AMD allegedly plans to shift low-end APU and Radeon GPU production to one of Samsung’s 4nm nodes. According to @Tech_Reve, Team Red will diversify its supply chain, offloading the manufacturing of its low-end Ryzen APUs (and Athlon?) to a cheaper Samsung node. This strategy will first be employed with entry-level desktop and notebook processors and will slowly be expanded to Radeon GPUs.
Following its split with Global Foundries, AMD has been sorely exclusively relying on TSMC for its CPU and GPU production. Lately, however, the capacity of advanced nodes has been limited at the Taiwanese foundry, with Apple being the sole 3nm client (for now).
AMD’s shift in policy to a dual foundry system is likely an attempt to carve out additional production capacity as competition at TSMC intensifies. Samsung offers several variants of its 4nm node, including SF4E, SF4, SF4P, and soon SF4X and SF4A. The SF4P or SF4A node will probably be used for the production of Ryzen/Athlon APUs and Radeon GPUs. If this plan works out, it will strengthen AMD’s supply chain and widen its profit margins.
