Meta, the parent company of Facebook, Instagram, and WhatsApp, has begun testing its first proprietary chip designed specifically for training artificial intelligence systems. This strategic move aims to reduce reliance on external semiconductor suppliers like Nvidia while addressing soaring operational costs, according to exclusive Reuters reporting.
A Long-Term Play for AI Independence
Two sources familiar with the project revealed that Meta initiated small-scale deployment of the chip, with plans for broader implementation contingent on successful testing. The development aligns with the company’s ambitious goal to curb its $114-119 billion projected 2025 expenses, which include up to $65 billion in AI infrastructure spending.
Technical Edge Over Conventional Hardware
The new chip operates as a dedicated accelerator optimized solely for AI computation tasks—a design choice experts say could significantly improve power efficiency compared to general-purpose GPUs traditionally used for such workloads. This technical pivot arrives as Meta intensifies investments in AI tools to drive future growth across its platforms.
