enjoyed this post or know someone who may find it useful, please share it with them and encourage them to subscribe: brandchemistry.co/p/openai-acquires-tbpn