Meta reorganises engineers around AI

Meta has started moving some of its strongest software engineers into a new artificial intelligence engineering group, and some engineers were reassigned into applied artificial intelligence teams even when they did not ask for the move. Reuters reported the reshuffle on April 9, and The Information said the goal was to improve Meta’s models faster and compete more directly with OpenAI, Google, and Anthropic. (reuters.com, theinformation.com) This is not a side project inside Facebook, Instagram, and WhatsApp owner Meta. The company has spent the past year rebuilding its artificial intelligence operation after Mark Zuckerberg decided Meta’s earlier model progress was not good enough. (reuters.com, cnbc.com) In June 2025, Meta put $14.3 billion into Scale AI and brought in Scale AI chief executive Alexandr Wang to help lead the push. By April 8, 2026, Meta was already presenting Muse Spark as the first major model from its rebuilt artificial intelligence effort. (cnbc.com, techcrunch.com) Meta’s old pitch on artificial intelligence was that open models like Llama would spread widely and pull developers into its ecosystem. Muse Spark broke with that pattern by launching as a closed model, which Bloomberg described as a shift away from Meta’s earlier open-source strategy. (bloomberg.com, cnbc.com) That helps explain why engineers are being moved around like emergency staff in a hospital. When a company changes both the people building the models and the rules for shipping them, it is usually trying to cut time between research, product, and release. (reuters.com, theinformation.com, bloomberg.com) Inside Meta, the push did not stop at staffing charts. Fortune reported that the company briefly built an internal dashboard that tracked how many artificial intelligence tokens employees were using, after leaders discussed treating artificial intelligence use as a productivity metric. (fortune.com) A token is a small chunk of text that an artificial intelligence system reads or writes, so token counts are a rough meter for how much employees are using chatbots and coding assistants. Measuring workers by token consumption is a little like judging a carpenter by how many nails came out of the box instead of whether the house stands up. (fortune.com) Meta later killed that dashboard, but the fact that it existed shows how far the company is willing to go to make artificial intelligence part of ordinary engineering work, not just a research lab project. Reuters and The Information describe the staffing changes the same way: artificial intelligence work is becoming the center of gravity for the company’s technical organization. (fortune.com, reuters.com, theinformation.com) This also fits Meta’s spending pattern outside the org chart. The Information reported on April 9 that Meta had expanded cloud deals with CoreWeave to $35 billion through 2032, after earlier chip rental commitments with Nebius, which means the company is buying not just talent but enormous amounts of computing power. (theinformation.com) Put together, the picture is simple: new leadership, new model strategy, new internal metrics, new compute contracts, and engineers pulled toward one priority. Meta is reorganizing itself so artificial intelligence is not one product line among many, but the thing the rest of the company is expected to orbit. (reuters.com, fortune.com, theinformation.com, cnbc.com)