Meta and Microsoft cut 23,000 jobs to fund AI spending as Big Tech reshapes workforce

Meta cuts 8,000 jobs and Microsoft offers first-ever buyouts as Big Tech converts payroll into AI capital expenditure

Summary: Meta and Microsoft announced workforce reductions on the same day, April 23, affecting up to 23,000 positions combined. Meta is cutting 8,000 jobs (10% of staff) and cancelling 6,000 open roles effective May 20, while Microsoft launched its first-ever voluntary retirement programme offering buyouts to up to 8,750 US employees whose age plus years of service equals 70. Both companies reported record revenues. Both are spending record amounts on AI infrastructure. The cuts are not about distress but about substituting human payroll for AI capital expenditure, a pattern that has now reached 96,000 tech workers in 2026. Meta and Microsoft announced workforce reductions on the same day, April 23, for what appears to be the same reason. Meta told employees it would cut approximately 8,000 jobs, 10% of its global workforce, effective May 20, and cancel 6,000 open positions. Microsoft disclosed its first voluntary retirement programme in 51 years, offering buyouts to up to 7% of its American employees, roughly 8,750 people, under a formula that requires a worker's age plus years of service to equal 70 or more. Between the two companies, up to 23,000 positions will be eliminated or never filled. Both companies reported record revenues in their most recent quarters. Both are spending more on artificial intelligence infrastructure than they have ever spent on anything. The cuts are not about financial distress. They are about what the money is for. Meta's chief people officer, Janelle Gale, wrote in an internal memo leaked before the company intended to release it that the layoffs were "part of our continued effort to run the company more efficiently and to allow us to offset the other investments we're making." The investments she was referring to are considerable. Meta has guided capital expenditure of $115 billion to $135 billion for 2026, nearly double the $72 billion it spent in 2025, directed almost entirely at data centres, Nvidia GPUs, custom silicon, and the infrastructure supporting its Llama model ecosystem and the newly created Meta Superintelligence Labs. Full-year 2025 revenue was $201 billion. Full-year 2025 net income was $22.8 billion in the fourth quarter alone. The company is not cutting because it cannot afford its workforce. It is cutting because it would rather spend the money on machines. Microsoft's calculus is quieter but structurally identical. The company's "Rule of 70" formula, which makes employees eligible for a buyout if the sum of their age and years of service reaches that threshold, disproportionately targets workers in their fifties and sixties who built the pre-AI Microsoft. Sales incentive plan employees are excluded. Full details will be communicated on May 7. CEO Satya Nadella warned in October 2025 that 2026 would be "messy" as the industry moved from AI demonstrations to AI integration. Second-quarter fiscal 2026 revenue was $81.3 billion, up 17% year over year. Azure grew 33%, with AI services contributing 16 percentage points of that growth. Microsoft's voluntary retirement programme for US workers is framed as a benefit. Its effect is to accelerate the departure of the employees least likely to transition into AI-native roles, using their severance as a rounding error against the hundreds of billions the company is committing to data centres, Copilot, and its OpenAI partnership. The 23,000 positions affected on April 23 are not an anomaly. They are the latest entries in a ledger that has been growing all year. Oracle eliminated up to 30,000 roles in March, roughly 18% of its workforce, to redirect an estimated $8 billion to $10 billion in annual cash flow toward a $156 billion AI infrastructure buildout. Amazon restructured 16,000 positions. Dell cut 11,000. Snap reduced headcount by 1,000, or 16%. According to industry trackers, more than 96,000 tech workers have lost their jobs in 2026 so far, a 40% increase over the same period in 2025. Oracle eliminated up to 30,000 roles to fund $156 billion in AI infrastructure, and its remaining performance obligations stood at $523 billion. It posted a 95% jump in net income the quarter before it made the cuts. The companies doing the firing are not the ones losing money. They are the ones making the most. Mark Zuckerberg has now cut approximately 25,000 jobs at Meta since 2022. The first two rounds, 11,000 in November 2022 and 10,000 in March 2023, were branded as the "Year of Efficiency" and came after the company's stock had collapsed on metaverse spending and a digital advertising downturn. Those cuts were defensive. The current round is offensive. Meta's stock is roughly flat year to date because investors already expect the headcount reduction to fund AI acceleration. Bank of America projects $7 billion to $8 billion in annualised savings. Wedbush's Dan Ives said the company is using AI to "automate tasks that once required large teams." CNBC's Jim Cramer called it a "screaming buy." Meta's earlier rounds of cuts across Reality Labs and recruiting in January and March 2026 eliminated approximately 2,200 positions and slashed the Reality Labs budget by 30%, suggesting the April announcement is the culmination of a restructuring that has been under way for months, not a sudden decision. What separates the Meta and Microsoft announcements is not the logic but the method, and the method reveals something about each company's relationship with its workforce. Meta is firing people. The memo from Gale described the cuts as involuntary and companywide, touching every major business unit. Engineers are being reassigned to a new Applied AI division and a small-business advertising group. The language shifted from the January 2025 round, when Zuckerberg framed the cuts as removing "low performers," to April 2026, when the framing became "contribution" and "efficiency," an acknowledgment that the people being let go are not necessarily underperforming. They are in the wrong part of the company. Hours before the April layoff memo leaked, Meta had awarded its six most senior executives stock options worth up to $921 million each, tied to a $9 trillion market capitalisation target by 2031. Meta awarded executives up to $921 million in stock options while cutting staff, and in the same period reduced stock-based compensation for rank-and-file employees by 5% to 10%. Microsoft, by contrast, is offering to pay people to leave. The voluntary retirement programme is unprecedented in the company's history and is designed to avoid the reputational damage of mass involuntary layoffs while achieving the same structural outcome. But calling it voluntary obscures the targeting. The Rule of 70 formula means a 55-year-old with 15 years of service qualifies. A 30-year-old with five years does not. The programme selects for age and tenure, not performance, and the divisions most affected, Azure cloud operations, gaming, and global sales, are the ones where automation via Copilot and AI agents is furthest advanced. Microsoft had already been tightening performance management throughout 2025, instructing managers to issue 30% more performance improvement plans and barring employees who failed benchmarks from reapplying for two years. The buyout is the softer instrument. It follows the harder ones. A survey of 1,000 US hiring managers by Resume.org found that 55% expect layoffs at their companies in 2026, and 44% identified AI as the primary driver. A Motion Recruitment study found that AI adoption is slowing hiring for entry-level and generalised IT roles while creating intense demand for AI specialists. The term analysts are using is the "AI employment paradox": companies are simultaneously cutting headcount and investing record sums in AI infrastructure, producing a labour market in which aggregate spending is rising and aggregate employment is falling. Economists quoted by CNBC described the situation as an AI-driven labour crisis that "is here, not coming in the future." A poll found that 57% of Americans think AI is advancing too fast, and 79% are concerned the government has no plan to protect workers from AI job losses. No legislative or regulatory response has materialised. The human cost of AI-driven tech layoffs is difficult to measure in aggregate because companies attribute cuts to AI without demonstrating that AI systems have actually absorbed the displaced work. The term "AI-washing" has emerged to describe this pattern: a company announces layoffs, cites its AI strategy in the same breath, and lets investors draw the inference that machines are replacing humans even when the company has no mature, scalable AI implementation capable of doing so. Meta's new Superintelligence Labs and Meta Compute division exist on paper and in press releases. Whether they can perform the work of 8,000 eliminated employees is a question that will not be answered for years. What can be answered now is where the money is going. Meta is spending $115 billion to $135 billion on AI capital expenditure. Microsoft spent $81 billion in the last fiscal year on capital investment and has committed to spending more. The salaries of the people being cut are a fraction of those figures. The cuts are not funding the AI buildout. They are a signal that the AI buildout has made the current workforce configuration obsolete in the eyes of management, whether or not the technology has made it obsolete in practice. Since 2020, nearly 900,000 tech workers have been laid off globally, according to the tracking site Layoffs.fyi. The first wave, in 2022 and 2023, was attributed to the unwinding of pandemic-era hiring. The second wave, in 2024 and 2025, was attributed to restructuring around AI. The third wave, now under way in 2026, no longer needs attribution because the companies are stating the connection explicitly. Oracle said it was cutting jobs to build AI data centres. Meta said it was cutting jobs to offset AI investments. Microsoft structured a buyout programme that selects against the employees least aligned with its AI future. The direction is not ambiguous. The question is whether the substitution is real, meaning AI genuinely performs the work the displaced employees did, or whether it is financial, meaning the companies are converting payroll into capital expenditure because Wall Street rewards the latter more than the former. On April 23, the market offered its answer. Meta's stock fell 2.3% on the layoff news. It would have fallen further if investors thought the cuts were a sign of weakness rather than a down payment on a thesis they already believe. Twenty-three thousand positions, eliminated or abandoned in a single day by two of the most profitable companies on earth. The reactor did not change. The spreadsheet did.

The Downgrading of the American Tech Worker

Everyone in tech is worried about layoffs. Since the beginning of the year, Block (formerly Square) cut its workforce nearly in half, Oracle began laying off up to 30,000 people, Amazon announced yet another reduction in force, Snap laid off around 15 percent of its staff, and Pinterest announced plans to do about the same thing. Microsoft recently announced its first-ever voluntary buyouts, targeting longtime employees, which some are reading as a prelude to deep cuts. Most of the firms are paring down workforces that grew enormously while interest rates were low. But all of them have at least something to do with AI. Some firms, like Oracle and Microsoft, are reallocating capital to massive investments in the technology. Block's leadership says it believes it can now run its firm with far fewer people, while Snap, in a more urgent cost-cutting mode, signaled that automation may save money. Job loss is the dominant fear as announcements pile up and employees absorb warnings about mass displacement from AI leaders like Anthropic CEO Dario Amodei. But two recent changes at Meta highlight another question: What does all this mean for the people who are left? The first, of course, is a layoff, which Meta characterized as part of a "continued effort to run the company more efficiently and to allow us to offset the other investments we're making." The firm has spent -- often haphazardly -- more than $70 billion developing AI so far, with plans to spend more in the coming year (in addition to spending a lot of its cash on AI and freeing up more through layoffs, the company has been selling tens of billions in bonds). The second change came just before the layoffs and was reported first by Reuters: Meta is installing new tracking software on U.S.-based employees' computers to capture mouse movements, clicks and keystrokes for use in training its artificial intelligence models, part of a broad initiative to build AI agents that can perform work tasks autonomously, the company told staffers. "The vision we are building towards is one where our agents primarily do the work and our role is to direct, review and help them improve," one memo said. "This is where all Meta employees can help our models get better simply by doing their daily work," said another. Just a few years ago in a healthier job market, installing screen-recording software on employee laptops would have been an enormous scandal for Meta and unthinkable for many of its employees, who were well compensated and in demand. Now, backed by the threat of more layoffs, their jobs are being both downgraded and assessed for possible automation (employees are indeed quite angry about this). These are common conditions in many industries in which workers have less leverage; indeed, even at Meta, subcontractors hired to moderate content and label data have long dealt with constant surveillance and threats of job loss and automation. Meta's new chief AI officer, Alexandr Wang, comes from this world: His startup, Scale AI, oversaw a vast workforce of contractors who did the "grunt work" involved in AI training and deployment and specialized in collecting the sorts of data Meta is now gathering by monitoring its own employees. It may be the case that collecting the mouse movements and keystrokes of thousands of Meta employees provides a real advantage in training AI models or designing agentic software. What's certain is that Meta believes it's in a position to dramatically renegotiate its relationship with its workforce. Into the early 2020s, jobs at big tech companies were extremely desirable, attracting qualified workers from across the economy and drawing ambitious young graduates away from jobs in finance, law, medicine, etc. Now, rather than a track to long-term job security in a futureproof industry, some of tech's most visible leaders are eagerly making a new offer to their employees: the prospect of downward mobility, limited freedom, and de-skilling. They're being told that they may be replaceable and that to find out if they are, they'll have to try to train their replacements. The destruction of the aspirational tech job is itself an aspirational project: We don't yet know if cutting a workforce to the bone and rolling out a bunch of AI tools is a path to success for an established tech firm, and anyone who tells you it's easy to guess what the workforce composition of a 2026 "technology" company will look like in 2029 is probably a little too confident in their forecasting skills. (It's also worth nothing here that Meta, a successful social-media company that is struggling to catch up with newer AI labs, has tended to chase industry trends, not lead or correctly anticipate them, for the past decade.) For now, though, the message being sent to many workers in America's most vibrant economic sector, a pillar of the economy's success for the past 50 years, is loud, clear, and being heard: We might lay you off soon, and even if we don't, your job isn't what it used to be.