Oracle Layoffs Hit Thousands as the Database Giant Bets Bigger on AI Than Anyone Expected

As is customary in the tech industry, the notifications began to go out on a Tuesday morning. An invitation to a meeting that nobody had scheduled the day before. HR sent a Slack message with a well crafted subject line that allows the recipient three seconds to brace. Internal estimates were circulating by midday that the cut at Oracle would be in the thousands, probably higher, and possibly spread across several regions. In the typical corporate manner that affirms everything by being silent, the company itself refuses to comment. By Tuesday night, Business Insider had already covered the wider story, and CNBC had confirmed the layoffs with two sources. The announcement did not come as a huge surprise to Oracle’s stock, which was already down significantly for the year.

More important than the headline statistic is the context. With about 162,000 employees as of May 2025, Oracle has been implementing one of the most aggressive AI infrastructure pushes in corporate software for the last eighteen months. Five years ago, when Larry Ellison was still building Oracle’s future around traditional database licensing, the business could not have imagined the kind of capital commitment that was announced in January 2026—a $50 billion debt and equity raise.

AI workloads, such as GPU clusters, data center growth, and the kind of capital-intensive infrastructure that hyperscalers have been developing for ten years, are currently the wager. Oracle is essentially attempting to leapfrog rather than catch up because it is smaller than AWS, Azure, and Google Cloud. That is costly. Additionally, it dilutes the share price, which has been rising throughout the year.

What really makes the entire concept intriguing is the OpenAI contract, which was revealed in September 2025. Following the transaction, Oracle’s residual performance obligations—the contracted revenue that hasn’t yet been recognized—rose 359% to $455 billion and have already surpassed $553 billion. The OpenAI component alone is valued at around $300 billion.

If that figure is taken at face value, it implies that Oracle has essentially taken on the role of back-end infrastructure provider for one of the most significant AI businesses globally. Depending on unresolved issues, the contract may or may not produce on that scale and within the timeframes investors anticipate. Looking at the figures, it seems like Oracle has positioned itself for either a generational disappointment or a generational rebound, and the market isn’t quite sure which.

A complex layer is added by the change in leadership. In late 2025, Mike Sicilia and Clay Magouyrk took over as co-CEOs of Oracle, replacing Safra Catz, who had led the company for years with the kind of exact operational discipline that won her quiet respect throughout the industry. In technology, co-CEO arrangements seldom endure.

When decisive strategic choices are required, the structure often causes conflict. The more prominent voice on AI strategy has been Magouyrk, who rose through Oracle’s cloud infrastructure division. Sicilia tends to focus more on industry-specific goods and applications. One of the unanswered topics that is discussed softly in Austin and not at all during earnings calls is whether the partnership will survive what is expected to be a challenging fiscal year.

According to a January TD Cowen research, eliminating 20,000–30,000 workers might result in an additional $8–$10 billion in cash flow. Whether the analysts want it to or not, that type of math that is recorded in study notes frequently finds its way into boardroom discussions. Although the current round of layoffs, which is in the thousands, is significantly below that level, it is also obviously a first step rather than a final one.

Speaking with those close to the company, it seems likely that the workforce downsizing will continue in waves until 2026, with each round focusing on certain business areas that don’t align with the AI-forward positioning the current leadership team is attempting to establish.

There’s a certain silence when you stroll through Oracle’s Austin offices during a layoff week. There are fewer people in the cafeteria. The little uneasiness of meetings that no one really wants to be in permeates the hallways.

Engineers are checking their phones more frequently than normal, hoping to avoid receiving the meeting invitation. While hiring for AI infrastructure continues at a rapid pace, layoffs are occurring in a number of departments, including sales, marketing, and certain legacy product teams. Given Oracle’s debt load and competitive pressure, it is the modern tech restructuring strategy, carried out with a specific intensity.

The AI era has hastened the loss of older software workforces, which is difficult to ignore. Google cut, Salesforce cut, Microsoft cut. Due in part to its later start and increased margin pressure, Oracle is currently making more drastic cuts than its competitors. As Magouyrk pointed out during the most recent earnings call, there is still more demand than supply for AI infrastructure.

The truly unanswered question is whether such demand supports the size of Oracle’s wager and the expense of getting there. There are still actual performance duties. There is also a serious risk of execution. As this develops, there’s a sense that Oracle’s position in the AI economy and whether it becomes a true fourth hyperscaler or a warning against debt-financed transformation will be decided over the course of the next two years.