Deep dive into Azure Maia 300 architecture(azure.microsoft.com)
16 points by azhenley | 2 comments
Microsoft Azure blog post giving a technical deep dive into the architecture of the Azure Maia 300 AI accelerator chip.
#ai-accelerators on Hacker News
5 stories tagged ai-accelerators, newest first.
Our eighth generation TPUs: two chips for the agentic era(blog.google)
420 points by xnx | 209 comments
Google blog post announcing its eighth-generation TPUs, TPU 8t and TPU 8i, custom AI chips designed for training and agentic inference at scale.
Ironwood: The first Google TPU for the age of inference(blog.google)
432 points by meetpateltech | 159 comments
Google blog post announcing Ironwood, its seventh-generation TPU designed for large-scale AI inference workloads.
Google's First Tensor Processing Unit: Architecture(thechipletter.substack.com)
324 points by c_joly | 172 comments
A Chip Letter Substack post examining the architecture of Google's first Tensor Processing Unit (TPU v1), following an earlier post on its origins.
Nvidia releases new AI chip with 480GB CPU RAM, 96GB GPU RAM(nvidia.com)
324 points by logicchains | 256 comments
NVIDIA product page for the Grace Hopper Superchip, an accelerated CPU-GPU module for giant-scale AI and HPC applications.