Oracle Just Fired 10,000 People for AI. One of Them Described the Email as “Thank You. Go F*** Yourself.”

By his third failed attempt to log into Oracle’s VPN on Tuesday morning, a decades-long employee of the company started to get a bad feeling. When Slack stopped responding too, it got worse. Then he checked his email.

“Basically it said ‘Thank you. Go fuck yourself,’” he told The Register, asking not to be named because he feared it might affect a severance package he had not yet received any information about.

He was one of an estimated 10,000 Oracle employees who received that email on April 1, 2026. Senior engineers. Architects. Operations leaders. Program managers. Technical specialists with decades of institutional knowledge. All of them gone, in a single morning, before most people had finished their first cup of coffee.

Oracle has not confirmed the number.It has not commented publicly at all. The figure of 10,000 comes from employees counting the drop in active accounts in Oracle’s internal Slack workspace overnight. That is the kind of detail that emerges when a company fires so many people that the workforce itself becomes the measurement tool.

The official reason, to the extent Oracle has offered one, is AI.

Bloomberg reported in March that Oracle was planning significant job cuts as it deals with a cash crunch from its aggressive push to build AI data centers. The company has committed to spending at least $50 billion on AI infrastructure this year, raised another $50 billion in debt to meet demand, and signed a contract with OpenAI worth over $300 billion for computing capacity as part of the Stargate initiative.

To pay for all of this, it needs to cut costs elsewhere. The most visible cost in any company is people.

Co-CEO Mike Sicilia had already laid out the logic earlier this month: AI tools, he said, allow Oracle to deliver solutions to customers with smaller engineering teams. Investment bank TD Cowen had been even more specific in January, estimating that cutting 20,000 to 30,000 employees could generate $8 billion to $10 billion in incremental free cash flow. There is nothing subtle about that calculation. It treats people as a line item that can be optimized away to fund a bet on infrastructure.

Senior Operations Manager Michael Shepherd, who was not among those cut, posted on LinkedIn the same day. His message was worth reading in full.

“Today, Oracle conducted a significant reduction in force that impacted some of the most talented, dedicated, and high-performing people I have had the privilege of working alongside. Let me be direct: this was not a performance action. The individuals affected were not let go because of anything they did or didn’t do. Many of them are the people you call when something is truly broken, the ones who show up early, stay late, and carry institutional knowledge that took years to build.”

They were made redundant because the company needed to redirect capital toward data centers, and they happened to be in the way of that redirection.

The decades-long employee who spoke to The Register was not bitter about the lack of a phone call. “When you are cutting that many people, it sort of makes sense,” he said. What he was clear about was the cause. It was AI, he said, that took his job. And it will take more.

There is a version of this story that Oracle would prefer you to read.

In that version, the company is making a bold strategic bet on the future of AI infrastructure. The job cuts are painful but necessary. The people who remain will be more productive, better compensated, working at the frontier of a technology that will reshape the world. The short-term disruption is the price of long-term transformation.

This version is not entirely false. Oracle is genuinely making a massive bet. The infrastructure it is building may well prove valuable. The people who remain may genuinely benefit.

But this version leaves out the 10,000 people who received an email on April 1 that told them their final working day was immediate. It leaves out the unvested restricted stock units that were forfeited the same morning. It leaves out the employees in India — an estimated 10,000 there alone, representing roughly 20% of Oracle’s local workforce — who now face not just job loss but visa uncertainty and the prospect of returning to an already strained job market. It leaves out the H-1B visa holders in the United States who have a limited window to find new employment before they are required to leave the country.

The strategic version of this story is written for investors. The other version is written at kitchen tables, at midnight, by people updating their CVs.

What makes the Oracle layoffs significant beyond their scale is what they reveal about the current moment.

This is not a company that tried AI, found it wanting, and cut jobs anyway. Oracle is not Klarna, which replaced 700 customer service workers, watched quality decline, and quietly started rehiring. Oracle is cutting people to fund the infrastructure that will make AI possible at scale in the future. The displacement is not a consequence of AI working. It is a down payment on AI working later.

That is a different and in some ways more troubling logic. The people who lost their jobs this week are not being replaced by a chatbot that can do what they did. They are being replaced by a capital allocation decision — a judgment that the money currently spent on their salaries is better deployed in data centers that will, eventually, power AI systems that will, eventually, change how software is built and sold.

The word eventually is doing a lot of work in that sentence. And the people at the kitchen tables do not have the luxury of eventually.

Anthropic’s research published in March 2026 found that the jobs with the highest AI exposure are concentrated in exactly the kinds of roles Oracle just eliminated: computer programming, technical operations, data management, customer support. The paper also noted that the pattern of AI displacement tends to be quiet and gradual rather than dramatic and sudden — a slowing of hiring rather than a wave of layoffs, an absence rather than a visible event.

Oracle on April 1 was visible. Unusually, jarringly visible. Ten thousand people, one morning, an email that said thank you and goodbye and here is your severance paperwork. The Register described one employee’s experience as the moment a decades-long loyalty to a company collapsed into a VPN that would not connect and a Slack that had gone silent.

That employee said he was not upset about the manner of it. He understood the logistics. What he was clear about was the meaning.

It was AI. And it will take more.

I have written elsewhere on this site about what it means to watch a profession transform from inside it. About my sister, who lost her translation work without seeing it coming. About my own legal work, where AI has given me something I had not had in a decade — evenings, a private life, the sense that the work does not have to consume everything. I have tried to be honest that the technology is genuinely useful, and that the disruption it causes is genuinely real, and that both of these things are true at the same time.

The Oracle story does not complicate that view. It confirms it.

The technology is coming. The timeline is uncertain. The people in its path did not choose to be there. And the emails will keep arriving at six in the morning, telling people that their institutional knowledge, their years of showing up early and staying late, their decades of loyalty to a company and a story — that all of this has been weighed against a capital allocation decision, and the capital allocation decision won.

Thank you. Go fuck yourself.