What Happened to Meta's AI Reorganization and its Aftermath?

Meta Platforms initiated a sweeping reorganization of its artificial intelligence efforts starting in 2025, aiming to re-center the company around AI and accelerate its development in a highly competitive landscape. This strategic pivot involved a significant restructuring of its workforce, leading to thousands of layoffs and the mandatory reassignment of approximately 7,000 employees to new AI-focused units, such as Applied AI Engineering (AAI) and Agent Transformation Accelerator (ATA) XFN. The rationale behind this aggressive shift was to streamline development, flatten management structures, and embed AI across Meta's entire product ecosystem, from internal workflows to consumer-facing features.

However, the implementation of this reorganization faced considerable internal challenges, leading to what many employees and even Meta's leadership described as a 'backfire.' CTO Andrew Bosworth publicly admitted in internal communications that the company had done an 'atrocious job explaining the vision' and supporting employees through the transition. Reports indicated widespread discontent, low morale, and a feeling among some reassigned engineers that their new roles, often involving data labeling, were 'mind-numbing' or 'gulag-like.' Issues cited included overly large teams, a lack of personal attention from managers, and poor communication regarding career paths.

Key turning points included the appointment of Alexandr Wang as Chief AI Officer in June 2025, following Meta's substantial investment in his company, Scale AI. Wang was given broad control over Meta's AI operations, leading to the dissolution of the AGI Foundations group and the creation of four new core teams. Despite aggressive recruitment efforts, including poaching talent from rivals like OpenAI, Meta's Superintelligence Labs also experienced a wave of high-profile departures, indicating ongoing talent retention challenges. The company also faced setbacks with its AI models; while Llama models were released, they initially trailed competitors, and the highly anticipated 'Avocado' text model, planned for early 2026, was delayed due to performance issues compared to rivals like Google Gemini 3.0.

As of June 2026, Meta is actively working to address the consequences of the initial reorg. CEO Mark Zuckerberg and CTO Andrew Bosworth have acknowledged past mistakes and pledged to improve internal stability, with Zuckerberg promising no further company-wide layoffs for the remainder of 2026. Reforms include capping managers at approximately 20 direct reports and minimizing employee transfers during restructurings. Financially, Meta continues to make massive investments in AI infrastructure, with projected capital expenditure for 2026 reaching up to US$145 billion. The company has seen strong Q1 2026 financial results, with AI-driven improvements leading to increased engagement on platforms like Instagram Reels and a significant rise in AI-translated video consumption. Meta launched new, cheaper AI smart glasses on June 24, 2026, powered by its Muse Spark model, and is developing an internal prediction market app called 'Arena,' signaling a continued push into consumer-facing AI and new ventures. While still striving to match the cutting-edge performance of some rival AI models, Meta's strategy is focused on leveraging its vast user base and integrating 'good enough' AI deeply into its core products and new hardware.