Foxconn is making waves in the tech world with its latest orders for AI clusters centered around Google’s TPUs. This development marks a pivotal moment for the Taiwanese company as it expands its partnership with leading tech giants.
Foxconn’s Dual Role with NVIDIA and Google
The tech industry is abuzz with the growing prominence of ASICs, especially following the release of Google’s Ironwood TPU platform. Reports suggest that Google’s TPUs are on the cusp of wide adoption, with Meta among the potential users. Foxconn has reportedly received orders for Google’s TPU compute trays and will work together on Google’s ‘Intrinsic’ robotics projects.
Industry insiders say that Google’s AI servers, built with its self-developed TPUs, are mainly divided into two racks: one rack for TPUs and the other for computing trays. According to Google’s proposed supply chain, for every rack of TPUs shipped, Foxconn will ship one rack of computing trays, resulting in a 1:1 supply ratio.
– Taiwan Economic Daily
The High-Powered Superpod Infrastructure
Google’s seventh-generation TPUs offer more than just chip configurations; they boast a scalable ‘rack’ infrastructure known as the ‘Superpod’. Each pod can house 9,216 chips, delivering a staggering 42.5 exaFLOPS for FP8 compute workloads. These TPUs utilize a 3D Torus layout, enabling high-density interconnectivity among numerous chips. Reports indicate that Foxconn will manufacture computing trays based on the orders it receives for Google’s TPU racks.
The Future of AI Workloads
With AI inferencing becoming increasingly dominant, corporations are eager to optimize their computing resources for superior performance while maintaining cost-effectiveness. Google’s TPUs are emerging as front-runners in AI applications. This shift fuels discussions about whether NVIDIA could be challenged by bespoke silicon from tech giants, but for now, the supply chain shows significant demand for Google’s TPU solutions.