According to AWS, the new G7 family promises gains over the G6 generation and is already available in two U.S. regions.
AWS announced the general availability of Amazon EC2 G7 instances, designed for AI inference, graphics, and data analytics workloads. According to the company, the new family is powered by Nvidia RTX PRO 4500 Blackwell Server Edition GPUs and succeeds the G6 line in scenarios requiring graphics computing and parallel processing.
According to AWS, G7 instances deliver up to 4.6 times higher performance in AI inference and up to 2.1 times higher performance in graphics compared to G6. The company cites machine translation, video and image analysis, speech recognition, recommendation systems, real-time rendering, game streaming, and large-scale data processing pipelines as use cases.
The configuration detailed by AWS includes up to eight Nvidia RTX PRO 4500 Blackwell Server Edition GPUs, each with 32 GB of memory, alongside custom Intel Xeon 6 processors. The company also states that the instances can offer up to 700 Gbps of network bandwidth via the Elastic Fabric Adapter.
According to AWS, EC2 G7 instances are now available in the US East (Ohio) and US West (Oregon) regions. They can be purchased through On-Demand, Savings Plans, and Spot Instances models, with access via the AWS Management Console, CLI, and the company's SDKs.
Amazon EC2 G7 instances are designed for AI inference, graphics, and data analytics workloads, including machine translation, video analysis, speech recognition, real-time rendering, and game streaming.
According to AWS, G7 instances deliver up to 4.6 times higher performance in AI inference and up to 2.1 times higher performance in graphics compared to the previous G6 generation.
EC2 G7 instances are powered by up to eight Nvidia RTX PRO 4500 Blackwell Server Edition GPUs, each featuring 32 GB of memory, paired with custom Intel Xeon 6 processors.