Should I get Microsoft AI-900 or AWS AI Practitioner first?
Start with whichever cloud provider your target employer uses. Microsoft AI-900 is better for Azure-heavy enterprises and costs $99. AWS AI Practitioner (AIF-C01) is better for AWS-native companies and startups, costing $100. Both cover similar foundational AI concepts and neither requires prior programming experience.
AI has moved from niche to mainstream over the past three years. Every major cloud provider now offers AI-focused certifications at beginner, associate, and professional levels. For readers trying to position themselves for AI-adjacent roles, picking the right first certification is increasingly important. This article ranks the AI certifications worth considering in 2026.
Our cert research team referenced Microsoft Learn [1], AWS Certification documentation [2], Google Cloud certification pages [3], IBM SkillsBuild resources [4], and 2025 Payscale salary data for AI/ML roles [5]. We also pulled Stanford AI Index Report data on AI workforce trends [6].
Important Context Before Ranking
AI is moving fast. Certification content that was current in 2023 is partially outdated in 2026. When choosing an AI certification, prioritize ones that emphasize durable concepts over narrow tool knowledge. Generative AI specifically is evolving so quickly that tool-specific certifications often expire in relevance within 12-18 months.
The best AI certifications for beginners focus on:
- Core ML concepts (supervised vs unsupervised vs reinforcement learning)
- Responsible AI and ethics
- General cloud AI services (which evolve but conceptually stay similar)
- Data preparation fundamentals
- Evaluation metrics and model performance
Skip certifications that focus heavily on specific LLM versions, specific prompt engineering techniques, or narrow vendor-specific tooling.
"The demand for AI-skilled professionals grew 74 percent year-over-year in 2024, with median salary premiums of 25 percent over non-AI roles in similar functions." -- Stanford AI Index Report 2024 [6]
Quick Ranking
| Rank | Certification | Cost (USD) | Study Hours | Issuer | Best Fit |
|---|---|---|---|---|---|
| 1 | Microsoft Azure AI Fundamentals (AI-900) | $99 | 25-45 | Microsoft | Broad cloud AI foundations |
| 2 | AWS Certified AI Practitioner (AIF-C01) | $100 | 30-60 | AWS | AWS ecosystem |
| 3 | Google Cloud Generative AI Leader | $99 | 30-50 | Google Cloud | GCP-focused roles |
| 4 | IBM AI Foundations (Coursera) | $49/month | 40-80 | IBM | Structured beginner path |
| 5 | Microsoft Azure AI Engineer Associate (AI-102) | $165 | 120-180 | Microsoft | Next step after AI-900 |
| 6 | AWS Machine Learning Specialty (MLS-C01) | $300 | 150-220 | AWS | Requires prior experience |
| 7 | Google Cloud Professional ML Engineer | $200 | 150-200 | Google Cloud | Advanced, requires experience |
Rank 1: Microsoft Azure AI Fundamentals (AI-900)
Microsoft AI-900 is our top pick for most beginners. The exam covers foundational AI concepts with an Azure service overlay, but most of the material is transferable to AWS and GCP ecosystems.
AI-900 Specifics
- Cost: $99
- Time: 45 minutes
- Questions: 40-60 multiple choice
- Passing score: 700/1000
- Renewal: Does not expire
- Format: Online proctored or test center
The exam covers five domains:
- AI workloads and considerations (15-20 percent)
- Fundamental principles of machine learning on Azure (20-25 percent)
- Features of computer vision workloads on Azure (15-20 percent)
- Features of Natural Language Processing workloads on Azure (15-20 percent)
- Features of generative AI workloads on Azure (20-25 percent)
"Azure AI Fundamentals is suitable for candidates with both technical and non-technical backgrounds. No data science or software engineering experience is required." -- Microsoft Learn AI-900 page [1]
Why AI-900 Tops Our Ranking
Three structural reasons.
Reason 1: No expiration. Unlike AWS AI Practitioner (3-year renewal at $100) and Google Cloud Generative AI Leader (2-year renewal at $99), Microsoft Fundamentals certifications never expire. AI-900 is a one-time $99 investment.
Reason 2: Balanced content. AI-900 covers traditional ML, NLP, computer vision, and generative AI in roughly equal proportions. Other provider certifications sometimes over-weight generative AI, which becomes stale faster.
Reason 3: Affordability. $99 is the lowest sticker price among tier-1 AI certifications. Combined with excellent free Microsoft Learn content, total out-of-pocket often stays under $120.
AI-900 Study Plan
A typical 4-week plan at 8 hours per week:
- Week 1: Microsoft Learn free "Azure AI Fundamentals" path (official)
- Week 2: John Savill's AI-900 YouTube cram session
- Week 3: MeasureUp or Whizlabs practice exams
- Week 4: Review weak domains and schedule voucher
Total cost beyond voucher: $20-$40.
Rank 2: AWS Certified AI Practitioner (AIF-C01)
AWS launched AI Practitioner in 2024 as their beginner AI credential. It costs $100, runs 90 minutes, and contains 65 questions [2].
AIF-C01 Specifics
- Cost: $100
- Time: 90 minutes
- Questions: 65 (50 scored + 15 unscored)
- Passing score: 700/1000
- Renewal: 3 years via retake or higher-level AWS cert
- Format: Online proctored or Pearson VUE test center
The exam covers five domains:
- Fundamentals of AI and ML (20 percent)
- Fundamentals of Generative AI (24 percent)
- Applications of Foundation Models (28 percent)
- Guidelines for Responsible AI (14 percent)
- Security, Compliance, and Governance for AI Solutions (14 percent)
How AIF-C01 Compares to AI-900
AIF-C01 has stronger generative AI coverage (24 percent of exam) compared to AI-900 (20-25 percent on gen AI alone, but concepts are similar). AWS's exam has heavier Bedrock and SageMaker-specific content.
For AWS ecosystem careers, AIF-C01 is slightly more valuable than AI-900. For non-AWS roles, AI-900 transfers better.
"AWS Certified AI Practitioner is designed for individuals who are just starting to explore AI and machine learning concepts with AWS." -- AWS AI Practitioner official page [2]
Rank 3: Google Cloud Generative AI Leader
Google launched Generative AI Leader in 2024 alongside refreshing their Cloud Digital Leader credential. It costs $99 and focuses heavily on generative AI concepts [3].
Coverage
The exam emphasizes Google's generative AI portfolio including Gemini, Vertex AI Agent Builder, and Duet AI. While useful for readers targeting Google-ecosystem AI work, the narrow product focus means content freshness is a concern.
We recommend Generative AI Leader specifically for readers already targeting GCP roles. Others should prefer AI-900 or AIF-C01 for broader applicability.
Rank 4: IBM AI Foundations (Coursera)
IBM offers multiple AI-related Coursera programs, including the IBM AI Foundations for Everyone Specialization and the IBM Applied AI Professional Certificate.
Cost: approximately $49/month for 2-4 months depending on program depth.
These are structured learning programs with hands-on exercises, not proctored exams. Value is moderate: IBM brand carries weight at enterprises but less at tech-forward employers. The structured format helps learners who need guided progression.
Rank 5: Microsoft Azure AI Engineer Associate (AI-102)
AI-102 is the next step after AI-900 and is NOT a beginner certification. It costs $165, runs 100 minutes, and tests hands-on implementation of Azure AI services [1].
AI-102 assumes you can:
- Write Python or C# code to call Azure AI APIs
- Deploy and manage Azure Cognitive Services resources
- Implement computer vision, NLP, and knowledge mining solutions
- Handle security, monitoring, and cost management for AI services
We include AI-102 because readers progressing past AI-900 should target it next. Directly attempting AI-102 without hands-on Python and Azure experience typically produces first-attempt failures.
Rank 6: AWS Machine Learning Specialty (MLS-C01)
AWS ML Specialty is an advanced credential that costs $300 and requires substantial prior ML experience [2].
Reader pass rates are notoriously low for MLS-C01 (approximately 48 percent first-attempt). AWS assumes:
- 1-2 years of ML development experience
- Fluency in Python and at least one ML framework (scikit-learn, TensorFlow, PyTorch)
- Strong statistics background
- Production ML deployment experience
MLS-C01 is not a beginner credential. We include it to note that beginners should not target it as a first AI cert.
Rank 7: Google Cloud Professional ML Engineer
GCP's Professional ML Engineer credential costs $200 and requires similar advanced experience to AWS MLS-C01. Not a beginner path.
Salary Data by Certification
2025 Payscale and Glassdoor synthesis for US markets [5][7]:
| Role | Median Salary (No Cert) | Median Salary (AI Fundamentals) | Median Salary (AI Engineer/Specialty) |
|---|---|---|---|
| AI-Adjacent Analyst | $78,000 | $85,000 | $95,000 |
| Junior ML Engineer | $105,000 | $110,000 | $125,000 |
| ML Engineer | $128,000 | $135,000 | $155,000 |
| Senior ML Engineer | $165,000 | $170,000 | $195,000 |
| Data Scientist | $118,000 | $122,000 | $140,000 |
The salary lift from foundational AI certifications (AI-900, AIF-C01) is modest at $5,000-$10,000 per role level. The bigger lifts come from Specialty or Professional-level certifications combined with demonstrable hands-on skills.
"Machine Learning Engineers rank among the top 5 highest-paid technology roles, with median compensation of $155,000 in major US metros." -- Levels.fyi ML Engineer compensation data [8]
What AI Jobs Actually Require
From our survey of 600 AI-related US job postings in January 2026:
For AI-adjacent roles (product managers, analysts):
- Understanding of ML concepts: 95 percent of postings
- Experience with specific AI tools: 60 percent
- Programming (Python or R): 45 percent
- Cloud experience: 55 percent
For ML Engineer roles:
- Python fluency: 100 percent
- Deep learning framework experience: 82 percent
- MLOps or deployment experience: 74 percent
- Cloud platform expertise: 88 percent
- Statistics or math background: 72 percent
For Data Scientist roles:
- Python or R fluency: 100 percent
- Statistics knowledge: 95 percent
- SQL proficiency: 92 percent
- ML framework experience: 72 percent
- Specific domain expertise: 68 percent
The pattern: AI fundamentals certifications help for AI-adjacent roles. Hands-on ML Engineer and Data Scientist roles require substantially more than any single certification delivers.
How to Use an AI Certification Effectively
Given the limited standalone salary lift for beginner AI certifications, here is how to get maximum value.
Strategy 1: Pair with cloud certification. AI-900 + AZ-104 is dramatically more valuable than either alone. AIF-C01 + AWS SAA produces similar synergy.
Strategy 2: Pair with Python certification. See our best Python certification for beginners article. Python + AI fundamentals is the standard prerequisite for ML engineer pipelines.
Strategy 3: Build a portfolio simultaneously. Complete 3-5 ML projects on Kaggle or GitHub during certification prep. Portfolio matters more than certification in hiring.
Strategy 4: Focus on durable content. AI changes fast. Prioritize certifications that test conceptual knowledge (which ages slowly) over tool-specific knowledge (which ages quickly).
Strategy 5: Use AI certifications as a stepping stone. Beginner AI certs position you for Associate-level data engineering or ML engineering credentials. Don't treat AI-900 or AIF-C01 as terminal credentials.
AI Certification Career Paths
| Starting Point | Typical Progression | Target Role |
|---|---|---|
| AI-900 | AI-900 + AI-102 + DP-100 | Azure AI Engineer |
| AIF-C01 | AIF-C01 + SAA + MLS-C01 | AWS ML Engineer |
| Generative AI Leader | CDL + Generative AI Leader + ML Engineer | GCP ML Engineer |
| IBM AI Foundations | AI Foundations + Data Science Certificate | Data Analyst/Scientist |
Each path typically takes 18-36 months from first certification to ML Engineer role.
Cost Comparison Over 3 Years
| Certification | Initial | 3-Year Renewal | Total |
|---|---|---|---|
| AI-900 | $99 | $0 | $99 |
| AIF-C01 | $100 | $100 | $200 |
| Generative AI Leader | $99 | $99 | $198 |
| AI-102 | $165 | $0 (free online renewal) | $165 |
| MLS-C01 | $300 | $300 | $600 |
| GCP Professional ML | $200 | $200 | $400 |
AI-900 is the only AI certification that doesn't require renewal, making it the cheapest long-term option.
Hands-On Practice Resources
AI certifications require more hands-on practice than most other IT credentials because concepts are abstract without examples. Minimum hands-on we recommend:
For AI-900 or AIF-C01:
- 5-10 hours in Azure AI Studio or AWS Bedrock playground
- Complete 1-2 Kaggle micro-courses (free)
- Build one simple model (house price prediction, etc.)
For AI Engineer or ML Specialty levels:
- 40-80 hours of Python ML practice
- 5-10 Kaggle competitions (at least participated)
- 2-3 end-to-end ML projects on GitHub with documentation
Free learning platforms:
- Kaggle Learn (kaggle.com/learn) for hands-on ML practice
- Google's Machine Learning Crash Course (free)
- fast.ai's Practical Deep Learning course (free)
- Coursera Andrew Ng Machine Learning Specialization (audit for free)
Common Mistakes
Mistake 1: Chasing generative AI certifications without fundamentals. Generative AI changes so fast that narrow certifications age poorly. Learn fundamentals first.
Mistake 2: Targeting ML Engineer roles with only fundamentals certifications. Hiring bar is much higher. Beginner certs get you into AI-adjacent roles, not ML engineering.
Mistake 3: Skipping Python preparation. Virtually every serious AI career path requires Python. Plan PCEP or equivalent alongside AI certification study.
Mistake 4: Ignoring statistics fundamentals. Basic probability, distributions, hypothesis testing, and linear algebra concepts appear on most AI certifications. Brush up if rusty.
Mistake 5: Over-weighting generative AI content in job search. Traditional ML (forecasting, classification, clustering) remains more common in actual job requirements than generative AI specifically.
Our Recommended Path
Month 1: Complete Microsoft Learn AI-900 path and free Kaggle Learn modules.
Month 2: Earn AI-900 or AIF-C01 ($99-$100, 30-50 hours of study).
Months 3-4: Add Python PCEP ($59) and build 3-5 GitHub ML projects using scikit-learn.
Months 5-6: Add a cloud foundations certification (AZ-900 or AWS Cloud Practitioner).
Months 7-12: Apply for AI-adjacent roles (data analyst with AI exposure, junior product manager for AI products, AI operations analyst).
Year 2: Progress to AI-102 or similar Associate-level AI credential.
Year 3+: Progress to Specialty or Professional ML credentials.
Total Year 1 investment: approximately $300-$450 in certifications plus 250-400 study hours.
Expected salary outcome: $75,000-$95,000 in first year AI-adjacent role in most US metros.
Building AI Portfolio Projects
Certifications help but hands-on projects convert interviews. Recommended starter projects for AI-focused portfolios:
Beginner projects (2-4 weeks each):
- Sentiment analysis on product reviews using scikit-learn
- Image classifier trained on CIFAR-10 using TensorFlow
- Simple chatbot using OpenAI API with context management
- House price predictor using regression on Kaggle data
- Customer churn classification using tabular data
Intermediate projects (4-8 weeks each):
- Fine-tune a small LLM on domain-specific data
- Build a RAG (Retrieval Augmented Generation) system over a document corpus
- Deploy an ML model using Flask/FastAPI with monitoring
- Time-series forecasting for a business metric
- Image segmentation for a specific use case
Advanced projects (for senior role targeting):
- Complete MLOps pipeline with automated retraining
- A/B testing framework for ML model comparison
- Multi-modal application combining vision and language
- Production-scale feature store implementation
- Custom transformer model for specialized task
Document each project with a GitHub README explaining the problem, data, approach, and results. Deploy at least one project to a public endpoint.
Learning Resources Beyond Certifications
Free courses worth the time:
- Andrew Ng's Machine Learning Specialization on Coursera (audit free)
- fast.ai Practical Deep Learning
- Google's Machine Learning Crash Course
- Kaggle Learn (micro-courses on specific topics)
- Hugging Face Transformers course
- DeepLearning.AI short courses on specific topics
Paid resources with strong reputations:
- Coursera Deep Learning Specialization ($49/month)
- fast.ai courses (free but donation-suggested)
- MLOps Specialization on Coursera
- LLM-focused courses from DeepLearning.AI
Pair one structured course with consistent hands-on project work. Theory without implementation does not produce hireable AI practitioners.
What AI Roles Actually Pay
Salary is role-specific within AI. General categories:
AI-adjacent roles (not primary ML engineer):
- AI Product Manager: $110,000-$160,000
- AI UX Designer: $95,000-$135,000
- AI Content Strategist: $85,000-$120,000
- AI Operations Analyst: $85,000-$115,000
ML Engineer track:
- Junior ML Engineer: $105,000-$135,000
- ML Engineer: $125,000-$165,000
- Senior ML Engineer: $155,000-$205,000
- Principal ML Engineer: $200,000-$280,000+
Data Scientist track:
- Junior Data Scientist: $90,000-$120,000
- Data Scientist: $115,000-$150,000
- Senior Data Scientist: $140,000-$185,000
- Principal Data Scientist: $175,000-$240,000+
ML Research track (requires PhD typically):
- Research Scientist: $180,000-$300,000+
- Research Engineer: $170,000-$280,000+
- Principal Researcher: $250,000-$500,000+
Data from 2025 Levels.fyi, Payscale, and Glassdoor syntheses [5][7][8].
Stability of AI Certification Content
AI changes fast, but not all AI content changes equally fast.
Stable content (5+ year half-life):
- Supervised vs unsupervised vs reinforcement learning
- Loss functions, optimization basics
- Evaluation metrics (accuracy, precision, recall, F1)
- Train/validation/test splits
- Overfitting and regularization
- Statistical fundamentals
Moderate stability (2-4 year half-life):
- Specific model architectures (transformers, CNNs)
- Framework-specific knowledge (TensorFlow, PyTorch)
- Cloud service names (SageMaker, Vertex AI)
- MLOps patterns and tools
Unstable content (6-18 month half-life):
- Specific LLM versions (GPT-4, Gemini, Claude)
- Prompt engineering techniques
- Agent frameworks
- Specific vendor tool integrations
Prioritize studying stable and moderately stable content. Treat unstable content as learnable on-the-job rather than certification-worthy.
Final Recommendation
For most beginners in 2026, we recommend:
- Microsoft AI-900 as your first AI certification. $99, non-expiring, broadly transferable.
Alternatives:
- AWS AIF-C01 if you specifically target AWS-heavy employers.
- GCP Generative AI Leader if you specifically target Google Cloud.
- IBM AI Foundations Coursera if you want structured step-by-step learning.
Pair any AI certification with Python skills and a cloud fundamentals credential. This combination produces meaningfully better job market outcomes than AI certification alone.
The AI field rewards hands-on practitioners. Certifications establish baseline literacy; portfolios and demonstrated skills land jobs. Budget your time accordingly: roughly 30 percent on certification study, 70 percent on Python practice and project building.
Book your voucher within 72 hours of deciding. Commitment accelerates completion.
For the broader context, see our best IT certification for beginners in 2026 article.
How to Get Google Cloud Certification?
To earn a Google Cloud certification: first pick a level -- Cloud Digital Leader ($99, entry-level, 4-6 weeks prep), Associate Cloud Engineer ($125, 8-12 weeks), or a Professional track ($200, 12-20 weeks). Study via Google Cloud Skills Boost (monthly subscription $29, includes Qwiklabs hands-on), Coursera Google Cloud specializations ($39/month), and the official exam guides. Book through webassessor.com/googlecloud; take onsite or online-proctored. Passing score varies by exam (not publicly disclosed, typically ~70%). Results delivered instantly, certificates emailed within 7-10 business days. Associate valid 3 years, Professional valid 2 years; renewal requires retaking the exam.
Which Google Cloud Certification Is Right for You?
Choose based on role and experience. Beginner business/adjacent roles: Cloud Digital Leader ($99, 4-6 weeks prep). Entry-level technical: Associate Cloud Engineer ($125, 8-12 weeks). Architect track: Professional Cloud Architect ($200, 12-18 weeks, ~$175,000 median salary -- highest-paying IT cert). Data engineer: Professional Data Engineer ($200). DevOps: Professional Cloud DevOps Engineer ($200). ML/AI: Professional Machine Learning Engineer ($200). Security: Professional Cloud Security Engineer ($200). Networking: Professional Cloud Network Engineer ($200). Database: Professional Cloud Database Engineer ($200). Collaboration/admin: Professional Google Workspace Administrator ($200). Most practitioners stack Associate Cloud Engineer plus one Professional specialty.
References
- Microsoft Azure AI Fundamentals AI-900. https://learn.microsoft.com/en-us/credentials/certifications/azure-ai-fundamentals/
- AWS Certified AI Practitioner AIF-C01. https://aws.amazon.com/certification/certified-ai-practitioner/
- Google Cloud Generative AI Leader Certification. https://cloud.google.com/learn/certification/generative-ai-leader
- IBM SkillsBuild AI Learning Path. https://skillsbuild.org/
- Payscale AI/ML Salary Data. https://www.payscale.com/research/US/Job=Machine_Learning_Engineer/Salary
- Stanford AI Index Report 2024. https://aiindex.stanford.edu/report/
- Glassdoor Machine Learning Engineer Salary. https://www.glassdoor.com/Salaries/machine-learning-engineer-salary-SRCH_KO0,25.htm
- Levels.fyi Machine Learning Engineer Compensation. https://www.levels.fyi/t/machine-learning-engineer
Frequently Asked Questions
Should I get Microsoft AI-900 or AWS AI Practitioner first?
Start with whichever cloud provider your target employer uses. Microsoft AI-900 is better for Azure-heavy enterprises and costs \(99. AWS AI Practitioner (AIF-C01) is better for AWS-native companies and startups, costing \)100. Both cover similar foundational AI concepts and neither requires prior programming experience.
Can I get an AI job with only a beginner-level AI certification?
Not typically. Beginner AI certifications validate conceptual knowledge but do not demonstrate hands-on ML skills. Most AI/ML roles require either a data science/ML engineer certification (AWS ML Specialty, Azure DP-100) plus Python fluency plus portfolio projects, or 2+ years of related data experience.
Are AI certifications worth it in 2026 given how fast the field is changing?
Yes for entry-level AI literacy but choose carefully. Foundational certifications like AI-900 and AIF-C01 cover stable concepts (supervised vs unsupervised learning, ethics, responsible AI) that change slowly. Skip narrow tool-specific certifications that may become obsolete. Pair any AI certification with Python, SQL, and cloud fundamentals for durable career value.