Meta Platforms shifts headcount to AI compute, using internal workflows as training data for superintelligence models, per leaked remarks.
Meta Platforms (NASDAQ:META) is reducing its workforce to fund artificial intelligence development, using employee output and internal workflows as proprietary training data for its superintelligence models. CEO Mark Zuckerberg outlined the strategy in a leaked April 30, 2026 all-hands meeting, stating that internal talent accelerates AI coding capabilities beyond industry peers.
The approach mirrors tactics likely used by competitors like Google and Amazon, though Meta is the first to publicly acknowledge it. Employees were reportedly sent home for “humanitarian reasons” before receiving termination notices, highlighting the company’s focus on operating leverage through reduced headcount.
Zuckerberg emphasized that Meta’s workforce possesses above-average intelligence, making their contributions critical to AI model performance. The shift underscores a broader industry trend of prioritizing compute over labor costs.