Microsoft expands Azure with AI infrastructure and agent-led tools

Microsoft is expanding Azure across infrastructure, application modernization and enterprise data, positioning the cloud platform as a foundation for agents and AI workloads. The announcements at Microsoft Ignite 2026 include Azure Cobalt 200 VMs, Azure Catalyst and Microsoft IQ.
- Azure Cobalt 200 VMs are now generally available.
- Microsoft says Cobalt 200 offers up to 50% better CPU performance than Cobalt 100.
- Azure Catalyst unifies agent-led modernization from discovery and planning through execution.
- Microsoft IQ is designed to manage enterprise context across apps and agents.
- Azure HorizonDB and Fabric HorizonDB provide two PostgreSQL offerings for different workloads.
Azure adds a managed foundation for AI workloads
Microsoft says its new AI capabilities across compute, storage and networking are designed to let customers retain control of their models and workloads while Azure manages more of the underlying infrastructure. The company is bringing these capabilities together as an AI foundation that supports Microsoft Foundry and Azure Kubernetes Service, with the latter offering more control over performance, security and cost.
According to Microsoft, Azure can set up and tune infrastructure, place workloads according to available capacity and enforce security centrally across agents, models and tools. The company also says the platform is built on open-source projects, helping keep workloads portable, including to servers outside Azure.
Microsoft also says Azure has added more datacenter capacity in 18 months than in its first decade and has more regions than any other cloud provider. Those claims are presented as support for scaling AI workloads globally as demand grows.
Cobalt 200 targets cloud-native and agentic workloads
Azure Cobalt 200 virtual machines are now generally available. Microsoft describes them as its highest-performing general-purpose VMs, designed for scale-out, cloud-native and agentic AI workloads.
The company says Cobalt 200 delivers up to 50% better CPU performance compared with Cobalt 100 VMs. Microsoft positions its own Maia and Cobalt chips alongside hardware from NVIDIA and AMD, saying this approach is intended to match different workloads with suitable silicon while improving performance, efficiency and security.
For customers running agents continuously, the announcement reflects Microsoft’s focus on the infrastructure beneath AI applications. The company says the combination of Cobalt and Maia provides infrastructure that organizations can plan around as their workloads change.
Azure Catalyst brings agents into modernization work
Microsoft is also introducing Azure Catalyst, a unified experience for agent-led modernization. It brings discovery, planning, migration, modernization and code intelligence into one workflow covering infrastructure, applications and data.
Azure Catalyst is intended to help IT teams modernize existing applications, code and databases, putting business logic in older systems to work in AI-powered experiences and agentic business processes. Microsoft says its IT-focused agent extends from migration planning into modernization and execution, while GitHub Copilot modernization carries plans into developers’ existing tools.
The company says Azure Catalyst includes agent governance rulebooks and support for customers’ own models. It also works with Microsoft Security to prioritize vulnerabilities across estates and code. Microsoft Perception and MDASH rank vulnerabilities, Azure Catalyst identifies code and databases to modernize first, and GitHub Copilot can turn those findings into fixes.
Other announcements include generally available Azure HorizonDB for mission-critical scale, Fabric HorizonDB for connecting PostgreSQL applications with analytics and AI, and Fabric apps for turning governed enterprise data into working applications. Microsoft IQ is described as an enterprise context management platform that helps organizations build and use their own business context across apps and agents. For users, the direction is clear: Azure is being developed to support AI while keeping organizational data, systems, security controls and business knowledge within a governed platform.
Microsoft Azure: Microsoft Ignite 2026: What’s new in Azure, the platform purpose-built for AI

