Meta Rewrites AI Bets Again: Zuckerberg Uses Claude to Write Code, Employees Kick Off a Token Consumption Battle to Hit KPIs

Meta is doubling down on AI, with Zuckerberg personally using Claude Code to write code. To hit KPIs, an internal Token consumption frenzy has reportedly erupted, wasting resources. Learning from the $80 billion loss lesson from the metaverse, Meta is actively acquiring startups, hoping to turn the technology into real, measurable value.

Meta doubles down on AI: Zuckerberg writes code using Claude Code

Meta, the tech giant that owns social media platforms like Facebook, Instagram, and Threads, is currently pouring the company’s resources entirely into the generative AI space.

Meta CEO Mark Zuckerberg has begun personally using an AI coding tool, Claude Code, to write code, breaking a record of many years during which he did not take part in hands-on development.

External reports also claim that Meta recently sparked an internal contest to burn through Tokens, with many engineers consuming massive numbers of Tokens to boost their personal performance metrics (KPI).

AI coding is all the rage, founders return to the development front

In March 2026, Zuckerberg submitted three code diffs to Meta’s single repository—his first substantive code contribution in 20 years.

He is using the Claude Code CLI, a coding assistant terminal developed by Anthropic. In one of his submissions, it received approvals from more than 200 engineers.

His actions reflect how AI coding tools are attracting corporate founders to return to system development. Y Combinator CEO Garry Tan also returned to coding work after 15 years and open-sourced a system that combines Claude Code.

According to internal documents from Meta that leaked in March 2026, the company set ambitious goals, planning that by mid-2026, 65% of engineers would use AI to write 75% or more of their code.

Image source: flickr, photographed by Niall KennedyMeta founder Zuckerberg speaking at the Facebook F8 developer conference in September 2011

Meta holds a Token consumption contest internally, turning KPIs into a show

To drive adoption of generative AI applications, Meta has seen a phenomenon internally that links Token usage to productivity. A Token is the smallest unit of text processing by large language models; in Chinese, it is often called “symbol units” or “tokens.”

The report from 《The Information》 reveals that Meta internally has created a leaderboard called Claudeonomics, tracking artificial intelligence token consumption for more than 85k employees. The data shows that employees consumed as many as 60 trillion tokens in just 30 days—putting them at the top. The average consumption of the top user was 281 billion tokens.

The leaderboard awards titles such as Token Legend, encouraging employees to integrate AI tools into their daily work.

A report by 《Forbes》 notes that Meta’s CTO Andrew Bosworth previously mentioned that a top engineer consumed a token amount equivalent to their annual salary. Nvidia CEO Jensen Huang has also said that if an engineer making $500k in salary fails to consume tokens worth $250k, he would be concerned.

However, this KPI system that drives “Token consumption competition” also brings drawbacks. Some Meta employees, to push up performance numbers, leave AI agent programs running idle for hours, wasting computing resources.

Moreover, treating employees’ Token consumption directly as a productivity metric turns consumption behavior into a performance, creating challenges for performance reviews that lack support from real business outcomes.

Learning from the metaverse setbacks, Meta’s next challenge in the AI race

Before going all-in on AI, Meta’s metaverse bet ended in failure. The company previously poured roughly $8 billion into building the virtual world Horizon Worlds and VR/MR devices, and even changed the company name to “Meta”—yet it still failed to reach the market-expected scale of users.

When Lily Liu, president of the Solana Foundation, commented in a discussion thread on a social platform about blockchain games and metaverse development, she also expressed a pessimistic view of virtual economy models that previously lacked real, substantive content support.

Image source: Meta Meta metaverse platform Horizon Worlds. In the original version shown, Zuckerberg’s virtual likeness

Now that Meta is shifting its focus to AI, it is actively laying out its market strategy. In addition to launching its own large language model, LLaMA, it is also gradually advancing an AI model initiative called “Avocado.”

Recently, a report from 《Axios》 also revealed that Meta has acquired Moltbook, an agent community often dubbed the “AI version of Reddit.” Moltbook’s founders, Matt Schlicht and Ben Parr, will join Meta’s team.

• **Related report:**Lobster community bought! Rumor: Meta acquires Moltbook, and the founder successfully switched careers from media to AI

The outside world is also watching to see whether Meta can avoid repeating the metaverse’s pattern of overinvesting without real applications—turning the current internal Token consumption frenzy and acquisition deals like Moltbook into actual products with commercial value, so it can stay above in the highly competitive generative AI market.