Oracle Just Cut 30,000 Employees. Here's What Engineers Need to Do Right Now.

On March 31, Oracle fired between 20,000 and 30,000 employees in a single day. Roughly 18% of the company's 162,000-person global workforce received termination emails from "Oracle Leadership" at 6 a.m. local time. No call from their manager. No heads-up from HR. Just an email saying their role had been eliminated as part of a "broader organizational change," followed by immediate revocation of all system access.

The day of the email was their last day.

Employees across the United States, India, Canada, Mexico, and Uruguay all got the email at the same time. India lost approximately 12,000 workers from Oracle's 30,000-person local workforce. In Seattle alone, nearly 500 positions were cut. Entire divisions bled out: Revenue and Health Sciences (RHS) and SaaS and Virtual Operations Services (SVOS) each saw workforce reductions of 30% or more. Software engineers, account executives, program managers, and staff across Oracle Health, Cloud, NetSuite, Sales, and Customer Success all got the same email.

The reason Oracle gave: the company needs $8 to $10 billion in freed-up cash flow to fund AI infrastructure. Capital expenditure has surged from roughly $8 billion in fiscal year 2024 to a projected $50 billion in fiscal year 2026. Oracle is building data centers for clients like OpenAI and Meta. Larry Ellison has said publicly that AI now does Oracle's coding.

Oracle's stock didn't collapse. The company's quarterly net income sits around $6 billion. It's not struggling. It disclosed a $2.1 billion restructuring plan in its March 2026 SEC filing, which means the cuts were premeditated and budgeted. This was not a company reacting to a crisis. It was a profitable company choosing to replace headcount with infrastructure.

This is not a normal layoff

When a company is losing money and cuts staff, the advice is straightforward: update your resume, tap your network, land somewhere stable. Oracle's situation is different. The company is making money. Lots of it. The threat it's acting on is not financial distress but a belief that certain categories of work no longer require the number of humans currently doing them.

This is the same pattern that played out at Block in February, where Jack Dorsey cut 40% of the workforce while posting 24% gross profit growth and called it "100 people + AI = 1,000 people." It is the same pattern that played out at Atlassian two weeks later, where 1,600 employees were cut with nearly identical AI-restructuring language. Pinterest followed with a 15% cut.

Oracle's version is bigger and blunter. No CEO letter framing it as a philosophical evolution. No pretense of partnership with the departing employees. A 6 a.m. email, a DocuSign for severance paperwork, and system access gone before breakfast.

One Principal Product Manager who was cut wrote on LinkedIn: "This was honestly a shock. Like many of us, I put in long hours and late nights, believing deeply in the importance of what we were building and the impact it would have."

On Reddit, someone shared that their father, a 20-year Oracle veteran, was let go without even a phone call. He has cancer and now has no health insurance.

IBTimes reported that employees with more than 30 years at Oracle were among those cut. Former employees claimed an algorithm picked who got fired. Oracle has not confirmed this.

Wall Street's read is simple: Oracle's free cash flow went from positive $11.8 billion in 2024 to a projected negative $23 billion in 2026, driven by AI infrastructure spending. Cutting 30,000 employees is how you close that gap without slowing down construction. The employees aren't being punished for poor performance. They are a line item being moved from one budget to another.

Ask the question nobody wants to ask: is my work the kind AI replaces?

Oracle's cuts didn't follow performance ratings. They followed role categories. If you were in a division that Oracle decided could operate with fewer humans, your individual track record didn't save you. The 34-year veteran got the same email as the person hired last year.

That distinction matters. The question is no longer "am I doing good work?" It is "is the category of work I do one that my company's leadership believes requires fewer people?"

Run this audit on your own role:

• Could an AI tool handle 80% of my daily tasks with a human reviewing the output? If your work involves writing boilerplate code, triaging support tickets, generating reports, managing routine customer interactions, or processing standard configurations, the answer might be yes. That doesn't mean your job disappears tomorrow. It means your role is in the bucket that CFOs are now putting dollar figures on.

• Does my work require judgment that resists templating? The engineers hardest to replace are the ones making calls that AI cannot: architectural tradeoffs where the requirements are ambiguous, debugging production incidents where the failure mode is something nobody has seen before, navigating organizational dependencies where the blocker is political rather than technical.

• Am I the person who knows why, or the person who executes what? Oracle kept people who hold context that cannot be transferred to a system. The engineer who understands why the legacy database integration works the way it does. The one who knows which customers will break if you change the API contract. If your value is executing defined tasks efficiently, you are competing directly with the thing your company is spending $50 billion on.

• Would anyone outside my immediate team notice if my role disappeared for a month? If the answer is no, your work is probably invisible to the people making headcount decisions. Visibility is not a vanity project. It is how decision-makers determine which roles are load-bearing.

The survivors at Oracle are already learning what comes next. Remaining employees were told by senior management to "stretch" and "ramp up efficiency" to cover the workload left by departed colleagues. On Reddit, the response was blunt: "Do not, I repeat do not give even an hour extra of what you would usually give to work." The reasoning: if survivors quietly absorb the extra load and keep hitting targets, they prove the cuts were justified and invite another round.

Reposition your work around what AI cannot do

If the audit made you uncomfortable, the move is to reposition. Oracle's bet may be too aggressive. But the market incentive behind it is real, and other companies are watching.

Judgment under ambiguity. AI can generate code. It cannot decide what to build when the requirements are vague, the stakeholders disagree, and the technical constraints are unclear. If you are the person who takes a messy problem and turns it into a clear technical direction, that is the work to lean into. Write the decision document. Present the tradeoffs to leadership. Own the outcome, not just the implementation.

End-to-end ownership. The engineer who spots a problem before anyone asks about it, scopes the fix, ships it, and measures the result is doing something that AI tools cannot replicate. Oracle's layoffs hit hardest in roles that executed pieces of a larger process. The people who owned outcomes from problem identification through production monitoring were harder to cut.

Cross-team influence. Negotiating priorities across teams, unblocking dependencies through relationships, understanding why an org-level decision was made and adjusting your team's strategy accordingly. None of this is automatable. If you are doing it, make sure the people making headcount decisions know about it. If you are not doing it, start.

Building what the company is betting on. Oracle is spending $50 billion on AI infrastructure. The people building, operating, and selling that infrastructure are not getting cut. Every company has a version of this: the thing leadership is funding aggressively. If your work is adjacent to that bet, you are safer. If your work is in a cost center that the bet is designed to shrink, you need to move.

Prepare for the version of this that hits your company

Oracle did it with a 6 a.m. email. Your company will probably do it with a slightly more polished process and a slightly less honest explanation. But the mechanics are the same: leadership decides certain functions need fewer people, models the savings, and acts.

Watch for the signals:

• Productivity metrics being applied to your team. When leadership starts measuring output-per-head and comparing it to AI-augmented benchmarks, headcount decisions follow. Block's CFO cited a 40% increase in code shipped per engineer before the cuts came. If your company starts tracking similar numbers, take it seriously.

• Hiring freezes next to aggressive investment. Oracle was hiring AI infrastructure engineers while eliminating thousands of other roles. If your company is freezing backfills on your team while aggressively hiring in AI/ML, the message is not subtle.

• Reorganizations that consolidate your function. When your team gets merged into another team, or your manager's scope expands to cover two functions, that is often the step before headcount reduction. Fewer managers with broader scope means fewer individual roles to justify.

• Your company posting record revenue while talking about "efficiency." Efficiency language from a profitable company means margin expansion. The fastest way to expand margins is to cut the people who are not building the thing leadership is betting on.

Know your financial runway. Oracle's severance terms drew criticism for falling short of industry standards. In India, the formula was N+2 (years worked plus two months). In the US, specific terms have not been fully disclosed, but reports suggest they were not generous. All unvested RSUs were forfeited immediately, wiping out years of expected compensation for long-tenured employees. If you don't know how many months of expenses you can cover without income, figure that out before the email arrives.

Keep your experience documented. The worst time to reconstruct your accomplishments is the morning after you lose access to your company's systems. Keep a running record of your projects, their outcomes, and the quantified impact of your work. Do this while you still have access to the data.

The Oracle layoffs are not an outlier. They are what happens next.

In the first three months of 2026, Block cut 4,000 employees. Atlassian cut 1,600. Pinterest cut 15%. Oracle cut 30,000. Every company cited AI. Every company was profitable. The stocks mostly went up.

Each round is bigger. Block's cuts were framed as a visionary bet. Atlassian's were framed as a strategic realignment. Oracle's were barely framed at all. Just an email at 6 a.m. and a DocuSign link. The pretense gets thinner each time because the market has made it clear that it rewards the outcome regardless of the framing.

If you were not affected by Oracle's cuts, this is not a reason to exhale. It is a reason to ask whether you would survive the version of this that your company runs in six months. The engineers who come through this period in good shape will not be the ones who worked harder or stayed late. They will be the ones who looked honestly at what they do every day, asked whether that work survives an AI restructuring, and repositioned while they still had the leverage to choose.

You cannot control whether your CEO decides that AI infrastructure is worth more than your team. But you can control whether you are doing work that keeps you on the human side of that math.

---

CareerClimb is an AI career coach that helps engineers document wins, build their case, and stay ready for whatever comes next. Download the app and start building your case today.