AWS Unveils Kiro: An AI Agent That Codes Independently for Days

Amazon Web Services kicked off its massive re:Invent conference with a wave of AI announcements. The cloud giant revealed several new tools and models designed for enterprise customers. A standout announcement was the introduction of three new “Frontier” AI agents.

One agent, named Kiro, is designed to write and manage code autonomously. AWS claims it can learn a development team’s patterns and then operate on its own for extended periods. This push into advanced AI agents underscores the industry’s shift from simple chatbots to automated workers.

New AI Chips and Customizable Models Mark Major Upgrades

AWS CEO Matt Garman set the tone in his keynote. He stated that AI agents represent the next phase of tangible business value from artificial intelligence. The event’s first day was packed with product launches supporting this vision.

A new AI training chip, Trainium3, was unveiled. AWS promises it delivers four times the performance for AI training and inference. It also uses 40% less energy. The company also teased future compatibility with Nvidia’s industry-leading hardware.

The Nova family of AI models also expanded. AWS introduced four new models, including one that generates both text and images. A related service, Nova Forge, lets companies fine-tune these models with their own private data. This focus on customization is a key theme.

Autonomous Agents Take Center Stage for Business Automation

The concept of AI agents that work independently was a central theme. AWS announced new features for its AgentCore platform, which helps build these agents. One feature lets developers set strict operational boundaries for safety and control.

Another allows agents to remember user preferences over time. The Kiro autonomous agent is the most ambitious example. It is built to handle coding tasks after learning a team’s style and can run for hours or days without human intervention.

Early customer testimonials highlighted the practical impact. Ride-sharing company Lyft shared data on its use of an AI agent for customer support. They reported an 87% reduction in average issue resolution time, according to their presentation.

The launch of Kiro and its sibling agents signals a bold step toward a future where AI doesn’t just assist with tasks but completes them independently. AWS is betting that this level of automation will define the next generation of enterprise software development.

Info at your fingertips-

What exactly does the Kiro AI agent do?

Kiro is an autonomous AI agent built for software development. It writes and manages code. It learns a development team’s specific workflows and can then operate independently for long stretches.

How is AWS ensuring these AI agents are secure?

AWS introduced a feature called Policy in AgentCore. This gives developers direct tools to set boundaries and rules for agent behavior. It is a core part of their approach to safe, controlled automation.

Are these new AWS AI agents available to use now?

Preview versions of the Frontier agents, including Kiro, are available now. The new Trainium3 chip and expanded Nova models are also being rolled out to AWS cloud customers.

What was the key takeaway from the AWS re:Invent keynote?

The main theme was the shift from AI assistants to AI agents. AWS CEO Matt Garman argued that agents which automate tasks are where businesses will see real material returns on their AI investments.

Why did AWS announce a new AI chip?

The Trainium3 chip is designed to improve performance and lower costs for customers training and running AI models. It offers significant speed gains and energy savings compared to previous generations.