The enterprise software giant slashed 13% of its workforce and recorded $1.84 billion in severance costs as it redirects capital toward data centers and AI infrastructure
Oracle just laid off roughly 21,000 employees, about 13% of its entire global workforce, leaving the company with 141,000 staff as of its latest fiscal year. The reason, stated plainly in Oracle’s own SEC 10-K filing: artificial intelligence deployment and a sweeping restructuring designed to funnel more money into data centers and compute infrastructure.
The price tag for showing those workers the door? A cool $1.84 billion in severance costs. Oracle is spending nearly $2 billion just to shrink itself so it can spend even more on chips, servers, and the physical plumbing of the AI economy.
The AI arms race gets expensive #
The big four, Alphabet, Microsoft, Amazon, and Meta, are projected to collectively pour between $600 billion and $650 billion into AI-related capital expenditures during 2026. That number has ballooned from previous years, and every dollar spent on GPUs and cooling systems is a dollar not going to headcount.
Oracle has reportedly secured a $300 billion contract with OpenAI for AI computing capacity, a deal that would make it one of the most critical infrastructure providers in the generative AI stack.
What this means beyond Silicon Valley #
Oracle’s filing explicitly names AI deployment as a driver of the layoffs. That’s notable because most companies prefer vaguely euphemistic language about “realignment” or “efficiency.” Oracle is essentially saying that the technology it’s building and selling is also eliminating roles within its own organization.
The $300 billion OpenAI deal, if it plays out at anything close to that scale, would represent one of the largest technology contracts in history. It would also explain why Oracle feels comfortable absorbing $1.84 billion in severance costs.
The crypto angle: parallel tracks, not intersecting ones #
Bitcoin miners have been quietly pivoting portions of their operations toward AI hosting and high-performance computing. The economics are compelling: the same power infrastructure and cooling systems that run mining rigs can serve AI workloads, often at higher and more predictable margins. Companies like Core Scientific and Hut 8 have already struck deals to provide data center capacity for AI firms.
The $600 billion-plus that major tech companies plan to spend on AI infrastructure in 2026 represents capital that’s competing for many of the same physical resources, electricity, land, cooling water, that Bitcoin mining depends on.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our