Meta announces the building of two 24k GPU clusters for AI workloads, and shares details on the hardware, network, storage, design, performance, and software of these clusters. Meta is committed to open compute and open source, and aims to grow its infrastructure build-out with 350,000 NVIDIA H100 GPUs by the end of 2024.
Table of contents
A peek into Meta’s large-scale AI clustersUnder the hoodCommitment to open AI innovationThe future of Meta’s AI infrastructureSort: