Meta Aims to Double AI Computing Power by Next Year with New 'Iris' Chips
Meta Platforms is reportedly planning to begin production of its custom "Iris" artificial intelligence chip in September. This move is part of a larger strategy to significantly expand its computing infrastructure. The company aims to double its AI computing capacity to 14 gigawatts by next year. For the current year, Meta intends to deploy 7 gigawatts of computing infrastructure. To support this expansion, Meta has secured long-term supply agreements with several key partners. These agreements include deals with Samsung Electronics for memory chips and Sandisk for flash storage. Additionally, Meta has established a protocol with Sumitomo Electric Industries for fiber optic equipment. These initiatives underscore Meta's commitment to advancing its AI capabilities and supporting its growing demands for computational power.
Meta's strategic investment in custom AI silicon, exemplified by the "Iris" chip and a planned doubling of computing capacity, reflects a broader industry trend toward in-house hardware development to optimize performance and cost for AI workloads. This vertical integration strategy aims to reduce reliance on external chipmakers and tailor hardware precisely to Meta's specific AI model requirements, potentially offering a competitive advantage. The multi-year supply agreements signal a long-term commitment and a proactive approach to securing necessary components, mitigating risks associated with supply chain volatility. As AI capabilities become increasingly central to platform functionality and future innovation, such infrastructure investments are critical for maintaining technological leadership and driving future revenue streams.
AI-generated to prompt reflection — not editorial opinion, not advice, not a statement of fact. How this works.