This month, three major US technology companies eliminated a total of 46,700 jobs, all driven by the rise of artificial intelligence.
On a regular Tuesday, tens of thousands of Oracle employees opened their work emails to find a message from 'Oracle Leadership.' Sent at six in the morning, the brief five-line email contained a single core message: "After careful consideration, your position has been eliminated. Today is your last working day."
There was no prior notice from HR, no advance conversation with direct managers, and no warning. Tens of thousands of people lost their jobs instantly. As they read the email, their system access was revoked, locking them out of company networks, and their office entry badges were deactivated. That morning, Oracle executed the largest layoff in its 48-year history, estimated to affect between 20,000 and 30,000 employees, representing approximately 18% of its global workforce of 162,000. Oracle forums on Reddit and the anonymous workplace platform Blind were quickly flooded with posts expressing shock and anger—some reported entire teams being eliminated within an hour, while others, after more than a decade of service, were denied even a face-to-face farewell.
On another ordinary weekday, a Thursday, Meta announced it was cutting about 8,000 jobs, roughly 10% of its global workforce, and eliminating 6,000 open positions that had not yet been filled. Unlike Oracle's sudden move, Meta's layoffs were anticipated by both employees and industry observers, following months of prior indications. On the same day, Microsoft made an unprecedented offer of early retirement packages to approximately 7% of its US workforce.
Employees have become casualties in the AI race. However, unlike previous large-scale layoffs, these three tech giants are not facing financial distress; in fact, their recent earnings reports have hit record revenues. So, what is driving this wave of job cuts?
The answer is evident in internal documents. Oracle co-CEO Mike Sicilia stated earlier this year, "The use of AI programming tools allows engineering teams to deliver more comprehensive solutions with fewer people." In other words, AI is replacing human labor, and companies are restructuring around this new reality.
Financially, Oracle is not struggling. Last quarter, its net profit surged 95% year-over-year to $6.1 billion, and its backlog of signed but unfulfilled contract revenue soared to $523 billion, a 433% increase—driven in part by a single contract with OpenAI worth over $300 billion. This is not a company in a revenue crisis but one leveraging its profitable present to make a massive bet on an AI-driven future. According to analysis by investment bank TD Cowen, these layoffs are expected to free up $8 to $10 billion in cash flow, which will be directed toward building AI data centers.
The logic at Meta and Microsoft is similar. While Meta's Chief People Officer, Janelle Gale, did not explicitly mention AI in an internal memo, she emphasized that the layoffs were part of "ongoing efforts to enhance operational efficiency and create space for other strategic investments." In essence, Meta aims to reduce labor costs to free up financial resources for its expensive technological transformation. Those "other strategic investments" can only mean AI. Meta's capital expenditure budget for this year is set at $115 to $135 billion, nearly double that of the previous year, with almost all of it directed toward AI. CEO Mark Zuckerberg has even scaled back his previously ambitious metaverse vision, cutting jobs and reducing investments in metaverse teams. Microsoft, for its part, has already committed $81 billion in capital investments this fiscal year and promises more.
The future of AI appears promising, but to fund this high-stakes arms race, tech giants are prioritizing layoffs to raise capital, especially as AI increasingly assumes many job functions. The reality is stark: the total salaries of laid-off workers amount to only a fraction of these AI investments. This is not about cost-saving in the traditional sense but a reallocation of resources—a signal to the market that AI has made existing staffing levels redundant. Last year, Salesforce CEO Marc Benioff announced a halt to engineering hires, explaining bluntly, "With AI, I don't need as many people now."
Oracle, Meta, and Microsoft are just the most visible waves in a broader tide of layoffs sweeping the U.S. tech industry. According to Layoffs.fyi, a site tracking tech industry job cuts, 98 tech companies have announced layoff plans this year, eliminating over 92,000 positions—an average of 864 people losing their jobs each day. The strategy for most of these companies is the same: slash labor costs on a large scale while funneling capital into the AI arena. Amazon, Alphabet (Google), Meta, and Microsoft are projected to spend a combined $700 billion in capital expenditures this year, nearly all flowing into AI-related infrastructure.
Analysts term this phenomenon the "AI employment paradox," where companies are simultaneously cutting staff and making record investments in AI. The victims in the labor market are those whose roles have become redundant due to efficiency gains from AI tools.
Employees are, in effect, training the AI that replaces them. A statement from Alphabet (Google) CEO Sundar Pichai is telling: AI tools have improved engineering efficiency by 10%. This efficiency gain easily translates to a 7% to 10% reduction in staff—not a coincidence but a management logic now openly used by tech giants.
Adding to the irony, these companies are accelerating efforts to have AI learn employees' tasks to prepare for further automation. Meta recently required US employees to install AI monitoring software on their computers, not just for surveillance but to enable AI to understand work processes in detail and overcome current technical limitations. Meta CTO Andrew Bosworth stated in an internal memo that the goal is to develop AI agents capable of autonomously performing work tasks, with a future vision where "AI agents do the primary work, and humans provide guidance and review." This initiative has sparked significant internal controversy, and employees have no option to refuse.
Meta is arguably the most aggressive in this push. Last November, the company's Chief People Officer sent a memo announcing that starting in 2026, "AI-driven impact" would become a core metric in performance reviews—in other words, failure to use AI could hinder promotions and raises. A Microsoft executive was even more direct in an internal notice to managers: "Using AI is no longer optional; it's a basic requirement for every role and level."
This has quickly devolved into absurdity. Meta employees created an internal "AI Token usage leaderboard," encouraging colleagues to compete on monthly token consumption. In one 30-day period, the company's total token usage exceeded 60 trillion, with the top individual user consuming 281 billion tokens in a single month—costing over $1.4 million at Claude's lowest pricing tier. This practice, dubbed "Tokenmaxxing" in Silicon Valley, treats AI usage as a proxy for productivity, rewarding behavior over results and fostering inefficiency, performative work, and internal waste.
Duolingo provides a rare case of corporate candor. In April of last year, CEO Luis von Ahn announced an "AI-first" mode, mandating employee AI use and assessment, which triggered user backlash and employee resistance. He later publicly admitted he "didn't anticipate such a strong reaction" and withdrew the mandatory requirement, conceding it "felt like asking people to use AI for the sake of using AI, not to do their jobs better."
The parallel is clear: forcing employees to use AI to boost efficiency generates data proving which roles are redundant. Requiring them to showcase AI usage effectively helps the company map out functions ripe for automation. This isn't conspiracy theory but straightforward business logic: when a company can quantify the efficiency gains from AI replacing human labor, layoff decisions gain irrefutable data support—data provided by the employees themselves.
The debate over whether AI will eliminate or create jobs features starkly contrasting views from two influential Silicon Valley CEOs. Anthropic CEO Dario Amodei stands as one of the most prominent pessimists. He stated in an interview, "I expect AI to have a significant impact on white-collar office work within the next one to five years." He estimates that up to 50% of entry-level white-collar jobs in tech, finance, law, and consulting could disappear within five years, potentially pushing US unemployment to 10-20%. At the World Economic Forum in Davos this January, he explained that AI's cognitive breadth means it won't replace jobs industry-by-industry like factory automation but will impact multiple sectors—finance, law, consulting, tech—simultaneously. "AI isn't replacing one type of job; it's acting as a general-purpose substitute for human labor," he said, noting that many AI CEOs discuss this privately but avoid public statements due to poor PR. "As producers of this technology, we have a responsibility to be honest with the public about what's coming."
In contrast, NVIDIA CEO Jensen Huang remains characteristically optimistic. In a talk at Stanford Graduate School of Business, he asserted, "You're not likely to lose your job to AI; you're more likely to lose your job to someone who uses AI better than you." His argument is that AI changes the tools for completing work, not the purpose of the work itself. He cited radiologists as an example: in 2020, AI vision technology achieved superhuman accuracy, leading to predictions that machines would replace radiologists. Instead, the number of radiologists has increased because AI allows each doctor to handle more cases, expanding demand. "That alarmism went too far. It scared away people who might have entered this socially vital field—that caused harm." Looking ahead, Huang predicts AI will "harass you, manage you, and make you busier than before," but he believes "we will ultimately create more jobs. There will be more employment at the end of the Industrial Revolution than at the beginning."
Naturally, both CEOs' views are informed by their positions: Amodei builds AI intended to augment or replace human tasks, while Huang sells the chips that power AI—the more widespread and powerful AI becomes, the more NVIDIA benefits. Their assessments of AI's future cannot be entirely separated from these contexts.
Nevertheless, the current layoff wave is an undeniable reality. The most ironic detail is that NVIDIA itself is the biggest beneficiary of this trend, as every tech giant cutting jobs is doing so to buy more chips. Layoffs in the tech sector are becoming the most potent, direct evidence for the proposition that AI is eliminating white-collar jobs.
The cost of this AI arms race is now being felt beyond corporate balance sheets, affecting universities and the job market. Computer science, once the hottest major, is becoming one of the hardest degrees for securing employment. A Federal Reserve report on the labor market last year showed unemployment rates of 7.5% for computer engineering graduates and 6.1% for computer science graduates—the latter even higher than for liberal arts graduates. Research from the Stanford Digital Economy Lab indicates that in occupation categories with the highest AI exposure (namely IT and software engineering), the employment rate for workers aged 22-25 fell by 6%, while employment for those aged 35-49 in the same roles grew by 9%. The entry barrier for young people in tech is being systematically raised.
Analysis from labor research firm SignalFire is more direct: tracking major public tech companies and mature startups from 2019 to 2024, the number of new positions open to recent graduates (within one year of graduation) has fallen by 50%. This decline spans core functions like sales, marketing, engineering, and design, indicating a broad trend. Handshake recruitment data corroborates this: this year, the average number of applicants for tech internships nearly doubled compared to the previous year, with each internship attracting 273 applications.
Amid the pessimism, another perspective exists. A more accurate description might be that AI is not directly eliminating jobs but changing the entry requirements and nature of work. Magdalena Balazinska, Chair of the University of Washington's Computer Science & Engineering department, wrote in an open letter to students: "AI is not eliminating your career options; it is expanding them." This is true, but with a caveat: it expands options for those who proactively embrace AI tools while possessing uniquely human judgment; for those who rely solely on "writing correct code" as their core competency, opportunities are indeed narrowing.
AI programming tools significantly boost the productivity of experienced engineers, with a direct consequence: the need to hire junior developers decreases, while those who remain take on heavier responsibilities and higher expectations. Tools like GitHub Copilot and Cursor enable tasks that once required three to five junior employees to be handled by two experienced engineers using AI. This isn't just layoffs; it's a structural contraction in hiring demand.
Intuit CTO Alex Balazs described this shift: "Five years ago, developers wrote every line of code. Today, engineers spend time on more complex problems—because they no longer need to spend countless hours on boilerplate code and routine implementation." This represents an efficiency dividend but also erodes the traditional value of junior engineers.
The most profound impact of this layoff wave may not be the lost jobs themselves, but how it is reshaping the industry's standard for "what kind of employee is worth hiring." In the past, the talent pyramid in tech companies was clear: numerous junior developers handled routine coding, mid-level engineers managed system design, and a few senior architects made top-level decisions. AI is flattening the base of this pyramid. Companies are increasingly favoring "small, elite" teams, drastically reducing entry-level and generalist positions, and prioritizing the recruitment of versatile talent capable of leveraging AI tools and solving problems across domains.
Comments