Microsoft AI Updates: November 18, 2026
Fully automated. Found, summarized and published by an automated AI pipeline. How posts are labeled
1. Microsoft Introduces Microsoft IQ, Combining Work IQ, Fabric IQ, Foundry IQ and Web IQ for Agents
Microsoft announced Microsoft IQ at Ignite, an enterprise context platform that combines Work IQ, Fabric IQ, Foundry IQ and Web IQ with more than 350 third-party connectors. Organizations can package context, skills and sources into governed business plugins that agents reuse in Microsoft Copilot, GitHub Copilot, Microsoft Foundry through MCP, and third-party assistants such as ChatGPT and Claude. Web IQ is generally available to every Azure customer starting at $1 per 1,000 calls, Foundry IQ knowledge graphs and Microsoft Retriever 1 are in preview, and Foundry IQ Serverless is planned for December 2026. Source
2. Microsoft Foundry Adds a Continuous Improvement Loop With Agent Optimizer and Programmed Training
Microsoft announced continuous improvement for Microsoft Foundry, in which agents learn from traces and user input and Foundry tests candidate improvements against evaluations before teams decide what ships. The Agent optimizer improves instructions, skills and model choice and tests candidates against evaluations, and Programmed Training (preview) lets engineers write Python training recipes that run on serverless GPUs. Other items include Insights and observability for LangGraph, LangChain and CrewAI agents (generally available), voice agents (generally available) and a Microsoft Copilot harness with reusable skills and sub-agents (preview). Source
3. Microsoft Introduces Azure Catalyst for Agent-Led Modernization and New AKS Infrastructure for Agents
Microsoft introduced Azure Catalyst, an agent-led approach to modernizing applications, data and infrastructure that covers discovery, migration planning and execution, with GitHub Copilot used to remediate vulnerabilities. The accompanying AI infrastructure announcements include an Agent Sandbox extension for AKS that runs agent-generated code in VM-level isolated sessions, Agent Grid for authenticated and encrypted connections between agents, models and tools, and an AKS AI Manager that places model deployments where GPU capacity is available. Flex nodes for AKS and Windows Server 2028 are in preview. Source
4. Azure Cobalt 200 VMs Reach General Availability in 14 Regions With Up to 50% Better Performance Than Cobalt 100
Microsoft announced general availability of Azure Cobalt 200 virtual machines, built on a 3nm Arm chip with 132 Neoverse V3 cores. Microsoft claims up to 50% better performance than Cobalt 100, including up to 1.5x for AI agent tool execution and up to 1.9x for database workloads. The VMs are available in 14 Azure regions, with expansion to 23 regions planned by the end of 2026. Source
5. Microsoft and Oracle Connect Oracle AI Database@Azure to Copilot, Excel and Foundry Agents
Microsoft and Oracle announced integrations that let AI agents work with Oracle enterprise data across finance, supply chain, HR and customer operations. Database calls from agents run through a managed MCP server with Microsoft Entra authentication, and Foundry developers can build agents that connect to Oracle AI Database using managed MCP servers and the A2A protocol. The Oracle AI Database plug-in for Excel is available now through Microsoft Marketplace, and Azure Workload Center for Oracle is in limited preview. Source