Amazon Web Services continues its aggressive expansion into the generative AI landscape with several high-impact releases targeting both developers and enterprises. The headline addition is NVIDIA Nemotron 3 Super on Amazon Bedrock, a sophisticated language model built for complex reasoning and summarization. This allows users to access high-performance AI through a single interface—known as an API—without the burden of managing their own physical servers or hardware infrastructure.

To help businesses build more specialized tools, AWS introduced the Nova Forge SDK. This software kit allows companies to take existing AI models and refine them using their own private data through a process called fine-tuning. By tailoring models to specific industry knowledge, enterprises can deploy more accurate and relevant AI solutions directly within the Bedrock ecosystem. This move significantly lowers the technical barrier for creating custom business intelligence tools.

Beyond models, the update highlights a breakthrough in AI reliability through Strands Steering Hooks. While most AI agents rely on simple instructions (prompt engineering), Steering Hooks offer a precise way to guide an AI's behavior, achieving 100% accuracy in controlled tests. As Swami Sivasubramanian (VP of Data and AI at AWS) noted, these tools empower builders to create more reliable autonomous systems. Coupled with a new 99.99% uptime guarantee for clusters, these updates provide the stability required for large-scale AI training in production environments.

Amazon Web Services continues its aggressive expansion into the generative AI landscape with several high-impact releases targeting both developers and enterprises. The headline addition is NVIDIA Nemotron 3 Super on Amazon Bedrock, a sophisticated language model built for complex reasoning and summarization. This allows users to access high-performance AI through a single interface—known as an API—without the burden of managing their own physical servers or hardware infrastructure.

To help businesses build more specialized tools, AWS introduced the Nova Forge SDK. This software kit allows companies to take existing AI models and refine them using their own private data through a process called fine-tuning. By tailoring models to specific industry knowledge, enterprises can deploy more accurate and relevant AI solutions directly within the Bedrock ecosystem. This move significantly lowers the technical barrier for creating custom business intelligence tools.

Beyond models, the update highlights a breakthrough in AI reliability through Strands Steering Hooks. While most AI agents rely on simple instructions (prompt engineering), Steering Hooks offer a precise way to guide an AI's behavior, achieving 100% accuracy in controlled tests. As Swami Sivasubramanian (VP of Data and AI at AWS) noted, these tools empower builders to create more reliable autonomous systems. Coupled with a new 99.99% uptime guarantee for clusters, these updates provide the stability required for large-scale AI training in production environments.