Meta doubles down on AI re-stakes: Zuckerberg uses Claude to write code, and employees launch a Token consumption battle to hit their KPIs

Meta is doubling down on AI, with Zuckerberg personally using Claude Code to write code. In pursuit of KPIs, an internal token consumption free-for-all has emerged, wasting resources. Learning from the lesson of the metaverse’s $80 billion loss, Meta is actively acquiring startups, hoping to turn its technology into real value.

Meta is doubling down on AI: Zuckerberg writes code with Claude Code

Meta, the tech giant that owns social platforms such as Facebook, Instagram, and Threads, is currently pouring all of the company’s resources into the field of generative AI.

Meta CEO Mark Zuckerberg has begun personally using an AI coding tool, Claude Code, to write code, breaking his record of not taking part in direct development for many years.

External media also report that Meta has recently kicked off a token-consuming contest internally—many engineers boost their individual performance indicator (KPI) metrics by burning through large quantities of tokens.

Coding with AI is all the rage—founders return to the development front line

In March 2026, Zuckerberg submitted three code diffs to Meta’s single repository—this is his first substantive code contribution in two decades.

The terminal coding assistant he used is Claude Code CLI, 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 company founders to re-engage with system development. Y Combinator CEO Garry Tan has also returned to code writing after 15 years and open-sourced a system that combines Claude Code.

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

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

Meta launches a token consumption contest internally—KPIs become mere theater

To drive the adoption of generative AI applications, Meta has also spawned a practice inside the company linking token usage to productivity. Tokens are the smallest unit that large language models use to process text; in Chinese, they’re often referred to as “syllable units” or “tokens.”

The report from 《The Information》 reveals that Meta has an internal leaderboard called Claudeonomics, tracking the AI token consumption of more than 85k employees. The data shows that employees consumed as many as 60 trillion tokens within just 30 days, with the top users averaging 281 billion tokens.

The leaderboard assigns titles such as Token Legend to encourage employees to integrate AI tools into their daily work.

A report by 《Forbes》 notes that Meta’s Chief Technology Officer 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 earning $500k fails to consume tokens worth $250k, he would be concerned.

However, this KPI system that drives token consumption into a competitive frenzy also brings downsides. Some Meta employees, in order to push their performance numbers higher, leave AI agent programs running for hours when they’re not actually needed, wasting compute resources.

Moreover, treating employees’ token consumption directly as a productivity metric makes token-burning behavior turn into performance. This puts performance reviews in a bind, lacking challenges supported by real business outcomes.

Learning from the metaverse’s failure experience—Meta’s next challenge in the AI race

Before making a large-scale bet on AI, Meta’s previous metaverse push ended in failure. The company once poured roughly $8 billion into building the virtual worlds Horizon Worlds and VR/MR devices, even changing the company name to “Meta,” yet it still failed to meet market expectations for user scale.

Solana Foundation president Lily Liu, in the comment section on a community platform, also expressed a pessimistic view about the virtual economy model tied to blockchain games and the metaverse, saying it previously lacked support from real substance.

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

Now that Meta has shifted 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.”

Recent reporting by 《Axios》 also reveals that Meta has acquired Moltbook, a community agent platform touted as the “AI version of Reddit.” Moltbook founders Matt Schlicht and Ben Parr will join Meta’s team.

• Related report:**“The lobster community has been bought!” It’s rumored that Meta acquired Moltbook—founders successfully transitioned from media to AI

The outside world is also watching whether Meta can avoid repeating the metaverse’s mistake of over-investing while lacking real applications, and instead convert the current internal token consumption frenzy and acquisition deals like Moltbook into actual products with commercial value—so it can secure a foothold in the fiercely competitive generative AI market.