🏠

Meta's AI Chaos 🤯: Project OT Failed 📉

August 28, 2026 | Author

AI

🎧 Audio Summaries
English flag
French flag
German flag
Japanese flag
Korean flag
Mandarin flag
Spanish flag
🛒 Shop on Amazon

🧠Quick Intel


  • Meta created “Project OT” in January to reduce team headcounts by up to 60%, aiming for an “AI native” structure.
  • The first round of layoffs from Project OT occurred in May, with Meta canceling the second planned round.
  • Meta executives, including Mark Zuckerberg, directed the implementation of Project OT, exploring scenarios where headcount could be reduced by 25% or more.
  • Meta began selling AI agents to third parties, starting in June, as part of its “AI native” strategy.
  • Changes to internal software platforms and infrastructure increased by 220% year-over-year in June.
  • Major technical and security incidents increased by 40% compared to the prior year, with employee time spent resolving them rising by as much as 70%.
  • Meta launched a pilot program restructuring teams in June, utilizing “agent-assisted analysis” and removing middle management.
  • The trajectory of AI agent development hadn’t accelerated as expected, according to Zuckerberg, over the last four months.

    • 📝Summary


    Earlier this year, Meta initiated a plan, codenamed Project OT, to reshape its operations, exploring potential reductions of up to 60 percent in team sizes. The company aimed to become “AI native,” leveraging AI tools to automate tasks and, initially, conducted scenario planning exercises involving layoffs. A first round of layoffs occurred in May, followed by a cancellation of a second planned reduction. Meta’s CEO, Mark Zuckerberg, directed the restructuring, which involved shifting thousands of employees to new priority teams and exploring AI-driven workflows. Internal documents revealed a focus on using AI agents, even considering selling them to third parties, alongside investments in high-performing AI engineering roles. Despite tracking employee activity and a surge in technical incidents, Zuckerberg ultimately halted the plan, highlighting a lack of accelerated progress in AI development, illustrating a significant, though ultimately unsuccessful, effort to integrate artificial intelligence into the company’s core functions.

    💡Insights

    ▼


    PROJECT OT: META’S AMBITIOUS AI TRANSFORMATION
    Meta initiated a radical restructuring plan, codenamed Project OT, aimed at transforming the company into an “AI native” organization. This involved exploring scenarios of up to 60% workforce reductions, driven by the potential of AI to automate tasks previously performed by human employees.

    THE INITIAL LAYOFFS AND SCENARIO PLANNING
    In May, the first wave of layoffs stemming from Project OT occurred, impacting thousands of employees. Meta’s approach involved teams conducting scenario planning exercises, evaluating redeployment, open role closures, and potential cuts. The company ultimately didn’t implement all scenarios, acknowledging the complexity of transitioning to an AI-driven model.

    MARK ZUCKERBERG’S DIRECTIVE AND SHIFTING COURSE
    Meta CEO Mark Zuckerberg spearheaded Project OT, directing executives to proceed with the changes to management structures. Following the initial layoffs, Zuckerberg abruptly canceled a planned November round of layoffs, leaving the reasons for the reversal unclear.

    AI-NATIVE STRATEGY AND INITIAL TEAM RESTructuring
    Meta’s “AI-native” strategy envisioned a company where AI tools and agents would handle much of the daily work, with small teams overseeing these systems. This led to restructuring efforts across engineering, research, and at least eight other teams, implementing a pilot program focused on removing middle management and leveraging “agent-assisted analysis.”

    EMPLOYEE MONITORING AND MORALE IMPACT
    Leading up to the layoffs, Meta tracked employees' keyboard and mouse input to train AI agents, a practice that negatively impacted employee morale. Internal reports highlighted a significant increase in code changes and technical incidents, alongside a rise in problem resolution time.

    AI’S UNEXPECTED CHALLENGES AND PROJECT SCALBACK
    Zuckerberg acknowledged in July that the agentic development hadn’t accelerated as expected. Internal data revealed a disparity between increased internal tool usage and new feature development, alongside a rise in major incidents and a 70% increase in time spent resolving them. This prompted a scaling back of Project OT.

    AI-POWERED SALES AND THIRD-PARTY AGENTS
    Meta began monetizing its AI agents by selling them to third parties, starting in June. The company launched a pilot program restructuring teams to facilitate this strategy, utilizing “agent-assisted analysis” to prioritize tasks.

    HUMAN DECISION-MAKING REMAINS
    Despite the push for AI automation, Meta emphasized that performance ratings and promotion decisions remained the domain of human leadership, rejecting the complete automation of these processes.

    A CASE STUDY IN AI IMPLEMENTATION RISK
    Meta’s experience with Project OT serves as a cautionary tale for organizations exploring AI, highlighting the potential challenges in replacing human workloads and the risks of overly ambitious AI projects.

    Related Articles