Michael Saylor says ChatGPT helped him raise $15 billion to buy more Bitcoin

Michael Saylor did not come on The Diary of a CEO to tell you ChatGPT made him 12 percent better at inbox zero. He says ChatGPT helped him raise about $15 billion so his company could buy more Bitcoin, which is either the best ad OpenAI never paid for or the most Saylor sentence ever assembled in a lab.

The claim arrives with the usual Saylor voltage, part MIT lecture, part CNBC fever dream, part wizard explaining why your broom has no yield. But underneath the flex is a real and specific argument about artificial intelligence: the people getting paid will not be the ones asking AI to summarize PDFs. They will be the ones using it to design weird new things that conservative institutions reject because nobody has blessed them yet.

لقد استخدمت الذكاء الاصطناعي لأربح 15 مليار دولار العام الماضي.

That sentence is doing some peacocking. Saylor did not say ChatGPT went out and raised the money, charmed investors, filed the paperwork, and rang the bell while wearing a Patagonia vest. What he describes is stranger and more useful: his company had already exhausted the obvious routes for raising capital to buy more Bitcoin. So he went looking for a new financial instrument.

ChatGPT as investment banker, apparently

Saylor says the problem was simple in the way only a multibillion-dollar Bitcoin balance sheet can be simple. His company had issued common equity. It had become, by his telling, the largest issuer of convertible bonds in the world. The machine needed more fuel, and the old pipes were maxed out.

The answer, he says, was preferred stock, a flexible security that can be made to resemble debt, equity, or some uncanny corporate-finance centaur in between. Specifically, he points to STRK, a convertible preferred stock instrument, and then to another short-term credit-style product designed to trade close to par while paying a distribution rate the company could adjust.

This is where the Saylor story gets its bite. Lawyers and bankers, in his telling, did what lawyers and bankers are paid to do: they looked at the new thing and said, essentially, nobody does that. Saylor treated that not as a stop sign, but as a prompt.

ذهبنا إلى الذكاء الاصطناعي. قلنا:” حسناً، هل يمكننا القيام بذلك؟ ” أجابوا:” بالطبع يمكنك القيام بذلك. فقط افعل هذا، وهذا، وهذا، وهذا ، وذاك ”.

شات جي بي تي (Chat GPT). أوبن إيه آي (OpenAI).

There is a self-serving gloss here, obviously. Saylor has every incentive to frame his Bitcoin financing strategy as visionary rather than maximalist borrowing with better branding. He also has every incentive to make ChatGPT sound less like a chatbot and more like a junior Goldman Sachs team that never sleeps, never bills, and never asks whether the whole thing is too spicy for compliance.

Still, the distinction matters. The exaggerated version is: ChatGPT made him $15 billion. The credible version is: Saylor used ChatGPT to explore a legal and financial design space faster than his human advisers were willing to. That is not as memeable. It is probably closer to the truth.

بشكل أساسي، بعنا 15 مليار دولار من الائتمان، وهو ما يعادل تقريباً تحقيق الشركة حوالي 15 مليار دولار.

The AI lesson is harsher than the headline

Steven Bartlett tries to bring this back to normal human scale, the person listening who wants a business idea and maybe does not currently control billions in Bitcoin. Saylor’s answer is not a cozy productivity tip. It is closer to career triage.

If you’re building something, Saylor says, the point is not to learn tasks AI can already perform. It is to learn how to ask for the thing civilization has not already solved. That sounds grandiose, because Saylor rarely uses a candle when a flamethrower is nearby, but the practical advice is crisp: don’t compete with AI at clerical work. Use it to search the boundary of what is newly possible.

This is why his college advice gets so brutal. He says he would not study surgery, law, accounting, or driving if he were starting now. He is not predicting that every surgeon is toast by Thursday. He is saying that high-status paths are dangerous if their core skill can be decomposed, modeled, and automated.

إذا عدت إلى المدرسة، فربما لن أرغب في دراسة 95%مما تعلمته.

The funny part is that Saylor’s own example is not exactly democratizing. A baker in Lagos, a firefighter in Los Angeles, and a public company with a giant Bitcoin treasury do not all get the same magic wand. AI may help anyone ask better questions. It helps a lot more when the answer can be converted into securities filings, investor demand, and another mountain of Bitcoin.

So yes, the claim is inflated. It is also the rare inflated podcast claim with an actual mechanism inside it. Saylor is not saying ChatGPT gave him a listicle called 10 Side Hustles for 2025. He is saying it helped him find a financing structure human gatekeepers treated as too unfamiliar to bless. If that version is right, the arbitrage is not ChatGPT. The arbitrage is asking it the question everyone else is still waiting to ask a committee.