United StatesMeta is preparing to begin mass production of its custom AI chips in September, aiming to reduce reliance on external GPU suppliers like NVIDIA and AMD. According to Reuters, the company has successfully tested at least one version of the chip within six weeks and is manufacturing them via TSMC while sourcing components from Samsung, Sandisk, and Sumitomo Electric. These chips, part of Meta's MTIA program, are designed for training large AI models and improving efficiency in recommendation systems and other AI tasks. The move comes as Meta continues to invest heavily in AI infrastructure, planning to deploy 7 gigawatts of compute this year and double that next. The company has also secured partnerships with ARM, AMD, and Amazon for additional compute resources. Other companies including OpenAI, Anthropic, and Google are similarly pursuing in-house AI chip development.
Bias read (Center): The article presents a factual update on Meta's technological strategy without overtly favoring any political ideology. While the development of AI technology can have economic and strategic implications, the focus remains on corporate operations and industry trends rather than ideological stances.
