91探花

Meta workers spent this week waiting to learn who would lose their jobs. The New York Times reported that staff exchanged salad emojis on internal forums as a salute to departing colleagues. Workers in Singapore received layoff emails at 4am local time, while employees in the United States and other countries received notices later that day.

The company told employees that 8,000 workers (which is 10% of the workforce) would lose their jobs as Meta rebuilt itself around AI. The New York Times also reported that 7,000 employees would move into new AI projects.

Meta plans to spend between $125 billion and $145 billion this year, more than double the amount spent in 2025. Much of that money will go into AI.

Mark Zuckerberg wrote to employees, 鈥淪uccess isn鈥檛 a given. A.I. is the most consequential technology of our lifetimes. The companies that lead the way will define the next generation.鈥�

Workers reacted with frustration after Meta reported record revenue last month. Employees asked why job losses were necessary when the company continued earning record income.

听

Are Businesses Using AI Headlines To Explain Ordinary Cuts?

听

James Lloyd, Digital Strategy Lead for Retail at THE LINE, NEOM, believes many companies are using AI language to explain decisions that started elsewhere.

He said, 鈥淵es, AI is becoming an easy headline for layoffs, but in some cases this is misleading. In many boardrooms the reality is more mundane: cost pressure, over-hiring corrections and the need to fund this massive AI investment.鈥�

Lloyd also said many companies are reorganising workers instead of removing them completely.

He said, 鈥淭he companies using AI well are usually redeploying and redesigning teams, not just simply replacing them. So when 鈥楢I-driven redundancies鈥� become the entire explanation, it deserves scrutiny.鈥�

The New York Times reported that Cisco announced 4,000 job losses while moving more resources into artificial intelligence. Microsoft, Block and Coinbase also announced layoffs or buyouts connected to AI programmes.

Meta also transferred hundreds of employees into a new Applied A.I. and Engineering team. The group already contains around 2,000 workers and fewer management levels than other divisions. Managers reportedly told employees that participation 鈥渨as not optional鈥�.

听

What Are Workers Saying About The AI Race?

听

Employees at Meta have openly criticised the company鈥檚 AI programmes. More than 1,000 workers signed a petition against employee data tracking used for AI training.

Meta software engineer Mack Ward wrote in an internal post, 鈥淎.I. is a freight train, but the future is not a foregone conclusion. It鈥檚 not too late to pump the brakes and consider how we, society, want to go about this.鈥�

Andrew Bosworth, Meta鈥檚 chief technology officer, also acknowledged the mood among staff during an internal meeting reviewed by The New York Times.

He said, 鈥淭here are a tremendous number of employees feeling anxieties about their futures. It鈥檚 all bad. I鈥檓 not going to try to sugarcoat that.鈥�

The company also offered retention packages to selected workers. The New York Times reported that one director level employee received additional equity worth roughly $500,000 to stay at Meta.

More experts chime in on whether AI is just the excuse behind something else鈥�

听

Our Experts:

听

• Matt Cockett, CEO, Dayshape
• Dima Beseda, Co-founder, Spiry
• Sayali Patil, AI Reliability Researcher & Practitioner
• Andrei Romanescu, CMO, LumaDock
• Jan Hendrik von Ahlen, Managing Director & Co-founder, Career Coach & Advisor, JobLeads

听

Matt Cockett, CEO, Dayshape

听

听

鈥淟ayoffs make for dramatic headlines, but they often hide the failures that led to them. The latest cuts have been framed as another sign that AI is reshaping work at breakneck speed. The truth is far more ordinary. Redundancies usually begin long before the announcement, in the months and years when organisations lose sight of how their own operations function. When leaders cannot see who is doing the work, how capacity matches demand, or where inefficiencies stack up, technology becomes a sticking plaster rather than a strategic tool.

鈥淎I amplifies this gap. It does not close it. The narrative centres on rapid technological change, but the underlying issue is a gap in workforce planning. Without a clear map of skills, strengths and potential pathways, firms struggle to redeploy people into emerging roles. If a firm does not understand how work gets delivered today, AI will not transform it tomorrow. Technology cannot compensate for a lack of operational clarity. When organisations automate processes they do not fully understand, they risk making poor decisions faster.鈥�

听

Dima Beseda, Co-founder, Spiry

听

听

鈥淐ompanies aren鈥檛 cutting people simply because AI can do their jobs, they鈥檙e raising the bar for what they expect humans to bring to the table. In today鈥檚 landscape, the real question is whether you know how to work alongside AI models in a way that makes you irreplaceable.

鈥淲e鈥檝e seen this firsthand. We now ask candidates to share their LLM chat history from test assignments. Of course, not to catch them using AI, but to see how they think with it. Everyone uses these tools now. The ones who stand out use AI as a thinking partner and bring judgment the model can鈥檛.

鈥淥n the flip side, we鈥檝e caught candidates who fabricated their entire CV with AI 鈥� skills, work history, all of it generated! That鈥檚 the dark side of this moment, and that鈥檚 why companies are very cautious.

鈥淎I makes it trivially easy to look qualified on paper, which means companies need new ways to verify who鈥檚 real. And yes, some will use 鈥楢I transformation鈥� as cover for layoffs they were planning anyway. The technology becomes the scapegoat for decisions that are really about cost-cutting.鈥�

听

听

Sayali Patil, AI Reliability Researcher & Practitioner

听

听

鈥淎I is being used as a narrative frame for decisions that were already made on spreadsheets.

鈥淚 have spent over a decade building automation systems inside large enterprise environments. What I know from that work is this: the organisations currently citing AI as the reason for workforce reductions are, in most cases, not the ones with the deepest AI integration. The causality is running in reverse. Headcount decisions driven by margin pressure, post-pandemic over-hiring, and rising capital costs are being announced alongside AI productivity narratives because it sounds more strategic than admitting the org grew too fast.

鈥淭he actual pattern I have seen repeatedly is different. When automation is genuinely embedded in production infrastructure, it tends to shift the composition of teams, not eliminate them. At Cisco, the automation platform I built and led processed over 1,200 tools across 20+ enterprise customers. It did not reduce engineering headcount. It eliminated the class of repetitive, low-judgment work that was quietly burning out senior engineers, and it freed capacity for the reliability and governance problems that actually required human decision-making.

鈥淲hat concerns me more than the layoffs themselves is the feedback loop this narrative creates. When enterprises see AI positioned as a cost-reduction lever, they start optimiSing for the wrong thing. They cut the observability and reliability engineering roles that keep AI systems honest in production. Six to twelve months later, they are dealing with model drift, silent failures in automated workflows, and governance gaps that are significantly more expensive to fix than the headcount they saved.

鈥淭he companies worth watching are not the ones announcing AI-driven cuts. They are the ones quietly redeploying technical talent into the infrastructure layer that makes AI systems trustworthy at scale. That work is harder to headline, but it is where the real competitive differentiation is being built.鈥�

听

Andrei Romanescu, CMO, LumaDock

听

听

听

鈥淭he 鈥淎I replaced these jobs鈥� framing is mostly PR cover for cost cuts that were already on the slide. A few specifics if useful:

First, the actual AI productivity gain inside large engineering teams runs around 10 to 20% on routine tasks per published studies, and that matches what we see in our own team. That鈥檚 a real number. It does not equal 鈥渨e cut 6,000 engineers because of AI鈥�. If a 6,000-person cut were justified by AI alone, the same companies would be visibly redirecting that headcount budget into compute spend and new GPU clusters. Most of them are quietly trimming compute spend at the same time. Obviously the two stories don鈥檛 add up.

鈥淪econd, the supply side data tells a different story than what鈥檚 in the press releases. DRAM and NVMe prices doubled in some configurations through Q1 2026 because AI inference demand exploded. Frontier-lab GPUs like the H100 and B200 have multi-quarter procurement lead times.

鈥淚f AI capacity were truly replacing 20% of an engineering org, the underlying hardware would be flying off shelves faster than it already is. The number of medium-sized companies still on waitlists for GPU capacity tells you real AI deployment is bottlenecked at the inference layer first, with the labour layer well behind that.

鈥淭hird, the cleaner explanation is post-2021 overhiring catching up with cost-of-capital. Microsoft and Meta both confirmed in 2023 they overhired during the zero-rate years. That correction is still working through the system. AI gives boards a more flattering narrative than 鈥渨e made expensive bets during a hiring boom and now we鈥檙e unwinding them鈥�.

鈥淭he AI framing lets a layoff announcement carry a forward-looking capex story in the same release. That paired narrative is the actual product the framing delivers for IR teams. Strip the AI line out and the cost-cut version hits the analyst call and the recruiting funnel a lot harder.鈥�

听

Jan Hendrik von Ahlen, Managing Director & Co-founder, Career Coach & Advisor, JobLeads

听

听

鈥淐ompanies are taking a massive risk with these AI-led layoffs. The main irony here is that the financial wins expected very rarely come true. Studies already show that aggressive AI automation hasn鈥檛 delivered the productivity companies expected, so many will end up hiring for the same roles again, often at higher cost than the roles they cut (Source: Fortune).

鈥淚t鈥檚 hard to say whether each layoff is really because of AI since 鈥淎I鈥� has become the most convenient label in the corporate communications world. It鈥檚 a buzzword that helps avoid harder explanations: e.g., why some companies overhired during the 2021 boom, slow growth, etc.

鈥淪o, a lot of very diverse layoffs are put down to AI automation. And don鈥檛 get me wrong, many companies are in fact reorganising their staff because of this. Others are just using it as cover. And in many cases the real reason is much more pragmatic: costs and financial planning.鈥�