G6 expansion to SageMaker Studio notebooks in Dubai and Malaysia — pricing implications

AI Impact Summary

Amazon announces general availability of EC2 G6 instances in the Middle East (Dubai) and Asia Pacific (Malaysia) for SageMaker Studio notebooks. G6 instances pack up to 8 NVIDIA L4 Tensor Core GPUs with 24 GB per GPU and 3rd-gen AMD EPYC CPUs, delivering roughly 2x the DL inference performance of G4dn. This enables interactive model deployment testing and training for workloads such as generative AI fine-tuning, NLP, translation, computer vision, and recommender systems in the new regions; teams should plan provisioning and cost in those regions and compare to existing G4dn usage. Expect pricing considerations to be surfaced on the pricing page and incorporate in regional budgeting and capacity planning.