Amazon EC2 G7 GPU instances expand to four more AWS Regions
AWS brought EC2 G7 instances to London, Mumbai, Sydney and Sao Paulo. G7 runs on NVIDIA RTX PRO 4500 Blackwell GPUs and offers up to 4.6 times the AI inference performance of G6.
By JetherVerse News

AWS announced on November 10, 2026 the arrival of Amazon EC2 G7 instances in four more Regions. The new Regions are Europe (London), Asia Pacific (Mumbai), Asia Pacific (Sydney) and South America (Sao Paulo). G7 instances run on NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs.
AWS says G7 delivers up to 4.6 times the AI inference performance of G6 and up to 2.1 times the graphics performance. GPU accelerated data analytics also run faster. The largest configuration offers 8 GPUs with 32 GB of memory each, custom Intel Xeon 6 processors, up to 192 vCPUs and up to 768 GiB of system memory. Networking reaches 700 Gbps through Elastic Fabric Adapter. Local NVMe SSD storage reaches 7.6 TB. AWS lists maximums only, so check the console for individual sizes.
G7 is now available in US West (Oregon), US East (N. Virginia and Ohio), Europe (London, Spain, Frankfurt and Stockholm), Asia Pacific (Mumbai and Sydney) and South America (Sao Paulo). You buy G7 through On-Demand Instances, Spot Instances and Savings Plans.
AWS points to three groups of workloads. The first is AI model deployment, including language translation, video and image analysis, speech recognition and recommender systems. The second is graphics work, such as real time cinematic quality rendering and game streaming. The third covers video transcoding, spatial computing and data analytics.
For teams serving AI features, Region choice matters. Placing inference close to your users cuts latency and keeps data in the Region your contracts require. London and Frankfurt serve UK and EU clients. Oregon and Virginia serve US clients. Mumbai and Sao Paulo bring GPU capacity closer to users in South Asia and South America. Compare Spot and On-Demand prices in each Region before you pick one, and test your model on a single instance before you plan a fleet.
If you host AI features for clients in the UK and US, GPU Region coverage now belongs in your architecture checklist.
Source: Amazon Web Services
Get the news in your inbox
The tech stories that matter from JetherVerse News, with the source on every one.

