Oracle is the latest American tech company to make massive layoffs because of AI. However, the goal is not to replace employees with machines, but to save money to offset the gargantuan expenses in data centers.
One more day, a new announcement of massive layoffs by an American tech company. Yesterday, it was the cloud and software giant Oracle which announced the elimination of “several thousand jobs”, citing AI as a motivation. But while the company said it was eliminating certain positions that it expected to become redundant with the rise of this technology, its decision is also and above all motivated by the desire to reduce its expenses to compensate for its massive investments in AI data centers.
Oracle’s crazy spending
Once considered one of the gray beards of American tech, Oracle has succeeded in recent years in a spectacular repositioning around AI which has allowed it to come back in force, notably by focusing on small data centers scattered everywhere, as well as on inference rather than training. In particular, it signed a $300 billion partnership with OpenAI, to build the IT infrastructure that Sam Altman’s company needs to scale up. It is also part of the Stargate project launched with great fanfare at the start of the second Trump administration.
But this spending on AI infrastructure naturally has a significant cost, especially as companies are engaged in an arms race aimed at putting in place ever more computing power to beat their competitors, while the positive impacts of AI are slow to appear. In December, Oracle had to revise upwards its capital expenditure forecasts for fiscal year 2026 by an additional $15 billion, which will therefore be added to the $35 billion previously communicated, an already enormous figure. For comparison, Oracle spent 21 billion in capex in 2025 and only 7 billion in 2024.
Faced with these staggering expenses, the cash flow of these companies is no longer sufficient, and they are now forced to go into debt. In February, Oracle announced plans to raise between $45 billion and $50 billion through debt and equity offerings in 2026 to fund its expansion. At the same time, Google indicated its desire to issue bonds over a century.
These expenses end up worrying investors, especially since Wall Street analysts predict that Oracle’s cash flow will remain negative for several years before AI-related spending begins to bear fruit, a turning point that they place (for the moment) around 2030. The company is therefore forced to lay off workers to regain financial margin.
Nvidia is hiring like crazy
And she is far from the only one to act this way. Since the start of the year, Salesforce has laid off a thousand people, Amazon 16,000 (after already 14,000 people laid off last year), while Meta announced its desire to reduce the staff of its division dedicated to augmented reality by 10% to focus on AI. Last year, Microsoft let go of 15,000 people. What these companies all have in common is having invested massively in AI infrastructure.
Although many experts fear that AI will destroy many jobs, the studies carried out on this subject do not currently allow us to draw conclusions in this direction. “Whatever way we analyze the data, we see no significant impact of AI on the labor market,” notes a study by the Economic Innovation Group published last summer. Will Raderman, a researcher specializing in the evolution of employment, concluded at the same period that, contrary to what several press articles assumed, AI was clearly not to blame in the weakening of the job market for young graduates.
As political scientist Ian Bremmer argues in a recent note, “other factors – trade wars, inflation, high interest rates – have played a more significant role in the languishing labor market.” However, these factors, and in particular the rise in interest rates, by increasing borrowing costs, also make investments by tech giants more costly and push them to lay off workers to regain financial margin.
It is interesting, in this regard, to observe what is happening within the companies which are already getting the most out of the fire thanks to AI, namely the specialists in the chips necessary for training and operating the algorithms, Nvidia in the lead. Far from having made massive layoffs, it now has 42,000 employees, compared to 36,000 at the start of 2025, 29,600 at the start of 2024 and less than 14,000 at the start of 2020. AMD, Nvidia’s great rival, has doubled its workforce in three years. Until proven otherwise, where AI pays off, it generates hiring; where it generates costs, it leads to layoffs…
