Run NVIDIA Nemotron and OpenAI GPT OSS models on Amazon Bedrock in AWS GovCloud (US)
TL;DR
Amazon Bedrock now supports OpenAI GPT OSS 120B and 20B plus NVIDIA Nemotron 3 Nano 9B v2, 12B v2, 30B and Super 120B in AWS GovCloud (US). Inference stays inside the GovCloud boundary: single-region in us-gov-west-1 or GovCloud cross-region across us-gov-west-1 and us-gov-east-1. Global cross-region routing is not available there.
Nauti's Take
This is a very AWS-style announcement: lots of compliance language, many model names, little independent performance evidence. The practical point is still strong.
If agencies and contractors can test GPT OSS or Nemotron inside existing GovCloud controls, the path to real AI applications gets shorter. The open question is whether these models actually beat smaller, cheaper or already approved alternatives in the workflows that matter.
Briefingshow
For regulated US workloads, this is less a model announcement than an infrastructure move: open weights, OpenAI-compatible APIs and GovCloud residency land in a managed Bedrock setup. Teams get more model choice without moving sensitive prompts and outputs into commercial regions or self-managed GPU stacks.
Sources
