AZ-900 vs AI-900: Which Microsoft Certification Should You Take First? (2026)
Same cost. Same passing score. Same exam format. Two very different skill sets. Here is exactly which one fits your role — and why.
The Short Answer
- • Take AZ-900 first if you work in IT, cloud infrastructure, development, or any technical role touching Azure
- • Take AI-900 first if you are in a business, analytical, or non-technical role where AI literacy is immediately relevant
- • Take both if you are building toward a cloud or AI-focused career — they complement each other with minimal overlap
AZ-900 vs AI-900: Side-by-Side Comparison
| Detail | AZ-900 | AI-900 |
|---|---|---|
| Full Name | Azure Fundamentals | Azure AI Fundamentals |
| Questions | 40–60 | 40–60 |
| Time Limit | 60 minutes | 60 minutes |
| Passing Score | 700 / 1000 | 700 / 1000 |
| Cost | $99 USD | $99 USD |
| Prerequisites | None | None |
| Study Time | 2–3 weeks | 1–2 weeks |
| Pass Rate (MSCertQuiz users) | 94% | 91% |
| Best For | Cloud/IT/developer roles | Business/analytical/AI roles |
| Next Certifications | AZ-104, AZ-305, AZ-204 | AI-102, DP-100, DP-203 |
What AZ-900 Actually Tests
AZ-900 validates that you understand how Microsoft Azure works as a cloud platform. Three domains:
Domain 1: Cloud Concepts (25–30%)
IaaS, PaaS, SaaS. Shared responsibility model. High availability, scalability, elasticity. CapEx vs OpEx. The consumption-based cloud model. Not Azure-specific — foundational cloud computing principles that apply universally. This domain is straightforward but cannot be skipped.
Domain 2: Azure Architecture and Services (35–40%)
The core of the exam and the highest-weighted domain. Azure compute (Virtual Machines, App Service, Functions, Container Instances, AKS), storage (Blob, Files, Tables, Queues, Disks), networking (Virtual Networks, VPN Gateway, ExpressRoute, Load Balancer), and platform services (Cosmos DB, Azure SQL, Entra ID). Questions test your ability to match workload requirements to the right Azure service — the most common question format on AZ-900.
Domain 3: Azure Management and Governance (30–35%)
Cost management and billing. Azure Policy, RBAC, resource locks. Microsoft Defender for Cloud. Azure Monitor, Log Analytics, Service Health. Azure Arc. Management groups, subscriptions, resource groups. This domain trips up candidates who study architecture but skip governance — do not skip it.
What AI-900 Actually Tests
AI-900 validates that you understand artificial intelligence concepts and the Azure services that implement them. Five domains:
Domain 1: AI Workloads and Considerations (15–20%)
Types of AI tasks: prediction, anomaly detection, computer vision, NLP, generative AI. Responsible AI principles: fairness, reliability, privacy, inclusiveness, transparency, accountability. You need to understand what AI does before studying how Azure does it.
Domain 2: Machine Learning on Azure (15–20%)
Core ML concepts: features, labels, training vs validation vs test data, regression vs classification vs clustering. Azure Machine Learning workspace — automated ML, designer, and model deployment. The exam tests conceptual understanding, not hands-on ML engineering.
Domain 3: Computer Vision on Azure (15–20%)
Azure AI Vision: image classification, object detection, OCR, face detection. Azure AI Video Indexer. Custom Vision for training custom image classifiers. Key question format: given a computer vision task, identify the correct Azure AI service.
Domain 4: NLP on Azure (15–20%)
Azure AI Language: sentiment analysis, key phrase extraction, entity recognition, language detection, Q&A, text summarization. Azure AI Speech: speech-to-text, text-to-speech, translation, speaker recognition. Azure AI Translator. Match the task to the service.
Domain 5: Generative AI on Azure (20–25%)
The highest-weighted domain and most recently updated. Azure OpenAI Service, large language models, prompt engineering basics, embeddings, Copilot in Microsoft products, responsible AI for generative systems, and content safety. Candidates using resources older than 12 months will be under-prepared for this domain.
Which Exam to Take First: By Role
Cloud / IT Administrator
AZ-900 firstYour day-to-day work involves Azure resources, subscriptions, policies, and governance. AZ-900 validates the foundational knowledge you already apply and opens the path to AZ-104. AI-900 is a useful addition later, but not your priority.
Software Developer
AZ-900 firstAzure architecture, compute services, and storage are directly relevant to what you build. AZ-900 leads to AZ-204, the developer-focused Azure certification. AI-900 becomes relevant when your work specifically involves Azure AI or Azure OpenAI integrations.
Business Analyst / Product Manager
AI-900 firstYour stakeholders are asking about AI capabilities, and AI-900 gives you the vocabulary, framework, and credibility to participate in those conversations. AZ-900's networking and infrastructure content is less immediately relevant to your role.
Data Professional (Analyst, Data Scientist, BI Developer)
AI-900 firstThe ML concepts, model types, and Azure AI service landscape in AI-900 are directly applicable to your work. Follow it with DP-900 (Azure Data Fundamentals) or DP-100 (Azure Data Scientist Associate) depending on your technical depth.
Complete Beginner to Microsoft Cloud
AZ-900 firstAZ-900 provides the cloud foundation — IaaS/PaaS/SaaS, Azure architecture, billing, governance — that makes every subsequent Microsoft certification more accessible. AI-900 assumes cloud context even though it does not formally test it.
Security / Compliance Professional
AZ-900, then SC-900Neither exam alone covers security in depth. AZ-900 gives you Azure context; SC-900 gives you the security-specific credential. AI-900 becomes relevant when your work intersects with AI governance.
Student / Recent Graduate
AZ-900 firstAZ-900 is the most recognized Microsoft fundamentals certification by employers, appears most frequently in job postings, and provides the broadest career optionality. Adding AI-900 as a second credential before entering the job market is a strong strategy.
Career Path Implications
AZ-900 Career Path
AZ-900 → Azure Fundamentals
→ AZ-104 (Azure Administrator) — cloud ops, infrastructure
→ AZ-204 (Azure Developer) — developer and DevOps roles
→ AZ-305 (Azure Solutions Architect Expert) — architecture
→ AZ-700 (Azure Network Engineer) — networking specialization
Salary Ranges
Azure Administrator: $90K–$130K
Solutions Architect: $130K–$180K
Cloud Engineer: $100K–$145K
AI-900 Career Path
AI-900 → Azure AI Fundamentals
→ AI-102 (Azure AI Engineer Associate) — AI app development
→ DP-100 (Azure Data Scientist Associate) — ML and data science
→ DP-203 (Azure Data Engineer Associate) — data pipelines
Salary Ranges
Azure AI Engineer: $120K–$165K
Data Scientist: $110K–$155K
ML Engineer: $130K–$175K
Content Overlap Between AZ-900 and AI-900
Less than you might expect — roughly 10–15%. Both exams reference Azure as a cloud platform, and AI-900 assumes basic cloud familiarity. They do not test the same domain content.
If you take AZ-900 first, the cloud concepts you learn provide useful context for AI-900's Azure AI service content. If you take AI-900 first, you will need to study AZ-900 from scratch — AI-900 does not prepare you for AZ-900's architecture, networking, and governance domains.
The minimal overlap means taking both does not involve significant duplicate study effort. Most candidates spend 3–4 weeks total to complete both certifications when taken sequentially.
Frequently Asked Questions: AZ-900 vs AI-900
What is the difference between AZ-900 and AI-900?
AZ-900 (Azure Fundamentals) tests knowledge of cloud computing concepts and the Azure platform — architecture, services, management, and governance. AI-900 (Azure AI Fundamentals) tests knowledge of AI concepts and the Azure AI services that implement them — machine learning, computer vision, NLP, and generative AI. Both cost $99, have 40–60 questions, take 60 minutes, and require 700/1000 to pass.
Should I take AZ-900 before AI-900?
For most people, yes. AZ-900 builds the cloud foundation that makes Azure AI service concepts in AI-900 easier to understand. It also unlocks a broader set of follow-on certifications. The exception: if you work in a business or analytical role where AI literacy is immediately relevant and Azure infrastructure is not, AI-900 first is a reasonable choice.
Is AI-900 easier than AZ-900?
Comparable difficulty, but different challenge types. AZ-900 has more content volume — three domains covering architecture, services, and governance. AI-900 has more unfamiliar terminology for non-technical candidates — ML concepts, AI service distinctions, and generative AI. Study time is typically slightly shorter for AI-900 (1–2 weeks vs 2–3 weeks), but that reflects content volume, not question difficulty.
Do AZ-900 and AI-900 overlap?
Minimal overlap — approximately 10–15%. Both exams reference the Azure cloud platform, and AI-900 assumes basic cloud familiarity. They do not test the same domain content. Taking both does not involve significant duplicate study effort.
How much does each exam cost?
Both AZ-900 and AI-900 cost $99 USD. Microsoft occasionally offers free exam vouchers through Microsoft Learn challenges, virtual training days, and partner programs. Check the Microsoft Learn events page before paying full price.
Which looks better on a resume: AZ-900 or AI-900?
It depends on the role. For cloud, infrastructure, and developer positions, AZ-900 is more recognized and more frequently required. For AI, ML, data, and business roles, AI-900 signals directly relevant expertise. Having both is stronger than either alone.
Can I pass AZ-900 or AI-900 with just Microsoft Learn?
Microsoft Learn covers the content accurately, but the real exams test application of knowledge under scenario-based question formats that Microsoft Learn's exercises do not fully prepare you for. Candidates who supplement with practice exams — particularly timed, full-length ones — pass at significantly higher rates. Start with Microsoft Learn, then practice under exam conditions before scheduling.
Practice AZ-900
500 questions. 40 free. Updated for 2026 exam objectives.
Start Free AZ-900 Practice →Practice AI-900
500 questions. 40 free. Generative AI domain fully updated.
Start Free AI-900 Practice →