Michael Saylor, executive chairman of Strategy, said the company’s roughly $15 billion in fundraising over the past year was tied to a new set of Bitcoin-backed preferred stock products that he says were developed with help from ChatGPT.
Speaking on the Aug. 6 episode of The Diary Of A CEO, Saylor told host Steven Bartlett, “I made $15 billion with AI last year.” The comment quickly drew attention across crypto circles. He later clarified that the figure did not refer to personal profit or accounting earnings at the company. It referred to capital raised through a series of preferred stock offerings, with most of that money used to buy more Bitcoin.
How Saylor described the $15 billion figure
According to Saylor’s breakdown, one preferred stock product alone raised about $10.5 billion. That total included roughly $2.5 billion from its initial public offering and another $8 billion from follow-on shelf issuance. Other related products contributed about $4 billion combined, bringing the total to around $15 billion.
Forbes currently estimates Saylor’s net worth at about $4 billion. The report says he turned MicroStrategy, once an enterprise software company, into a Bitcoin accumulation vehicle in 2020. The company now operates as Strategy.
Over the past six years, Strategy has bought nearly 850,000 BTC. Its market capitalization stands a little above $50 billion, and the value of its Bitcoin holdings at one point approached $55 billion. The report says most of those purchases were not funded by operating profits. They were financed through repeated sales of debt, common stock and preferred stock.
Why Strategy looked beyond converts and common shares
By early 2025, the report says, Strategy had become one of the world’s largest issuers of convertible debt. That route was nearing its limit, while additional common equity issuance would have brought heavier dilution. With about $30 billion worth of Bitcoin on hand at the time and plans to keep adding to its position, the company needed a new funding structure.
Saylor said he did not start with another brainstorming session led by investment banks. Instead, he opened ChatGPT and asked whether it was possible to design a security that sat somewhere between common equity and traditional debt. He later recalled that the AI said it could be done and outlined a path.
The structure that emerged was a variable-dividend preferred stock, which the report describes as something that had not previously been put into practice in this form. The key idea was to adjust the dividend rate monthly, later shifted to semi-monthly, with the aim of keeping the preferred share price close to its $100 par value over time. In the report’s framing, that gave the instrument a different profile from common stock, which can move sharply, and from traditional bonds, which come with maturity repayment pressure, while still feeding fresh capital into the company for Bitcoin purchases.
Saylor said bankers and lawyers initially pushed back, arguing that no one had done this before. His response, as quoted in the report, was: “Is it illegal? No. Why has no one done it? Because no one had a reason to do it before.” The report also says AI helped the team think through language for handling regulatory and legal objections. Those products were launched during the first half of 2025.
STRK, STRC and STRF
The report identifies three core preferred stock products.
- STRK (Perpetual Strike Preferred Stock): pays an 8% fixed dividend, can pay in cash or common stock, is convertible into common shares, and offers some upside participation.
- STRC (Perpetual Stretch Preferred Stock): carries a variable dividend, currently adjusted to a 12% annualized rate, pays semi-monthly, and is designed to keep trading as close as possible to its $100 par value while offering a steadier income profile.
- STRF (Perpetual Strife Preferred Stock): is not convertible, sits higher in the capital structure, and is aimed more at conservative income-focused investors.
STRC is described as both the most controversial and the largest by fundraising size. The report says it fell from around par value to about $75 by the end of June, then rebounded in recent weeks and moved back toward the $100 target. Saylor’s team has openly framed a return to the $99-$100 range as one of its most important current operating goals.
Common stock fell while the company sold some Bitcoin and built cash reserves
At the same time that preferred-share fundraising accelerated, Strategy’s common stock came under pressure alongside Bitcoin. The report says the company’s common equity market value fell about 80% over the past year.
Strategy also started selling Bitcoin. In its latest public disclosure, the company said it sold 1,638 BTC in one week for about $104.7 million. Part of the proceeds went to pay preferred dividends, and part was used to repurchase its own preferred shares.
Saylor offered two reasons for those sales. One was to “vaccinate the market,” in his words, so investors would get used to the idea that Strategy may sell Bitcoin. The other was straightforward funding for debt and dividend obligations.
The report adds that, besides Bitcoin valued at about $55 billion, Strategy currently holds around $4 billion in cash reserves. Management said that amount is enough to cover roughly two years of dividend and interest payments without the need for frequent refinancing.
Reaction split across investors and executives
Reaction was divided after the remarks circulated. Jason Calacanis, a technology investor and host of the All In podcast, wrote on X: “Is this Saylor’s terrifying confession, or a brilliant use of AI?”
Abra CEO Bill Barhydt took the opposite view. He replied on X: “History will show this is absolutely brilliant financial engineering at the end of the U.S. debt cycle. If you can borrow dollars and put them into an asset with fixed supply and very fast adoption, you should do it. Saylor is legally barred from investing this capital into securities (because he is not a company under the Investment Company Act of 1940), so that leaves Bitcoin and gold. He will make some mistakes, but the strategy itself is smart.”
Saylor framed the episode as a broader lesson in how to use AI. “Don’t learn to do the things AI already knows how to do,” he said. “What you need to learn is how to get AI to do something nobody has ever done before. If you want an earthshaking success, you have to find that magical opportunity.”
Not the first time Saylor has credited ChatGPT
This was not the first time Saylor publicly credited ChatGPT. The report says that in May 2025, at the company’s developer conference, he said he had used ChatGPT’s deep research mode to design the company’s convertible preferred stock products.
From turning the company into a Bitcoin treasury vehicle in 2020 to using AI in 2025 to rethink funding structures, Saylor has again drawn attention to one question: whether AI can help companies build financial products that had not previously been brought to market in this form. The report does not say where that $15 billion effort will ultimately lead Strategy or Bitcoin.