Meta sets AI coding targets, with some teams aiming for 75% usage

Meta has set ambitious internal targets for AI-assisted coding, with some engineering teams expected to generate as much as 75% of their code using artificial intelligence tools, signalling a deeper shift towards AI-led workflows.

Internal documents reviewed by Business Insider show that the targets vary across teams but form part of a broader push to embed AI into everyday engineering work. The move comes as the company restructures parts of its business and reduces headcount.

TARGETS VARY ACROSS TEAMS

In Meta’s creation organisation—responsible for core platforms such as Facebook, WhatsApp and Messenger—65% of engineers are expected to write more than 75% of their code using AI tools in the first half of 2026, according to Business Insider.

The Scalable Machine Learning division, which builds AI infrastructure, has set targets ranging from 50% to 80% AI-assisted coding. Earlier, a company-wide goal for late 2025 required 55% of code changes in central product teams to be supported by AI agents.

The company is also pushing for wider adoption of tools. Around 80% of mid-to-senior engineers are expected to use AI platforms such as Gemini, DevMate and Metamate, with the focus placed on adoption and impact rather than just output levels.

AI AS A PERFORMANCE EXPECTATION

While it remains unclear whether these targets are directly tied to formal performance reviews, internal communication suggests that AI usage is becoming a core expectation.

Meta’s head of people, Janelle Gale, told employees in a memo last year that “AI-driven impact” would become central to performance from 2026. Employees have also been encouraged to highlight their use of AI tools in self-assessments, supported by an internal “AI Performance Assistant”.

A Meta spokesperson told Business Insider that performance systems are designed to reward outcomes driven by AI, rather than usage alone.

RESTRUCTURING AND WORKPLACE SHIFT

The coding targets sit alongside a broader organisational overhaul. A separate internal memo, cited by Business Insider, showed that a 1,000-person team within Reality Labs has been reorganised into smaller, cross-functional “pods”.

Employees in these teams have been assigned new roles such as AI Builder, AI Pod Lead and AI Org Lead, reflecting a shift towards flatter structures and AI-led collaboration. Engineers may now take on hybrid responsibilities, including design work.

Meta’s Chief Technology Officer, Andrew Bosworth, is also overseeing the company’s “AI for Work” initiative, aimed at scaling internal adoption of AI tools.

LAYOFFS AND PRODUCTIVITY PRESSURE

The push towards AI comes amid recent job cuts across multiple divisions, including Reality Labs and recruitment. While the company has said the layoffs are not directly linked to its AI strategy, the timing underscores a broader shift towards leaner teams supported by automation.

Chief executive Mark Zuckerberg has previously indicated that advances in AI could allow smaller teams—or even individuals—to deliver work that once required larger groups.