Bedrock Brief 05 Nov 2025

Pleri AI

Pleri AI

— 4 min read
Bedrock Brief 05 Nov 2025

Well, well, well. Looks like OpenAI's been playing the field this week, and AWS just scored the prom date. In a move that's got tech nerds and Wall Street suits alike buzzing, OpenAI inked a $38 billion deal with Amazon Web Services. This isn't just any old cloud agreement—we're talking hundreds of thousands of NVIDIA GPUs and the potential for millions of CPUs. It's like OpenAI just ordered the all-you-can-eat buffet of computing power, with AWS as the chef de cuisine.

But hold your horses, Microsoft fanboys. This doesn't mean Redmond's getting kicked to the curb. OpenAI's still committed to spending a cool $250 billion on Azure services. It's more like they're going from a monogamous cloud relationship to an open one. And they're not stopping there—Oracle, Google, and even Verizon are getting a piece of the AI infrastructure pie. It's a $1.4 trillion spending spree that's making some folks wonder if we're watching an AI bubble inflate faster than a bounce house at a kid's birthday party.

Speaking of infrastructure, Verizon's jumping on the AI bandwagon too. They're partnering with AWS to lay down some serious fiber, aiming to create the backbone for the next wave of AI innovation. It's like they're building the autobahn for data, and AI applications are the sports cars revving their engines, ready to zoom. With all these deals flying around, one thing's clear: the AI arms race is heating up, and everyone wants a seat at the table. Or should I say, a slot in the data center?

Fresh Cut

  • AWS Config expands its governance capabilities with 42 new managed rules, enabling developers to assess security, cost, and operations across various AWS services like Amazon EKS, EC2, Cognito, and Lightsail. Read announcement →
  • Amazon Bedrock AgentCore Runtime introduces direct code-zip file upload for AI agents, enabling faster prototyping and iteration without sacrificing enterprise-grade security and scalability. Read announcement →
  • AWS Config expands monitoring capabilities with 52 new resource types across EC2, Bedrock, and SageMaker, enabling developers to better track and audit a wider range of AWS resources for improved security and compliance. Read announcement →
  • SAP Cloud ERP is available in Europe (Frankfurt), offering faster implementation, AI-powered processes, and energy-efficient AWS Graviton processors for organizations looking to modernize their enterprise resource planning. Read announcement →
  • AWS Marketplace introduces flexible pricing and simplified deployment for AI agents, allowing developers to easily integrate and scale AI tools with options like usage-based pricing and streamlined OAuth credential management. Read announcement →
  • The Model Context Protocol (MCP) Proxy for AWS enables developers to securely connect AI tools to AWS services using SigV4 authentication, expanding possibilities for integrating AWS resources into AI workflows. Read announcement →
  • TwelveLabs' Pegasus 1.2, a powerful video-to-text AI model, is available in three new AWS regions, enabling developers to build video-intelligence applications with reduced latency and improved geographic coverage. Read announcement →
  • Amazon EKS users can now import up to 50 Kubernetes custom labels per pod as cost allocation tags, enabling more precise cost attribution for containerized applications and improved financial management of EKS clusters. Read announcement →
  • New EC2 API allows developers to view and manage the topology of their capacity reservations, enabling more efficient scheduling and performance optimization for distributed AI, ML, and HPC workloads across thousands of instances. Read announcement →
  • AI-powered tools in AWS Serverless MCP Server simplify Lambda event source mapping setup, offering automated configuration, optimization, and troubleshooting for developers building event-driven serverless applications. Read announcement →

The Quarry

Build reliable AI systems with Automated Reasoning on Amazon Bedrock – Part 1

Amazon Bedrock Guardrails now offers Automated Reasoning checks, a game-changer for regulated industries needing ironclad AI compliance. Unlike traditional QA methods that rely on statistical sampling, these checks use formal verification techniques to mathematically prove that every single AI response adheres to predefined policies and domain knowledge. This approach leverages symbolic execution and constraint solving to analyze all possible paths through the AI's decision logic, ensuring 100% coverage and eliminating the uncertainty inherent in probabilistic testing methods. Read blog →

More posts:

Core Sample

AI in the Wild: Lovable | What Happens When Anyone Can Build with AI?

Lovable, an AI startup powered by AWS, has revolutionized software development by enabling users to create fully functional applications through natural language descriptions. By leveraging AI to bridge the gap between imagination and execution, Lovable achieved an impressive $100M ARR in just months, showcasing the transformative potential of AI-driven, no-code development. This approach not only accelerates the development process but also democratizes software creation, empowering a wide range of users from novice entrepreneurs to seasoned teams to build more efficiently and creatively. Watch video →

More videos:

Read more