Latest about HBM3E
SK Hynix says new high bandwidth memory for GPUs on track for 2024 - HBM4 with 2048-bit interface and 1.5TB/s per stack is on the way
HBM4 memory with a 2048-bit interface on track for production in 2026.
China's CXMT reportedly aims to make HBM memory for AI chips
CXMT to compete against Micron, Samsung, and SK Hynix for lucrative HBM memory market, says report.
Nvidia reportedly races to secure memory supply for next-gen H200 AI GPUs
Nvidia allegedly pre-purchases HBM3E memory from Micron and SK Hynix, says report.
AWS and Nvidia build a supercomputer with 16,384 Superchips
Amazon Web Services to offer Nvidia-powered supercomputing infrastructure for generative AI, including 65 ExaFLOPS AI supercomputer.
Nvidia's Grace Hopper GH200 Powers 1 ExaFLOPS Jupiter Supercomputer
Nvidia's Grace Hopper now powers 40 AI supercomputers globally.
21 ExaFLOP Isambard-AI Supercomputer Uses 5,448 GH200 Grace Hopper Superchips
U.K.'s Isambard AI supercomputer for AI and HPC workloads boasts 200 PetaFLOPS for HPC and 21 ExaFLOPS for AI workloads.
Samsung Expects HBM4 Memory to Arrive by 2025
Samsung starts talking about HBM4, next-generation memory for AI and HPC GPUs.