Micron reportedly planning HBM-style GDDR stack for future accelerators [might be big for local compute/LLM]

r/singularity
Generative AI

Submitted by /u/Mochila-Mochila [link] [comments]