AWS says Amazon Bedrock Managed Agents powered by OpenAI are now in public preview, letting customers build agents using OpenAI models inside AWS. The practical draw is that those agents can use AWS resources while remaining within the identities, permissions and governance controls customers already use.
AWS AI Watch analysis
What happened
AWS says the managed agents run on a customised version of OpenAI’s Agents API, adapted for AWS. Customers can choose self-hosted compute, such as an existing development machine or container, or Amazon Bedrock AgentCore Runtime, AWS’s managed runtime option with configurable storage in the customer’s account. The announcement appeared in AWS’s weekly roundup, which says the public preview was announced last week.
Why it matters
This gives teams building with OpenAI models a route to run agents within AWS’s environment, rather than treating the model provider and cloud infrastructure as separate choices. Keeping existing AWS identities and permissions in the picture may make the setup more familiar to organisations already operating there. The two execution options also matter: teams can use their own compute or choose AWS-managed runtime sessions.
Our read
The notable change is the combination of OpenAI models with AWS deployment and governance machinery, not a promise that agents will suddenly become dependable employees. For AWS customers, the choice of self-hosted or managed runtime is a concrete option to evaluate. The announcement does not spell out pricing or the preview’s full availability, so those details will matter before teams plan a production rollout.
What to watch
- Which AWS regions and customers can access the public preview.
- How pricing works across the two execution options.
- What controls and integrations are available in the preview, and what changes before general availability.
Discussion spark: For an agent built on OpenAI models, would you rather run it within AWS’s managed environment or keep control of the compute yourself?
Sources and evidence
not affiliated with or endorsed by Amazon Web Services (AWS)