NVIDIA's next-gen 'Feynman' GPUs to be built with TSMC's 1.6 nm A16 process - TweakTown

NVIDIA is planning to launch its next-generation Feynman GPUs in the second half of 2028, utilizing TSMC's cutting-edge A16 1.6nm process. This move skips TSMC's N2 or 2nm nodes, representing a signif... Read original →

NVIDIA’s upcoming Feynman GPU, set for 2028, will leverage TSMC’s 1.6nm A16 process, bypassing the N2 node and signaling a major leap from current Rubin architectures. This move catalyzes adoption of 3D chiplets, SoIC, and backside power delivery, pushing upstream suppliers to accelerate innovation. From a compliance standpoint, the strategy reduces geopolitical exposure but increases reliance on a single foundry, raising supply chain vulnerability. Competitors like AMD and Intel may expedite their own architectures to counter NVIDIA’s performance edge. In the longer term, this development accelerates the data center shift toward ultra-high bandwidth, with HBM and CPO becoming industry standards, reshaping the AI infrastructure landscape.