Layoffs.fyi shows over 157,000 tech workers lost their jobs in 2025 through November—and December isn't finished yet. That's down from 2024's 262,000 but still brutal, especially when you consider the context: companies are posting record profits while citing "AI efficiency gains" to justify cutting headcount. The disconnect between corporate performance and worker security has never been starker.
The Numbers Tell a Story
157,000 layoffs across 587 companies. Largest cuts came from: Amazon (16,000+ workers), Google (12,000+), Microsoft (10,000+), Meta (5,000+), and dozens of smaller firms making "strategic workforce adjustments."
But here's the thing: most of these companies are printing money. Amazon's Q3 revenue was up 11% year-over-year. Google's ad business is stronger than ever. Microsoft's cloud growth continues. These aren't struggling companies cutting costs to survive. They're profitable companies optimizing margins.
The stated reasons always include some variation of "restructuring for AI efficiency," "focusing on core strategic priorities," or "rightsizing for current market conditions." Translated: we can do more with fewer people now that AI handles some tasks, and we'd rather boost profits than maintain headcount.
The AI Productivity Paradox
CEOs keep touting how AI makes their companies more productive. Microsoft's Satya Nadella talks about Copilot transforming knowledge work. Google's Sundar Pichai emphasizes AI-driven efficiency gains. Meta's Mark Zuckerberg calls 2024 "the year of efficiency" with AI as the enabler.
If AI makes workers more productive, why are companies laying people off instead of growing output? The logical conclusion: AI isn't making workers more productive—it's making workers redundant.
A support agent with AI tools can handle 2x the ticket volume. Great! Instead of hiring more agents to serve more customers, companies keep the same customer base and fire half the agents. The productivity gain goes to shareholders, not workers.
This is the pattern across functions: customer service, content moderation, basic coding, data entry, bookkeeping. AI doesn't eliminate the work, it reduces the number of humans needed to do it.
Who's Actually Getting Hit
Entry-level positions are disappearing. Junior developers, customer support reps, content moderators, QA testers—roles where AI can handle routine tasks with minimal supervision.
Companies used to hire entry-level workers and train them into mid-level roles. Now they're hiring senior people with AI augmentation skills and skipping the junior positions entirely. That breaks the career ladder for anyone trying to break into tech.
Recruiters and HR roles are getting cut heavily. Companies use AI for resume screening, initial outreach, and candidate evaluation. Human recruiters are reserved for final-stage interviews. Some companies have cut recruiting teams by 50%+ while maintaining similar hiring volumes.
Mid-level managers are also vulnerable. If AI can coordinate work and provide status updates, you need fewer managers. Flatter org structures with wider spans of control become viable when AI handles routine coordination tasks.
The "Efficiency" Narrative
Every company frames layoffs as strategic optimization, not cost-cutting. "We're focusing resources on high-priority areas." "We're streamlining operations for faster execution." "We're positioning for long-term growth."
But look at the financials: stock buybacks, dividend increases, and executive compensation are all up. If companies were genuinely struggling, they'd cut buybacks first. Instead, they're cutting workers while rewarding shareholders and executives.
AI gives companies cover for layoffs that are really about margin expansion. "We're implementing AI tools that improve efficiency" sounds better than "we're cutting staff to boost profits." Both statements can be true simultaneously.
The Severance Package Reality
Large tech companies still offer decent severance—typically 2-3 months salary plus benefits continuation and sometimes equity acceleration. That's better than nothing, but it's cold comfort when the job market is tight and hundreds of thousands of others are competing for the same roles.
Smaller startups are less generous. Some offer minimal severance or none at all, especially if they're citing financial difficulties. The WeWork model of "sorry, we're shutting down, good luck" is becoming more common.
What almost no company offers: retraining assistance for roles impacted by AI. If your job was automated away, you're on your own to upskill or pivot. Some companies provide resume services or job placement support, but actual retraining? Rare.
The Mental Health Impact
Beyond the financial stress, mass layoffs create psychological damage that doesn't show up in spreadsheets. Survivors deal with increased workload, fear of future cuts, and guilt about still having jobs. Those laid off face identity crises, loss of community, and the grinding stress of job searching in a competitive market.
LinkedIn is full of "I've been affected by reductions" posts where people perform public job searching. It's dehumanizing but necessary—the algorithm favors engagement, and posting about layoffs gets engagement.
The mental health services most companies offer are capped at a few free therapy sessions, barely enough to process the initial shock, let alone deal with prolonged unemployment stress.
What Companies Should Be Doing
If AI genuinely makes companies more productive, the gains should be shared. Options include: reduced workweeks with maintained pay, profit-sharing bonuses, investment in upskilling programs, or expanded benefits.
Instead, we're seeing the opposite: productivity gains go to shareholders, workers get laid off, and everyone who remains is expected to do more with AI tools for the same or lower compensation.
Some European countries are experimenting with work-sharing programs where companies reduce hours instead of cutting headcount. That spreads the pain more evenly and maintains employment. U.S. companies have little incentive to pursue that approach when cutting workers is cheaper.
The Policy Vacuum
There's no meaningful policy response to AI-driven workforce displacement. Retraining programs are underfunded and disconnected from actual market needs. Unemployment benefits haven't kept pace with the cost of living. And nobody's seriously discussing more dramatic interventions like universal basic income or mandatory work-sharing.
Politicians talk about preparing workers for the "jobs of the future," but what are those jobs? If AI handles code generation, customer service, data analysis, and creative content, what's left for average-skilled workers to do?
The optimistic case is that new job categories emerge we can't imagine yet, just as the internet created entirely new professions. The pessimistic case is that AI eliminates more jobs than it creates, and we're not prepared for that transition.
The Hiring Contradiction
Here's the weird thing: many companies laying off workers in one department are simultaneously hiring in others. They're cutting customer support while hiring AI engineers. Eliminating junior developers while recruiting senior AI/ML specialists.
This isn't downsizing—it's workforce transformation. Companies are shedding roles that AI can replace and hiring roles needed to implement and maintain AI systems. That's economically rational but socially destructive when workers can't easily transition between those roles.
A customer support rep can't just "learn AI engineering" and slot into a newly created ML role. The skills are completely different. Telling people to upskill is meaningless without accessible, effective programs that actually lead to employment.
What Workers Can Do
The practical advice is depressing: develop skills that AI can't easily replicate. Creativity, strategic thinking, complex problem-solving, and interpersonal skills. Easier said than done when companies are cutting training budgets and expecting workers to self-educate while handling increased workloads.
Building multiple income streams helps but isn't accessible to everyone. Not everyone can freelance, consult, or start side businesses. Many people just need a stable job that pays the bills.
The harsh reality is that individual workers have limited ability to protect themselves from macro trends. When entire job categories disappear, no amount of personal hustle prevents displacement.
My Take
Tech companies claim to be building AI for humanity's benefit, then immediately use it to cut costs and boost margins at workers' expense. The cognitive dissonance is stunning.
I don't blame companies for pursuing efficiency—that's what capitalism incentivizes. I blame the system that allows companies to reap all the gains from AI productivity while workers bear all the risk.
We need structural changes: stronger safety nets, mandatory profit-sharing when productivity gains come from automation, retraining programs that actually work, and possibly more radical interventions if displacement accelerates.
But instead, we're getting LinkedIn posts about resilience and networking while 157,000 people hunt for jobs in an increasingly AI-optimized economy that has less need for human labor.
That's not a sustainable path. Something has to give.