AIF-C01 vs AI-901: AWS AI Practitioner or Azure AI Fundamentals?
By Kyle Tice · Updated October 19, 2026 · Facts checked against official exam pages
Pick AIF-C01 if you work in an AWS shop, or if you want an AI credential that tests understanding rather than code. Pick AI-901 if your employer runs on Azure and you are comfortable reading basic Python, because the 2026 version of Microsoft’s fundamentals exam is no longer a pure concepts test. If you have no cloud preference at all, AIF-C01 is the safer default: it asks less of you technically and scores higher on employer demand in our rankings.
| AIF-C01 | AI-901 | |
|---|---|---|
| Provider | AWS | Microsoft |
| Code | AIF-C01 | AI-901 |
| Level | foundational | foundational |
| Cost | $100 | $99 |
| Format | Proctored exam | Proctored exam |
| Questions | 65 | — |
| Duration | 90 min | 45 min |
| Validity | 3 yrs | No expiry |
| Prerequisites | None | None |
| Study time | 20–40 h | 15–30 h |
| Difficulty | 2/5 | 2/5 |
| Employer demand | 4/5 | 3/5 |
| Atlas score | 85/100 | 75/100 |
| Status | active | active |
The big change: AI-901 is not AI-900
Most comparisons online still pit AIF-C01 against AI-900. That exam retired on June 30, 2026. Its replacement, AI-901, changed the deal.
Microsoft’s exam page now says you need “knowledge of Python coding syntax and programming techniques” and familiarity with Azure resources. The skills outline splits into two parts:
| AI-901 area | Weight |
|---|---|
| Identify AI concepts and capabilities | 40–45% |
| Implement AI solutions by using Microsoft Foundry | 55–60% |
More than half of the exam is about building things in Microsoft Foundry. AI-900 was something a salesperson could pass after a weekend. AI-901 is aimed at people “at the beginning of their career in AI solution development.”
AIF-C01 went the other way. The AWS exam guide explicitly lists developing or coding AI/ML models, feature engineering, hyperparameter tuning and building pipelines as out of scope. The target candidate “uses but does not necessarily build” AI solutions on AWS.
So the real question is not AWS versus Azure. It is whether you want a literacy credential (AIF-C01) or a first step toward a builder role (AI-901).
Content and domains
AIF-C01 has five weighted domains:
| AIF-C01 domain | Weight |
|---|---|
| Fundamentals of AI and ML | 20% |
| Fundamentals of GenAI | 24% |
| Applications of foundation models | 28% |
| Guidelines for responsible AI | 14% |
| Security, compliance and governance for AI | 14% |
Notice that 28% of AIF-C01 is about applying foundation models, mostly through Amazon Bedrock: prompt engineering, RAG at a conceptual level, fine-tuning trade-offs and evaluation. Another 28% combined goes to responsible AI and governance. That governance slice is larger than you would expect from an entry exam, and it makes AIF-C01 useful for compliance-adjacent roles.
AI-901’s concept half covers familiar ground: AI workloads, machine learning principles, computer vision, natural language processing and generative AI. The implementation half expects you to recognise how to deploy and call models in Foundry, work with SDKs and REST APIs, and wire up simple solutions. Microsoft notes the exam may include preview features if they are commonly used, which means study material goes stale quickly.
Difficulty
Both sit at 2 out of 5 on our difficulty scale, but they are hard in different ways.
- AIF-C01 is wide. You must tell apart a lot of AWS services (Bedrock, SageMaker AI, Comprehend, Rekognition, Textract and more) and pick the right one for a scenario. AWS also uses ordering and matching questions, not only multiple choice. The exam has 65 questions, 50 scored and 15 unscored, in 90 minutes, and you need 700 out of 1,000.
- AI-901 is narrower on services but deeper on hands-on detail. If you have never written Python, the implementation questions will feel unfamiliar even when the concept is easy. You need 700 to pass.
If you are non-technical, AIF-C01 is clearly easier. If you are a junior developer, the gap closes and AI-901 may feel more concrete.
Cost and validity
As of October 2026:
| AIF-C01 | AI-901 | |
|---|---|---|
| Exam fee (US) | $100 | $99 (varies by country) |
| Validity | 3 years | Does not expire |
| Renewal | Retake, or earn ML Engineer – Associate | None needed |
| Cost per year of validity | About $33 | One-time |
Microsoft Fundamentals certifications do not expire, so AI-901 is a one-time $99. AIF-C01 costs about $33 per year over its three-year life. A neat detail: AWS lets you renew AIF-C01 by passing the ML Engineer – Associate, so if you plan to move up, renewal is free in practice.
The non-expiring credential sounds like a win for Microsoft, but think about what it signals. An AI fundamentals pass from 2026 will look dated by 2029 whether or not it technically expires. Check /vouchers/ before you book; both vendors run discount campaigns, though the 2026 AWS AIF-C01 promo has closed.
Employer recognition
We rate AIF-C01 at 4 out of 5 for employer demand and AI-901 at 3. The reasons:
- AWS is the most widely used public cloud, and “AWS certified” is a common recruiter filter. AIF-C01 also signals Bedrock awareness, which is what AWS customers are buying.
- AI-901 is new. Job postings in 2026 still name AI-900 or AI-102, and recruiters will not always know that AI-901 replaced AI-900. The upside is that AI-901 now proves more, so a hiring manager who reads the outline will rate it higher than its predecessor.
Neither cert gets you hired on its own. For non-technical roles, they show you can talk sensibly about AI. For technical roles, a small project on GitHub beats either one. See what employers want for detail.
Study effort
- AIF-C01: 20–40 hours. AWS Skill Builder has a free exam prep plan and an official practice question set. Many candidates add a paid practice set such as Tutorials Dojo’s AIF-C01 practice exams.
- AI-901: 15–30 hours if you already read Python, more if you do not. Microsoft Learn’s study guide and free practice assessment (now on AI Skills Navigator) are the core. Spend real time in a Foundry project; reading alone will not cover the implementation half.
Our AWS study guide and Microsoft study guide list resources in order.
Who should take both
Very few people. Taking both makes sense if:
- You are a consultant or pre-sales engineer who works across AWS and Azure clients.
- Your company runs multi-cloud and wants staff fluent in both Bedrock and Foundry.
For everyone else, the second fundamentals cert adds little. Put that time toward an associate exam instead: AI-103 on Azure or the ML Engineer – Associate on AWS.
Who should skip both
If you already build AI applications, skip straight to an associate or professional exam. And if your goal is a job in AI engineering, a portfolio of two or three working projects will do more than any fundamentals badge. Our portfolio projects guide has ideas.
FAQ
Is AI-900 still worth taking?
You cannot. AI-900 retired on June 30, 2026. If you already hold it, it remains valid because Fundamentals certs do not expire, but new candidates must take AI-901.
Do I need to code for AIF-C01?
No. AWS lists coding models and building pipelines as out of scope. You need to understand what services do and when to use them.
Which is better for a complete beginner?
AIF-C01, unless you specifically need Azure. AI-901 now assumes some Python, and beginners with no coding background will find the Foundry questions harder.
Can AIF-C01 renew automatically?
Not automatically, but passing the AWS ML Engineer – Associate also recertifies AIF-C01, according to AWS. Otherwise, retake the current exam within three years.
Still unsure? Try our cert picker or see how both rank on /rankings/.