### AWS Weekly Roundup: Key Launches and Innovations
Last week, AWS brought together global innovators for the AWS Heroes Summit, an exclusive event focused on collaboration and technical deep-dives with product teams. During this gathering, AWS CEO Matt Garman provided an inspiring kickoff, while attendees engaged in sessions ranging from AI to serverless and containers. The event highlighted the power of community as builders connected, shared knowledge, and sparked new collaborations.
Among the standout announcements last week were several major updates aimed at enhancing AI capabilities and modernization efforts. Web Search on Amazon Bedrock now supports OpenAI models, enabling real-time internet browsing without data egress. Runtime Instances on Amazon Bedrock AgentCore offer more control over AI agent execution, while Vector Search for Amazon DynamoDB provides semantic retrieval alongside existing data. Additionally, AWS Transform continuous modernization is now generally available, and Lambda function bandwidth has increased significantly.
Beyond these launches, AWS open-sourced Dogwood, a governance language for AI agents, announced support for Agent Plugins as an open standard, and introduced Kiro Crew—a persistent, collaborative workspace for multi-agent development.
For a complete overview of AWS announcements and resources, the [AWS What’s New page](https://aws.amazon.com/about-aws/whats-new/) and [AWS Blogs page](https://aws.amazon.com/blogs/) are excellent references.
—
### FAQ
**Q1: What was the main highlight of the AWS Heroes Summit?**
The AWS Heroes Summit facilitated direct collaboration between global experts and AWS product teams, focusing on AI, serverless, and containers. It provided a platform for technical deep-dives and feedback sessions that turned conversations into real-world collaborations.
**Q2: Which AI models are now available for web search on Amazon Bedrock?**
Amazon Bedrock now supports OpenAI models, including GPT-5.4, GPT-5.5, and GPT-5.6 Sol/Terra/Luna, for web search capabilities. This allows AI applications to access up-to-date information while maintaining data residency within AWS.
**Q3: What is the purpose of AWS Transform continuous modernization?**
AWS Transform helps engineering teams analyze and remediate technical debt across source code repositories at scale. It enables ongoing, automated modernization of mainframe and legacy workloads rather than one-time migration events.
**Q4: What does Dogwood do in the context of AI agents?**
Dogwood is a purpose-built governance language for AI agents that supports Cedar policies and adds temporal conditions. It allows decisions to depend on the history of an agent’s actions within a session, enhancing control and governance.
**Q5: How has AWS Lambda network bandwidth been improved?**
AWS Lambda functions now support up to 3,000 Mbps of network bandwidth for functions configured with 2 GB of memory or more, enabling data-intensive workloads and faster communication between functions and other AWS services.
—
### Conclusion
The recent AWS updates reflect a strong push toward enhancing AI capabilities, streamlining modernization efforts, and empowering developers with greater control and collaboration tools. From real-time web search on Amazon Bedrock to open governance standards for AI agents, these innovations are set to accelerate development and deployment. As always, AWS continues to provide a robust ecosystem for builders to innovate and scale efficiently. Keep an eye on upcoming AWS events and resources to stay engaged with the latest advancements.
