Inside Meta, the layoff day followed a familiar ritual: North American employees were asked to work from home while notifications went out. But this round came with a twist — the company is also steering roughly 7,000 employees into AI-focused roles as part of a sweeping reorganization, according to internal documents reported by Bloomberg, even as it cuts about 8,000 jobs and eliminates roughly 6,000 open positions. The moves, part of what Chief Executive Mark Zuckerberg has called the “Year of Efficiency,” shift staff from non-essential departments and experimental projects onto work supporting the company’s Llama models.
The company counted close to 78,000 employees globally at the end of March, according to company filings, meaning the reassignments and cuts together touch roughly a fifth of the workforce. The restructuring was first reported by The Information, which said Meta would begin the layoffs in waves and direct the reassigned employees onto projects supporting Llama and other AI infrastructure. Meta has not publicly detailed the changes.
The reorganization is structural rather than a pure cost cut. In a memo to employees, Meta’s human resources chief, Janelle Gale, said many teams were being redesigned around “AI-native design principles” to create flatter structures and smaller groups that can move faster with greater ownership. A large share of the cuts is expected to hit managerial positions as Meta strips out layers of middle management. The company has projected capital expenditure of $125 billion to $145 billion for 2026, largely directed at AI data centers, custom chips and model training, with total AI spending well above $100 billion.
Zuckerberg has framed the reorganization as making Meta faster, not just smaller: teams are being regrouped into small units with broad ownership that are expected to integrate AI into every level of operations. In the company’s telling, those pods are the mechanism by which AI reshapes workflows from the inside, rather than a layer of automation imposed on top of existing structures. The reassigned employees, many from departments deemed non-essential or from experimental Reality Labs projects, are being directed to work that bolsters the infrastructure around Llama, Meta’s family of large language models.
The message inside the company, according to people familiar with the matter, is that AI is no longer a research project confined to the FAIR lab: it is the backbone of every Meta service, from ad targeting to content generation in Reels. The shift echoes the efficiency drive Zuckerberg announced in 2023, when the company cut more than 20,000 jobs in two rounds and declared the “Year of Efficiency” — except this time the savings are being reinvested in a single technology line rather than returned to shareholders as a margin story.
The move has triggered an internal debate that spilled onto Hacker News, where commenters called it an “internal migration”: with 7,000 people moved into AI roles, who staffs the business lines they left behind? The answer, so far, is partly no one — Meta is simultaneously cutting roughly 10% of its workforce and eliminating open roles, betting that AI-assisted workflows will absorb the workload.
Employees face pressure to upskill in AI technologies — internal messaging has emphasized training in LLM technologies for the reassigned staff, according to the reports — and some have raised concerns about the monitoring tools used to train AI systems and how those tools are applied. Zuckerberg reportedly told employees that data collected through these efforts was intended only to improve AI systems, not for surveillance purposes. The transition is not seamless: employees moved into AI roles face a steep learning curve, and the cuts remove managers whose job was coordination, which means remaining teams are expected to absorb more scope with less supervision — a model that works if AI tools deliver the promised productivity gains, and a risk if they do not.
Investors have reacted positively to the efficiency focus, even as the human cost plays out internally. Analysts say the combination of massive AI investment and headcount reduction reflects a broader pattern across the industry: companies are spending on AI infrastructure while rebalancing costs by trimming staff and scaling automation. The market’s positive reaction reflects the arithmetic: with AI spending expected to exceed $100 billion this year, investors want to see that money concentrated where it can compound — model training, data centers and custom silicon — rather than spread across businesses with unclear AI relevance.
The reorganization is also a competitive response. Meta is trying to close ground with OpenAI and Google in frontier AI, and Zuckerberg has argued internally that the company’s distribution advantages — billions of users across Facebook, Instagram and WhatsApp — give it a path that pure research labs do not have. The move also signals how Meta plans to build its AI bench at a moment when every lab is bidding for researchers: retrain from within rather than hire from outside. The question for the rest of 2026 is whether the AI investments translate into products and revenue faster than the restructuring disrupts the businesses that currently pay the bills.
