Artificial intelligence news.
Straight from the original announcements.

Microsoft Introduces Azure Catalyst for Agent-Led Modernization

November 17, 2026, 19:35By the AI News Bites Editorial TeamSource: Microsoft Azure
Microsoft Introduces Azure Catalyst for Agent-Led Modernization
Image: Microsoft Azure

Microsoft is introducing Azure Catalyst, an agent-led solution designed to modernize applications, data and infrastructure, alongside new Azure capabilities for building and running AI workloads. The company says the updates target the growing demands of complex, continuously operating AI agents.

Key points
  • Microsoft introduced Azure Catalyst for agent-led application, data and infrastructure modernization.
  • Agent Sandbox for AKS provides virtual machine-level isolation for agent-generated code.
  • Agent Grid adds authenticated, encrypted connections and default-deny network isolation.
  • Microsoft says Azure Catalyst has modernized 1.4 million applications.
  • Flex nodes for AKS are in preview, with support for Kubernetes, vLLM and Ray.

Azure expands its infrastructure for AI agents

Microsoft says the latest Azure updates cover isolated execution, agent networking, context management and inference efficiency. The company describes these capabilities as a way to support agents that reason across multi-agent systems, use rich context, act in real time and run continuously.

A new connected path between Microsoft Foundry and Azure Kubernetes Service lets customers build AI applications in Foundry and run them at scale through AKS. According to Microsoft, open-weight models and agents can move between the services and be deployed from the cloud to the edge with consistent infrastructure, security policies and governance.

The Agent Sandbox extension for AKS runs agent-generated code in managed sandboxes with virtual machine-level isolation for every session. Microsoft also describes pre-warmed pools for faster startup and a serverless option with no infrastructure to manage. Azure confidential computing is intended to protect sensitive data and workloads while they are in use.

New tools target agent networking and context

Microsoft says Agent Grid allows agents to connect to other agents, models, tools, data and Azure services through authenticated and encrypted connections. Identity-driven policies and default-deny network isolation control how agents communicate and which resources they can access.

For long-running agents, Azure can reuse processed prompt context instead of recomputing it on the GPU. Microsoft says this can make long tasks faster and less expensive, while cached context remains durable and controlled by the customer in its Azure tenant.

AKS AI Manager translates the needs of each model into infrastructure decisions, including where deployments should run based on available GPU capacity. Microsoft is also highlighting support for Kubernetes, vLLM and Ray, as well as flex nodes for AKS in preview and supported GPU environments beyond Azure.

Azure Catalyst aims to reduce modernization work

Azure Catalyst uses agents to analyze legacy code, map dependencies, and plan and execute modernization projects. Microsoft says work that previously took months or years can be completed in weeks or days, including in organizations with limited resources and skills.

The service connects discovery, design and execution in one experience. Its capabilities include estate insights, AI-powered application discovery, modernization strategies, cost comparisons and AI-generated migration waves. Execution covers applications, containers, storage, networking and databases, while the platform also supports Java and .NET modernization and migration for VMware, Hyper-V and physical servers.

Microsoft says Azure Catalyst has modernized 1.4 million applications. The company cites Levi Strauss as a customer modernizing enterprise systems and deploying more than 1,000 agents across them.

For Azure users, the announcement combines tools for creating AI applications with a route for adapting existing systems and data to those workloads. Microsoft is positioning Azure Catalyst and its infrastructure updates as connected parts of an Azure platform built for AI development, deployment and modernization.

Related news

Microsoft expands Foundry hosted agents with enterprise controls
Products and toolsNovember 18, 2026, 18:00

Microsoft expands Foundry hosted agents with enterprise controls

Microsoft is expanding hosted agents in Foundry Agent Service with networking, governance and deployment features aimed at moving enterprise projects from pilots into…

Microsoft Makes Azure Cobalt 200 VMs Generally Available
Chips and infrastructureNovember 17, 2026, 19:20

Microsoft Makes Azure Cobalt 200 VMs Generally Available

Microsoft has made Azure Cobalt 200-based virtual machines generally available, offering up to 50% better performance than Cobalt 100 VMs, according to the company. The…

Datadog expands Azure monitoring for AI workloads and cloud costs
Products and toolsNovember 17, 2026

Datadog expands Azure monitoring for AI workloads and cloud costs

Datadog is expanding its Azure monitoring coverage with 55 additional integrations for Azure Monitor services, alongside tools for tracking AI workloads and cloud…