AWS M8i lifts AI recommendations 40%

AWS says its M8i and M8i-flex instances deliver up to 40% better performance for AI deep-learning recommendation models than M7i equivalents, with up to 15% better price-performance and 2.5x the memory bandwidth, according to the EC2 availability announcement. The Intel Xeon 6 instances are now available in the Canada West region, so teams serving memory-heavy ranking or retrieval workloads should benchmark their own batch size and utilization before paying for GPUs by habit, alongside the archive’s analysis of AI inference power economics.