Oracle carried out a large-scale round of layoffs on Tuesday. Dozens of employees shared their experiences on LinkedIn, reporting that the layoffs were not performance-related. Estimates suggest that around 10,000 employees were affected. The tech company attributes the cost-cutting measures to its aggressive AI expansion, which is expected to come with a hefty price tag in the coming years.
On Tuesday, the “significant” reorganization, as Oracle managers described it on LinkedIn, took place, resulting in thousands of employees losing their jobs. Dozens of employees shared their experiences on LinkedIn, Reddit, and elsewhere online. Senior manager Michael Shepherd wrote that “senior engineers, architects, operations leaders, program managers, and technical specialists” were laid off. Shepherd himself was not affected, but emphasized that the “significant reduction in force” was not based on performance. “The individuals affected were not let go because of anything they did or didn’t do,” he wrote.
One employee estimates that about 10,000 people lost their jobs, based on a drop in the number of active users on Oracle’s Slack. Former Oracle employee Kendall Levin posted on LinkedIn that her position had been eliminated as part of a massive downsizing. Several others reported receiving an email early in the morning, often at 6 a.m., informing them that they were no longer employed, with one month’s severance pay.
In the United States, employment law is largely based on at-will employment, allowing companies to lay off large numbers of employees in a relatively short period of time. It is not known how the layoffs are distributed geographically.
AI as justification for layoffs
Oracle explicitly links the cost-cutting measures to AI tools. Co-CEO Mike Silicia stated earlier this month that AI coding tools enable Oracle to “deliver more complete solutions to customers faster with smaller engineering teams.” He is far from alone in delivering this type of rhetoric. Meta and Block CEO Jack Dorsey also attribute AI-driven efficiency to layoffs at their own companies. Meta previously denied rumors of an AI-driven round of layoffs affecting 20 percent of its workforce, but a smaller reduction has already taken place in the metaverse division.
In any case, layoffs reported to be AI-driven are often suspected of being a disguised push toward efficiency. By putting AI tools on a pedestal and believing claims these tools are capable enough to render thousands of employees redundant, investors view the reduction in a more positive light. For now, there are other reasons to believe Oracle is drastically reducing its own workforce.
Billions in AI Infrastructure
The layoffs coincide with massive AI investments by Oracle. The company plans to spend at least $50 billion on infrastructure this year and has raised an additional $50 billion in debt to meet the demand. As early as February, Techzine reported that Oracle was considering drastic measures to finance its AI spending, including a potential round of layoffs affecting 30,000 employees. Based on the limited information available about this round of layoffs, those numbers have not yet been reached. Nevertheless, it could involve a phased reduction, for example by allowing contracts to expire or enabling layoffs under less flexible labor laws than those in the U.S.
Oracle is also participating in the Stargate initiative, a $500 billion project to build data center capacity in the U.S. In this effort, it is collaborating with OpenAI, SoftBank, and MGX. OpenAI previously purchased $300 billion worth of computing power from Oracle for this project. Co-CEO Clayton Magouyrk said earlier this month that investments in AI infrastructure are simply capital-intensive.
The profitability Magouyrk predicts based on those investments is not yet materializing. The logic also seems fundamentally flawed. The expansion of AI data centers is typically limited by power supply, causing extremely expensive equipment to gather dust in a warehouse and thus generate no revenue. In a few years, that hardware will already be obsolete for the most ambitious AI workloads, requiring massive new investments. We estimate that this cycle cannot repeat itself indefinitely, but so far, it has served companies like Nvidia well.