OpenAI Runs an $852B Company Like a Group Chat

I ship products on OpenAI's API. Have for over a year. I don't have strong opinions about Sam Altman as a person, I don't follow the palace intrigue for entertainment, and org chart reshuffles normally don't make me open a text editor. But when your most critical vendor rearranges four senior roles in a single memo — three days after closing the largest private funding round in history — you start taking notes.

So here are mine.

What Happened

On April 3, Fidji Simo announced she's taking medical leave. Simo is OpenAI's CEO of Applications — which means she runs ChatGPT, the API, and basically everything that touches a paying customer. She has POTS, diagnosed in 2019, and it relapsed. The leave is expected to last several weeks. I genuinely wish her a full recovery. This piece isn't about her health.

It's about the chain reaction.

Within hours of Simo's announcement, OpenAI published what amounts to a full leadership reshuffle. Greg Brockman, co-founder and President, takes over product. Brad Lightcap, the COO, moves to something called "Special Projects" — which involves complex deals, investments, and a PE joint venture called DeployCo. I've been in tech long enough to know that "Special Projects" means either something genuinely important or a soft landing before an exit. We'll see which one.

Denise Dresser, the CRO — she was CEO of Slack before this, joined OpenAI in December 2025 — absorbs most of Lightcap's operational work. Jason Kwon, Chief Strategy Officer, picks up government affairs and "OpenAI for Countries." And Kate Rouch, the CMO, steps down entirely to focus on cancer recovery. Late-stage breast cancer, diagnosed about a year and a half ago. Gary Briggs, formerly Meta's CMO, comes in as interim.

That's four senior roles reshuffled, vacated, or redefined. In one day. At a company that closed a $122 billion funding round at an $852 billion valuation on March 31. Three days earlier.

I should put those numbers in context, because they're easy to skim past. The round was co-led by SoftBank, Andreessen Horowitz, and D.E. Shaw Ventures. Amazon put in $50 billion. Nvidia and SoftBank each put in $30 billion. OpenAI says it's generating $2 billion per month in revenue with over 900 million weekly active users. They're planning an IPO for Q4 2026 — though their own CFO, Sarah Friar, has reportedly told Altman the timeline is too aggressive.

And they've done seven acquisitions this year. Seven. In four months. Including — and I've reread this three times to make sure — a daily tech talk show called TBPN, for what the FT described as a "low hundreds of millions."

Why This Matters If You Build on Their Stack

Here's the thing that I don't see anyone saying plainly: if you're building on OpenAI, the application layer is your entire relationship with the company. You don't interact with the research team. You don't care about the board composition. You care about the API, the pricing, the rate limits, the deprecation schedule. Simo owned all of that.

Brockman is a good engineer. I have no doubt about that. But he's been away from day-to-day product work for a while, and stepping into this mid-sprint — with an IPO filing coming, enterprise customers who expect stability, and a product surface that now covers chat, API, voice, agents, coding tools, image generation, and whatever TBPN is supposed to be — that's not a lateral move. That's a different job entirely.

What actually concerns me isn't that one person left. It's the pattern. I started counting the leadership changes at OpenAI since 2023 and stopped because the list got depressing. The board crisis. Murati's departure. The nonprofit-to-profit conversion drama. The Altman consolidation. And now this. Every single time, the company's response is the same: everything's fine, look at the growth.

Growth is not a management strategy.

What Nobody's Saying Out Loud

Let me be clear: OpenAI is not collapsing. The products work. Revenue is real. The user base is massive. I'm not being dramatic here.

But there's a failure mode that hits companies at exactly this stage, and I've seen it enough times to recognize the shape. The infrastructure of decision-making doesn't scale with the infrastructure of compute. You can raise $122 billion and still not have a coherent answer to "who decides what the API does next quarter."

The TBPN thing is revealing. On April 2, OpenAI bought a daily live tech show hosted by Jordi Hays and John Coogan. It sits in OpenAI's Strategy org, reporting to Chris Lehane. Ben Thompson wrote on Stratechery that the purchase "makes no sense." CNN ran the headline "OpenAI isn't just buying a podcast — it's buying influence" and drew a comparison to RCA creating NBC back in 1926. Jessica Lessin at The Information was more direct: Musk has X, now Altman has TBPN.

I think they're all basically right. A company about to go through the most scrutinized IPO in tech history decided that buying a talk show was a good use of its time. That tells you something about what leadership is actually optimizing for — and it's not the API.

Altman is building a conglomerate. Hardware partnerships. A media property. Developer tools. Enterprise contracts. Consumer subscriptions. The acquisition list this year alone: Convogo, Torch Health, Crixet, an OpenClaw acqui-hire, Promptfoo, Astral, TBPN. Plus the $6.4 billion io deal with Jony Ive. And the person who was supposed to hold the product layer together just went on medical leave with no permanent replacement named.

Look — I run a Laravel stack. Nothing fancy. And I wouldn't push a major feature without knowing who owns the rollout plan. That's just basic. OpenAI is running an $852 billion company where the answer to "who's in charge of the product" changed overnight and could change again next month.

The Question Every CTO Should Be Asking

If you're evaluating a multi-year OpenAI enterprise deal right now, here's what your risk slide looks like: the CEO of Applications is on indefinite leave. The COO moved to a role nobody can quite explain. The CMO is gone. The President is doing product now. The CFO is reportedly telling the CEO the IPO timeline is too aggressive. And the CEO himself is running fundraising, acquisitions, hardware partnerships, a media company, and — presumably — sleeping sometimes.

Enterprise buyers don't care about your valuation. They care about whether the person they shook hands with will still be around when something breaks at 2 AM.

I've seen this pattern at smaller companies. Not at this scale, obviously — nobody's ever seen it at this scale. But the shape is the same: a company moving so fast that the org chart becomes a fiction you maintain for the investor deck, while actual decisions flow through whoever happens to be in the room with the founder.

That works when you're 50 people. It stops working well before you hit 900 million weekly active users. And it definitely stops working when you're about to become a public company with quarterly reporting and a board that has real fiduciary duties.

OpenAI will probably pull this off. They have enough money, enough talent, and enough market position to absorb a remarkable amount of chaos. But "probably" is doing a lot of heavy lifting in that sentence. And if you're building your business on their platform, "probably" isn't an architecture decision.

It's a prayer.