August 26, 2026, (Inside AI) — Meta Platforms scaled back a sweeping restructuring plan in May that would have cut up to 60% of employees across many teams, after internal data showed AI agents were not delivering expected productivity gains and workers raised alarms, according to sources familiar with the matter.
The company proceeded with layoffs affecting 10% of its workforce but canceled a second wave of restructuring. The reversal highlights the gap between AI investment rhetoric and operational reality at one of the world’s largest internet companies.
Global AI investment will exceed $1 trillion this year, estimates Goldman Sachs. Much of that spending bets on automation replacing human labor. Programming was expected to be an early win, since AI generates code faster than people. As a $1.4 trillion company and a leading data center spender, Meta serves as a crucial test case.
Meta’s headcount is shrinking while revenue per worker rises. Extending that trend could reduce total worker compensation, which topped $30 billion in the first half of the year. But data centers are expensive. Capital expenditure has consumed Meta’s formerly strong free cash flow, and investment is projected to reach nearly $170 billion next year, according to LSEG data.
Higher depreciation and amortization could offset any employment savings. Total employee compensation is actually rising. Excluding severance costs, Meta spent nearly 30% more in the first half of 2026 than in 2025. AI researchers command high salaries, and Mark Zuckerberg has bid up pay aggressively.
Problems extend beyond money. Changes to the codebase are proliferating faster than features reaching users. Security and reliability of automated work appear suspect. Plans to reorganize workers into “pods” and focus on top talent bred confusion. Resentment grew over job reassignments and the possibility that humans were training their AI replacements.
Ultimately, plans for firings and team restructurings were scaled way back. It is rarely obvious how best to use new technologies. Factories and offices took decades to redesign work around electricity and computers. Drastically refashioning a giant firm like Meta presents inherent problems.
The Innovator’s Dilemma Meets Meta’s Reality
Clayton Christensen‘s classic book, The Innovator’s Dilemma, explains the issue. Existing products, employees, and managerial structures are adapted to the status quo. Changing them during revolutionary industry shifts can cause them to fall apart.
“Moving fast and breaking things works for startups with little to lose. Zuckerberg might not be able to afford that mindset,” a source close to the matter said.
The canceled restructuring reflects a broader industry tension. Tech giants are pouring billions into AI while struggling to integrate it into legacy operations. Meta’s experience may foreshadow similar challenges at other large firms attempting rapid AI-driven transformation.
Meta did not immediately respond to a request for comment. The company has previously said it remains committed to AI investment and operational efficiency.
