Season 5 Episode 29: EC2 Turns 20, Leaked AWS Keys, and the AWS-Nvidia GPU Deal

Season 5, Episode 29 of the LogiCast AWS News podcast brings together hosts Karl Robinson, CEO and co-founder of Logicata, and Jon Goodall, principal cloud engineer at Logicata, with guest Khushi Nandwani, a recent graduate and AWS Community Builder in the containers category. They discuss EC2’s 20th anniversary, a new cost anomaly detection feature for third-party models, the AWS Agent Toolkit for the CLI, a report uncovering hundreds of leaked AWS keys, and the expanded AWS - Nvidia GPU partnership. Along the way, the conversation somehow veers into the plight of PC gamers, cloud-gaming latency, and Karl’s Mahjong addiction - before everyone remembers they are supposed to be talking about Nvidia GPUs.

 

03:29 - Amazon EC2 Turns 20 

The hosts celebrate EC2’s 20th anniversary and reflect on its evolution from a single instance type into a global compute platform. They highlight innovations such as Graviton and Auto Scaling and discuss EC2’s role in popularizing pay-as-you-go cloud computing.

11:07 - AWS Cost Anomaly Detection and Third-Party Bedrock Models 

AWS Cost Anomaly Detection can now identify unusual spending on third-party models available through Amazon Bedrock. The discussion explores the growing importance of monitoring AI costs and using carefully chosen thresholds to prevent alert fatigue.

15:41 - AWS Agent Toolkit for the AWS CLI 

The AWS Agent Toolkit gives coding agents access to current information about AWS services instead of relying entirely on potentially outdated training data. The hosts discuss its significance as AI agents begin interacting directly with cloud infrastructure.

21:01 - Hundreds of Leaked AWS Keys Exposed 

A security report uncovered hundreds of exposed AWS credentials, including long-lived root and administrator keys. The conversation highlights credential rotation, least-privilege access, IAM roles, secret scanning, and maintaining accurate security contacts.

31:41 - AWS and Nvidia: Two Million Additional GPUs 

AWS and Nvidia plan to deploy two million additional GPUs between 2027 and 2028 to support agentic and physical AI workloads. The hosts consider growing AI demand, manufacturing capacity, government projects, consumer GPU availability, and AWS’s investment in its own AI chips.