After a big event in the market, there’s a certain kind of silence—not panic, but something more unsettling. That’s about how it felt to see Oracle’s stock expire after having its worst week in 25 years, falling 19% over five days straight and reaching levels that made analysts look for historical comparisons. Oracle’s share price dropped this much in a single week last time in August 2001, when the dot-com era was still coming apart. Just thinking about that one detail is worth some time.
Larry Ellison is one of the most famous names in enterprise technology and a co-founder of Oracle. He has seen his personal wealth drop from a high point of about $388 billion in September 2025 to about $175 billion by mid-July 2026. Over $200 billion has been lost in less than ten months. Larry Page and Sergey Brin, co-founders of Google, Jeff Bezos, and Michael Dell have all made more money than him. Ellison is still very rich—worth more than $200 billion—but it’s impossible to ignore the way things are going.
Oracle has put a huge amount of money into building infrastructure for artificial intelligence, especially through its commitments to OpenAI. In fiscal 2026, capital spending went through the roof, reaching almost $56 billion, far exceeding the company’s own $50 billion goal. This meant that free cash flow went into the red by almost $24 billion. Oracle has about $130 billion in debt right now, and they want to get another $40 billion through debt and stock sales in the next fiscal year. As a direct competitor to Amazon, Microsoft, and Google, the company is rushing to open data centers in Michigan, New Mexico, and Texas. These companies can sell a much wider range of technologies than Oracle.
It’s possible that these huge commitments made sense when Oracle’s market value was close to $900 billion in the fall. At that time, interest in AI was high, and Oracle had a huge backlog of contracted future revenue—unfulfilled performance obligations rose 363% year over year and reached $638 billion. This week, Evercore analysts said that the demand signals are still strong. But demand and cash flow are not the same thing. Oracle is spending a lot of money at a rate that would have been unthinkable five years ago because those contracts won’t turn into cash for years.
The most recent Oracle earnings call didn’t have Ellison on it. Instead, co-CEOs Clay Magouyrk and Mike Sicilia, along with new finance chief Hilary Maxson, took questions. It’s possible that Magouyrk’s casual comment during the call, “Hilary has a tough life,” was either a sign of friendship or a quiet admission that the company’s finances were getting worse. Maxson, for her part, kept a calm attitude and promised to follow strict rules when allocating capital and to protect Oracle’s investment-grade credit rating.
To make things even more complicated, something is happening that has nothing to do with Oracle’s quarterly results. The Ellison family has promised $45.7 billion in equity financing for a big media deal. Larry Ellison is personally responsible for $40.4 billion of that amount, which is mostly backed by Oracle shares that are worth about half of what they were when he made the promise. Late in July, a temporary restraining order was issued that stopped work on the deal. A court hearing in early August could make things even slower. Any way you look at it, the timing is not good.
Wall Street analysts still believe in Oracle’s long-term thesis—71% of analysts who cover the stock still say you should buy it, which is the highest percentage of bullish analysts in 15 years. That number means something. But in the short term, things look messier. There is rising debt, a drop in headcount (Oracle’s workforce dropped 13% to 141,000 in fiscal 2026), and a 16% drop in the software sector as a whole because investors are worried that AI models will make people less interested in traditional software products. Oracle has dropped even more. Before this week’s losses, it was down 24% for the year.
It’s still not clear if Oracle’s “infrastructure-first” strategy will look great in three years or terrible. Businesses have been worse and come back better. But seeing a $45 billion change in wealth in just one week reminds me that even the most experienced tech builders can lose money when they move too quickly, especially when they have debt and not enough cash on hand yet.