AZ-204 isn't quietly disappearing. It's being rebuilt into an exam for a different job entirely.
Microsoft has confirmed that AZ-204: Developing Solutions for Microsoft Azure retires on July 31, 2026. Its replacement, AI-200: Developing AI Cloud Solutions on Azure, moves the credential away from general cloud development and toward building AI-powered applications.
This guide breaks down exactly what changed, what stayed the same, and how to decide based on where you are right now.
What AZ-204 Actually Tests
AZ-204 is the certification behind the Microsoft Certified: Azure Developer Associate credential. It has been the standard certification for Azure developers for years and is one of the most widely held associate-level credentials in the Microsoft ecosystem.
The exam covers a broad range of practical developer skills, including:
- Azure compute solutions, including App Service, Azure Functions, and Container Apps.
- Azure storage solutions, including Cosmos DB, Blob Storage, and Azure Cache for Redis
- Azure security implementation, including the Microsoft Identity Platform, Key Vault, and managed identities
- Monitoring and optimization, including Application Insights and autoscaling
- Integration with Azure services and third-party services, including API Management, Event Grid, Event Hubs, and Service Bus
Most candidates report needing 80 to 120 hours of hands-on preparation to be exam-ready, since AZ-204 tests practical, scenario-based knowledge rather than simple recall.
AI-200: The Exam Replacing It
AI-200 leads to the Microsoft Certified: Azure AI Cloud Developer Associate credential. Instead of testing the full breadth of Azure development, it narrows the focus specifically to building AI-powered applications on Azure.
Based on Microsoft's published exam guide, AI-200 centers on four main areas:
- Developing containerized solutions on Azure (20–25%)
- Developing AI solutions using Azure data management services, including vector search across Cosmos DB, PostgreSQL with pgvector, and Azure Managed Redis (25–30%)
- Connecting to and consuming Azure services (20–25%)
- Securing, monitoring, and troubleshooting Azure solutions (20–25%)
The exam also expects familiarity with retrieval-augmented generation, Azure OpenAI Service, AI agent development, and responsible AI practices such as content safety and governance controls. Candidates are expected to know Python, which is the only programming language named in the official study guide, unlike AZ-204, which supports both C# and Python.
What Changed Between AZ-204 and AI-200 (And What Didn't)
The shift is not a name change. Roughly 60 percent of AZ-204 content carries forward, but several core topics are being removed entirely while new AI-focused domains take their place.
Dropped from the exam:
- Blob Storage SDK
- Microsoft Identity Platform
- Microsoft Graph
- API Management
- Azure Event Hubs
- Azure Queue Storage
- Application Insights SDK, replaced by OpenTelemetry
New in AI-200:
- Vector search across Cosmos DB NoSQL, PostgreSQL with pgvector, and Azure Managed Redis
- Kubernetes Event Driven Autoscaling (KEDA) in Container Apps
- Full AKS deployment and management using manifest files
- OpenTelemetry for distributed tracing
- KQL queries for logs and metrics
- Retrieval-augmented generation (RAG) pattern implementation
AZ-204 vs AI-200 at a Glance
| Area | AZ-204 (retiring July 31, 2026) | AI-200 (replacement) |
| Credential | Azure Developer Associate | Azure AI Cloud Developer Associate |
| Focus | General Azure development | AI-powered application development |
| Languages | C# and Python | Python only |
| Storage/Data | Cosmos DB, Blob Storage, Azure Cache for Redis | Cosmos DB with vector search, PostgreSQL with pgvector, Azure Managed Redis |
| Compute | App Service, Azure Functions, Container Apps | Containerized solutions, AKS with manifest files, KEDA autoscaling |
| Security | Microsoft Identity Platform, Key Vault, managed identities | Securing and monitoring AI solutions, responsible AI/content safety |
| Monitoring | Application Insights SDK | OpenTelemetry, KQL queries |
| Integration | API Management, Event Grid, Event Hubs, Service Bus | Azure OpenAI Service, RAG pattern, AI agent development |
| Content overlap | Baseline | Roughly 60% carried over from AZ-204 |
| Prep time | 80-120 hours | Not yet established (exam in beta) |
Existing AZ-204 study material, whether books, courses, or practice tests, will not fully prepare you for AI-200. The overlap is real but partial. Treating AI-200 prep as a small add-on to AZ-204 knowledge is a mistake.
Should You Still Finish AZ-204 Before It's Gone?
If part of you is thinking "what's the point of finishing an exam that's about to be retired anyway," the answer depends entirely on where you are in your prep right now, not on which exam sounds more modern.
Take AZ-204 if:
- You're close to exam-ready and can test before July 31, 2026. Once earned, the credential stays valid through its full renewal period even after the exam retires, so you lose nothing by finishing now.
- You need a credential recruiters and hiring managers recognize immediately. AZ-204 has years of established presence in job descriptions and hiring criteria.
- Your day-to-day work covers the full Azure developer stack, including compute, storage, security, and integration, rather than being primarily AI-focused.
Before booking your exam, validate your readiness with Microsoft AZ-204 Exam Practice Questions to confirm you're genuinely prepared.
Move to AI-200 if:
- You're just getting started with your preparation. There's little value in spending 80+ hours studying topics such as API Management or Event Hubs that won't appear on any future exam.
- Your work is already moving toward building generative AI applications, RAG-based solutions, copilots, or AI agents.
- You want a certification path that reflects where enterprise Azure development is heading rather than where it's been.
The official AZ-204T training course already retired on May 29, 2026, and Microsoft Learn has shifted its focus to AI-200 as the current learning path.
One practical note: AI-200 is currently in beta. Practice exams and mature study materials are not fully available yet, and beta exam scores are typically delayed while Microsoft evaluates question quality. Factor that timeline in if you'd rather study with a full set of resources before testing.
This Isn't Just an AZ-204 Story
Microsoft is retiring a cluster of established certifications in 2026 and replacing them with AI-focused alternatives across multiple tracks:
- SC-500, Cloud and AI Security Engineer, is replacing AZ-500
- AI-300, Machine Learning Operations Engineer is replacing DP-100
- AI-900 was already updated to AI-901 earlier in the year
The pattern is the same across every transition. Microsoft is embedding AI competency directly into role-based certifications rather than treating AI as a separate, optional specialization. For Azure developers, the expectation is no longer just knowing how to build applications on Azure. It's knowing how to build applications that use AI effectively, securely, and responsibly.
If you're mapping out your broader certification path, our full library of Microsoft certification exams covers these transitions across every track.
So Which One Should You Take?
If you can realistically finish AZ-204 before July 31, 2026, and your current work spans the broader Azure developer stack, finishing it is still a smart, low-risk move that protects your existing preparation time. If you're starting from zero, or your work is already centered on AI-powered application development, building toward AI-200 from day one is the better long-term investment.
Either path is reasonable. What matters is deciding deliberately, based on your actual timeline and career direction, not defaulting to whichever exam happens to be more familiar.
Frequently Asked Questions
Is AZ-204 still valid after it retires?
Yes. Once you earn the AZ-204 credential, it remains valid through its full renewal period even after the exam itself retires on July 31, 2026. Only the ability to sit for the exam goes away, not the certification you already hold.
Can I use AZ-204 study material to prepare for AI-200?
Partially. About 60 percent of the content carries over, but core AZ-204 topics like API Management, Event Hubs, Microsoft Identity Platform, and Microsoft Graph are dropped entirely. AI-200 adds vector search, RAG implementation, AKS deployment, and OpenTelemetry, none of which existing AZ-204 material covers. Treat AI-200 as a distinct prep effort, not an extension of AZ-204 study.
Should I still start studying for AZ-204 if I haven't begun yet?
Generally no, unless your work is squarely focused on the broader Azure developer stack and you can realistically finish before July 31, 2026. If you're starting from zero, AI-200 better reflects where Azure development and hiring are heading.
When will AI-200 practice exams and full study materials be available?
AI-200 is currently in beta, so mature practice tests and study guides are still limited, and beta exam scores are typically delayed while Microsoft evaluates question quality. Factor this into your timeline if you prefer studying with a complete resource set before testing.