Microsoft Ignite 2026: What’s new in Azure, the platform purpose-built for AI
Azure at Microsoft Ignite 2026: Key Highlights
In the last year, my discussions with clients have evolved significantly. Previously, the main question was about choosing the right model; now, the focus has shifted to integrating agents into business processes and how to ensure they can be trusted.
With frontier models accessible to all, the true advantage lies in what sets your business apart: your unique context. This encompasses how your business operates, the expertise of your team, and the intricate systems and data you’ve meticulously developed over the years. This is your intellectual capital, and it should remain exclusively yours, functioning seamlessly alongside your team and agents. Our mission is to enhance every foundational layer beneath your agents, from the infrastructure that supports them to the applications they utilise, all rooted in the specific context of your business.
Azure is at the forefront of this Frontier Transformation. We designed Azure as a unified system, ensuring that the knowledge agents require is securely governed while extending protection measures across the agents that represent your workforce. As we advance the open, cloud-native platform you’re already dependent upon, there’s no need to start from scratch.
As Satya Nadella stated at Microsoft Build in June, we have expanded our datacentre capacity in just 18 months more than we did in the first ten years of Azure. With more regions than any other cloud provider, you can confidently scale your AI workloads globally as demand increases.
During Microsoft Ignite this week, we’re excited to unveil the latest advancements across every layer of our platform. New features in compute, storage, and networking empower you to run AI at scale while maintaining control over your own infrastructure, without the burden of building, securing, and managing everything yourself. Azure Catalyst offers a streamlined experience for agent-led modernization, covering everything from existing systems to code.
PostgreSQL now has a dual presence at Microsoft: Azure HorizonDB, which is generally available for mission-critical applications, and the newly introduced Fabric HorizonDB, connecting your PostgreSQL applications to analytics and AI. With the availability of Fabric apps, your teams can transform trusted enterprise data into operational applications.
We are also launching Microsoft IQ, a context management platform for enterprises designed to help you develop your unique IQ and leverage it across all apps and agents. This platform provides your teams with a pathway on Azure to create, operate, and continuously enhance those agents.
Running AI at Scale with Your Preferred Control
The journey begins with the infrastructure behind your agents. As they take on real tasks, it’s essential to run them according to your specifications, selecting your models, data, and runtime locations. This freedom allows you to strike the ideal balance between performance and cost for every workload, adapting as models and requirements evolve.
However, having control also introduces complexity and risk. Platform teams can find themselves piecing together and managing a considerable portion of the AI stack, often stretching GPUs to meet fluctuating demands. When agents interface with models, tools, and data independently, it may also create new security vulnerabilities, endangering sensitive information.
Teams seek the autonomy of managing their AI infrastructure without being burdened by the extensive manual labour involved in construction, security, and maintenance.
Creating the AI Core on Azure
That’s why we’re enhancing Azure. New functionalities in compute, storage, and networking merge to create a singular AI foundation, allowing you to retain control over your models and workloads while Azure manages the underlying components. Develop AI efficiently with Microsoft Foundry in a fully managed environment and operate it at scale using Azure Kubernetes Service (AKS) when you need greater control over performance, security, and costs.
Ignite Session: AI Infrastructure from Cloud to Edge
Azure manages and optimises the infrastructure for you, allowing your teams to devote less time to configuration and more time to putting models and agents into production. Workloads are allocated according to the available capacities, making the most of your GPU investments as demand increases. Centralised security is enforced across agents, models, and tools, equipping security teams with consistent controls, no matter where agents are deployed. Furthermore, because it’s founded on open-source projects, your workloads remain portable, potentially extending even to your own servers outside of Azure, keeping your options open as your requirements change.
The outcome is AI that scales with your business, operating on infrastructure you control.
To discover how organisations are successfully implementing AI at scale, join Mark Russinovich and Pavan Davuluri in the enlightening session “AI Infrastructure from Cloud to Edge.”
Enhancing Every Layer of Your AI Landscape with Azure Cobalt 200
As agents learn to reason, make decisions, and run continuously, the requirements for cloud infrastructure are evolving. That’s why we’ve designed Azure as an integrated system, from silicon to servers to services. Combining our own Maia and Cobalt chips with cutting-edge technology from NVIDIA and AMD, we ensure every workload is executed on the most suitable silicon. Our strategy focuses on embedding performance, efficiency, and security at every level, resulting in AI that’s faster and more cost-effective at scale, allowing you to strategise around a clear roadmap.
Azure Cobalt 200 virtual machines, our highest-performing general-purpose VMs, are now widely available. Specifically designed for scale-out, cloud-native, and agent-driven AI workloads, these VMs provide up to 50% improved CPU performance compared to our Cobalt 100 VMs. When used alongside Maia, they create a robust infrastructure for your planning needs.
Agent-led Modernization and Operations: From Existing Systems to Code
Agents are only as effective as the systems they can interact with. Much of what differentiates your business is embedded in your applications and databases, along with years of business logic relied upon by your teams. Modernisation allows you to leverage existing apps and data in new AI-enhanced experiences, enabling teams to innovate faster while also reducing the costs and risks associated with legacy systems, thus creating a more flexible foundation for growth.
However, this journey isn’t straightforward. Modernisation often requires navigating through complex systems and altering code that few truly comprehend. Ensuring that your resources are managed efficiently means filtering through constant signals with limited understanding of what aspects are most critical. Furthermore, as agents take on additional responsibilities, maintaining necessary oversight is crucial.
This is why we are introducing agents across the spectrum, from existing systems to code. Azure Catalyst assists you in modernising applications, code, and databases from inception to implementation. Enhanced features within Azure Copilot, Azure SRE Agent, and Azure Monitor ensure your cloud and databases are optimised, while keeping human oversight intact.
Modernise for AI with Azure Catalyst
Azure Catalyst represents a cohesive approach to agent-driven modernisation. It unites discovery, planning, migration, modernisation, and code intelligence, allowing IT teams and developers to collaborate on the same blueprint, covering infrastructure, applications, and data. Azure Catalyst transforms your existing technology into a foundational system for AI, using the business logic inherent in your applications within new AI-driven experiences and agent-led business operations.
For IT teams, the Azure Catalyst agent facilitates the journey from migration planning to comprehensive modernisation and execution. Developers benefit from GitHub Copilot’s modernisation features that translate plans into code using their current tools, supplemented by governance frameworks tailored to customisation needs, ensuring every modernisation project meets specific business requirements. Azure Catalyst also empowers teams with deeper insights into enterprise applications, enabling rapid modernisation and identification of areas where legacy processes could be enhanced through AI.
Moreover, modernisation helps to mitigate security vulnerabilities. Outdated runtimes, unpatched dependencies, and ageing code widen your threat landscape. Azure Catalyst collaborates with Microsoft Security to address these issues: Microsoft Perception and MDASH assess vulnerabilities across your estate and code, Azure Catalyst identifies critical code and databases needing urgent modernisation, and GitHub Copilot assists in rendering repairs concurrently. Each application emerges from the process more secure than it entered.
Maintain Cloud Performance with Agent-led Operations
As applications expand across hybrid infrastructures, microservices, and AI workloads, leading businesses are adopting operational models where insight seamlessly translates into action. This is the essence of agentic operations: AI agents observe, reason, and facilitate action throughout the cloud lifecycle, guided by your intent. Signals enable coordinated workflows that enhance performance, cost-efficiency, and reliability during operation, while ensuring that human oversight is preserved.
At Microsoft Ignite, we are extending this model across more of your systems. The Azure Copilot deployment agent, now generally available, helps transition from natural language requirements to deployed cloud workloads. You can even optimise agent-led operations involving Oracle and SAP with Azure Workload Center, currently in limited preview.
The Azure SRE Agent integrates agentic operations within secure production environments, offering new enterprise management controls in preview. Meanwhile, Microsoft Entra Agent ID and Agent 365 provide central teams with a consistent method for discovering, governing, and auditing agents throughout their lifecycle. We are enhancing agents’ capacity for more profound insights, enabling them to detect issues sooner, diagnose problems quicker, and respond more effectively. The new Azure Monitor health models, generally available, grant agents a real-time overview of your application architecture and business objectives, allowing them to propose solutions.
Creating the Next Generation of AI Applications
Next, we consider the applications powered by your agents, where business data transforms into business value. For every organisation, the crucial question is how swiftly it can transition from a concept to an application that the entire organisation can depend on. Today, individuals creating AI applications span the entire organisation. Developers are crafting systems of critical importance and new applications utilising agents, while teams across the business articulate the applications they require using Copilot. While each group has different needs from the platform, every application must connect to business data and function under reliable governance.
We are constructing a singular foundation for all of them, commencing with PostgreSQL, the native database that a growing number of developers favour. With Azure HorizonDB and Fabric HorizonDB, you can deploy PostgreSQL wherever your applications require it, from essential systems to analytics and AI. Fabric apps leverage this foundation, enabling teams to convert business data into functional applications, whether they originate from code or conversations.
Empower Your AI Applications with PostgreSQL on Azure HorizonDB and Fabric HorizonDB
Modern applications developed by developers are now capable of reasoning over data, calling models, and acting on behalf of users, so the underlying database must efficiently handle transactions, vector data, and AI at a mission-critical scale. Azure HorizonDB, now widely available, is a fully managed, PostgreSQL-compatible database designed specifically for this purpose. It offers millisecond latency with integrated vector indexing, enabling your applications to locate, reason, and act on data efficiently within a single platform. Top-tier security is inherent, and developers can continue using familiar tools, including GitHub Copilot and Visual Studio Code.
Ignite Session: Data Foundations for Modern Applications…
We’re enhancing PostgreSQL’s capabilities directly within Microsoft Fabric. Fabric HorizonDB, now in public preview, delivers a serverless, Fabric-native PostgreSQL experience alongside SQL databases in Fabric and Cosmos DB in Fabric. By combining operational databases, analytics, and AI on a single unified, governed platform, developers can create intelligent applications with reduced data movement and complexity. Fabric HorizonDB will serve as the PostgreSQL backend for Fabric apps.
To streamline database management across various environments, Database Hub in Fabric is now widely available. It offers a cohesive experience to discover, monitor, and manage databases across Azure, on-premises settings, and Fabric, with Fabric Copilot highlighting areas needing attention.
Transform Ideas into Enterprise-Ready Applications with Fabric Apps, Now Generally Available
Innovative app ideas emerge from all sectors of the business, and with Copilot, individuals can create an application swiftly. However, turning a prototype into a production-ready solution using trusted data remains a significant challenge. Fabric apps, now generally available, seamlessly integrate backend services required by applications with direct access to trusted data within Fabric, featuring built-in security and governance suited for enterprises. Developers and coding agents can build in code with open-source tools, while Fabric provisions everything, from databases to authentication, eliminating infrastructure management burdens. This facilitates the transformation of concepts into enterprise-ready applications that can scale across the organisation.
Additionally, we are expanding Fabric’s role as the data foundation for AI-driven applications. The new Microsoft Copilot integration, currently in public preview, enables anyone to articulate the application they need and bring it to life, with the possibility of using Fabric as a readily available data backend, providing databases, storage, and data APIs. We’re also introducing Copilot Credits as a new usage-based billing option for Microsoft Fabric, offering greater flexibility alongside existing capacity-based choices.
Moreover, Power BI creators can go beyond traditional static reports. With an innovative agent-led app creation experience in Power BI Desktop, now in preview, they can convert existing semantic models into interactive applications that gather user inputs, update data, and facilitate operational workflows.
Ground Every Agent in Your Business Context with Microsoft IQ
Most organisations already possess the knowledge necessary for their agents. This valuable information exists in the semantic models built by analysts in Power BI, the documents and procedures relied upon by teams, the ongoing conversations across Microsoft 365, and the record-keeping systems such as SAP and Oracle that you manage daily. The challenge lies in its distribution; every new agent begins with a blank slate, and each team is redundantly recreating similar contexts.
That’s why we are unveiling Microsoft IQ, an enterprise context management platform aimed at constructing your organisation’s IQ. It consolidates Work IQ, Fabric IQ, Foundry IQ, Web IQ, and over 350 connectors, ensuring the information your agents require is captured once, governed centrally, and reused across all agents and applications.
At the core of Microsoft IQ are innovative business plugins, now available as part of the Frontier programme. These plugins transform your team’s expertise into reliable context, skills, and tools that anyone within the organisation can utilise. For instance, consider a supply chain disruption: a business plugin can instruct an agent to use the components and orders semantic model in Fabric IQ to identify affected orders, the supplier contracts in Foundry IQ to verify late-delivery terms, and Work IQ to identify the individual responsible for the decision. Domain experts draft it once, and every agent across the company inherits the knowledge.
Additionally, teams can replicate proven evaluations across agents instead of redundantly rebuilding them. The underlying foundation of Microsoft IQ brings Microsoft Foundry and Microsoft Fabric together, facilitating a cohesive pathway to develop, deploy, and continuously optimise agents. Agent 365 provides IT teams with a centralized solution for monitoring, governing, and securing these agents.
Your IQ remains your intellectual property; it isn’t confined to a single assistant. The same context and business plugins are accessible to your agents no matter where they are running, whether it’s in Microsoft Copilot, GitHub Copilot, Microsoft Foundry, or other assistants such as Claude and ChatGPT, all managed with permissions and governance. Furthermore, agents are empowered to leverage business context from your record-keeping systems like SAP and Oracle, ensuring they operate from the data your teams already trust. Explore the latest integrations with Microsoft IQ and Oracle here.
Find out how the Australian Energy Market Operator uses Microsoft IQ to maintain the security of Australia’s power grid:
Currently, nearly 60,000 customers are actively using Microsoft IQ. As every agent is built on a shared context, each new addition enhances the overall capabilities of your organisation.
Discover These Innovations at Microsoft Ignite
Every agent you create is only as effective as its understanding of your business context. This is why we’re committed to ensuring your IQ—the essence that makes your enterprise distinct—is accessible to your agents throughout every layer of Azure, with built-in governance and security. It remains yours, and its value only increases with everything you develop upon it. To delve deeper into Microsoft IQ, I invite you to attend my session at Microsoft Ignite, “Harnessing Your IQ with a Continuously Improving Agent Platform,” either in person or online for free, or watch the recorded session afterwards.
For a comprehensive look at all the innovations we are introducing to Azure, check out the complete list of sessions in the Microsoft Ignite session catalog.
Share this content:
Discover more from Qureshi
Subscribe to get the latest posts sent to your email.