Meta reportedly invests $6.5 billion with Samsung Foundry to produce third-generation MTIA AI chips on a 2nm process, reducing reliance on TSMC and NVIDIA.
Meta Platforms is reportedly investing $6.5 billion with Samsung Foundry to produce its third-generation MTIA (Meta Training and Inference Accelerator) chips using a 2nm process. This strategic move signifies a shift from TSMC and aims to reduce reliance on NVIDIA GPUs, lower supply chain risks, and support Meta's ambitious goal of 5 gigawatts of computing capacity by 2030 for its AI and cloud initiatives. The deal is expected to bolster Meta's competitive position in the rapidly evolving AI and cloud computing markets. Context Meta has been increasingly focused on artificial intelligence and cloud services, necessitating advanced computing power. The MTIA chips represent Meta's third generation of in-house processors, designed to optimize performance for AI workloads. The shift to Samsung Foundry marks a strategic pivot in Meta's manufacturing partnerships, reflecting broader industry trends towards vertical integration. Why this matters Meta's $6.5 billion investment in Samsung Foundry is a significant step towards enhancing its capabilities in AI and cloud computing. By developing its own 2nm chips, Meta aims to reduce dependence on external suppliers like TSMC and NVIDIA. This move could improve supply chain stability and operational efficiency, which are critical in the fast-paced tech landscape. Implications This deal could enhance Meta's market position by enabling it to deliver more efficient AI services. It may also influence other tech companies to reconsider their supply chains and partnerships in light of Meta's strategic shift. If successful, this initiative could lead to increased investment in domestic semiconductor manufacturing and innovation within the tech sector. What to watch In the coming months, observers should monitor the progress of the chip development and production timelines. Any announcements regarding partnerships or technological advancements from Meta or Samsung could signal the effectiveness of this collaboration. Additionally, industry reactions from competitors and suppliers will provide insights into the competitive landscape.
Meta's custom AI chips (MTIA) will begin production in September, aiming to reduce GPU costs. The chips are designed with Broadcom and manufactured by TSMC, part of Meta's strategy to secure compute capacity while still purchasing from Nvidia and AMD.
The article discusses Meta's strategy of deploying four MTIA chips in two years to scale AI experiences for billions of users while managing infrastructure costs.
Microsoft is in talks to supply its custom Maia 200 AI chips to Anthropic following a $5 billion investment, as Anthropic seeks to address its growing compute needs amid rising popularity of its Claude assistant and Claude Code tool.