Meta's AI Overhaul: Zuckerberg Mandates Code Restructuring for AI Agents
Meta Restructures Code for AI Agents Under Zuckerberg Directive

Meta's AI Overhaul: Zuckerberg Mandates Code Restructuring for AI Agents

In a bold strategic move, Meta CEO Mark Zuckerberg has issued a directive requiring the company's engineering teams to undertake comprehensive restructuring of entire codebases. This initiative is not aimed at fixing bugs or shipping new features, but rather to enable artificial intelligence agents to autonomously read, navigate, and modify the code themselves.

According to an exclusive report by The Information, this internal mandate has been described as a "bold ask" that underscores Zuckerberg's commitment to pushing an AI-native vision throughout the organization. The directive represents a significant shift in how Meta approaches software development and engineering workflows.

AI Adoption Becomes Mandatory at Meta

The code restructuring directive fits into a broader pattern of mandatory AI adoption at Meta. The company has established explicit targets for how much engineers should rely on AI tools in their daily work. Meta has also reorganized teams around small, focused "pods" dedicated to AI development and implementation.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

CTO Andrew Bosworth has recently taken charge of Meta's internal "AI for Work" initiative, according to Business Insider. This move signals the company's serious commitment to integrating AI across all operational aspects.

Meta Tracks AI Usage Down to the Token Level

The AI push extends far beyond code restructuring. The Information reports that Meta now maintains an internal dashboard that meticulously logs token consumption across different employee roles. Tokens represent the amount of data processed through AI tools, providing a quantitative measure of AI engagement.

During a recent 30-day period, total token usage reportedly exceeded 60 trillion tokens across the organization. Using Anthropic's public pricing as a benchmark, this volume of usage could translate to approximately $900 million in monthly costs. While Meta's actual expenses likely differ due to internal infrastructure and custom pricing agreements, the sheer scale reveals the company's substantial investment in AI technologies.

Internal Competition and Cultural Shifts

Employees have transformed token usage into a competitive activity within the company. An internal leaderboard called "Claudeonomics"—a reference to Anthropic's Claude models—ranks the top 250 users among more than 85,000 employees. This system awards badges such as "Token Legend" and "Session Immortal" to high-performing users.

The cultural transformation runs deeper than mere dashboards and competitions. Business Insider reported that a 1,000-person team within Reality Labs has been restructured into AI-native pods, with employees receiving titles like "AI Builder" and "AI Pod Lead." This reorganization aims to flatten Meta's traditional hierarchy and encourage cross-functional collaboration, potentially enabling engineers to take on design roles when project requirements demand it.

AI Transformation Weeks and Ambitious Goals

Meta has been conducting intensive "AI Transformation Weeks" where employees across various roles participate in hackathons, watch demonstrations, and build with tools like Anthropic's Claude Code. An internal document reviewed by Business Insider revealed that Meta's creation organization has set a first-half 2026 goal for 65% of engineers to write more than 75% of their committed code using AI assistance.

Questions About Effectiveness and Metrics

Not everyone within Meta is convinced that this approach is yielding optimal results. Some employees have reportedly extended AI sessions or executed unnecessary tasks specifically to boost their visible token usage—a behavior that prioritizes engagement metrics over genuine productivity and output quality.

While Meta has not formally linked token consumption to performance reviews, the pressure to demonstrate AI fluency is palpable throughout the organization. As one employee told The Information, there exists a meaningful distinction between using AI effectively and simply using large quantities of it.

Pickt after-article banner — collaborative shopping lists app with family illustration

The company's aggressive push toward AI integration represents a fundamental transformation in how work gets done at one of the world's largest technology companies. As Meta continues to restructure its technical foundations and organizational culture around artificial intelligence, the tech industry watches closely to see whether this bold strategy will deliver the promised efficiency gains and innovation breakthroughs.