AWS AI Updates: July 11, 2026

1. Amazon EMR on EKS Gains a Generative-AI Spark Troubleshooting Agent

AWS. Amazon EMR on EKS added an Apache Spark troubleshooting agent that lets data engineers diagnose job failures in natural language, returning automated root cause analysis and PySpark code recommendations by analyzing Spark History Server data and distributed logs. The agent is available directly from the EMR on EKS console through a “Troubleshoot with AI” option, and it can also be wired into compatible AI coding agents such as Claude Code through the Model Context Protocol. This removes much of the manual log-navigation work that typically slows down debugging failed Spark jobs. Source

2. Amazon EC2 G7 GPU Instances Arrive in US East (N. Virginia)

AWS. Amazon EC2 G7 instances, built on NVIDIA RTX PRO 4500 Blackwell GPUs, are now available in the US East (N. Virginia) Region, delivering up to roughly 4.6x AI inference performance over the prior generation for workloads such as language models, video analysis, and graphics rendering. The instances target teams running inference and mixed AI-plus-graphics workloads that want newer Blackwell-class GPUs without moving to the largest accelerator tiers. Source