Cloud certifications: AWS, Azure or Google Cloud — which to choose in 2026

• Why do cloud certifications still matter?
• AWS, Azure or Google Cloud: how do you decide?
• What are each provider's tracks?
• How do you study efficiently for the exam?
• When is a certification NOT the best investment?

The question "which cloud certification should I take" almost never has a technical answer. It has a market answer: certify in the provider that shows up in the jobs you want, at the company where you already are, or in the ecosystem your clients use.

Why do cloud certifications still matter?

A cloud certification is a signal, not proof of competence — and that is precisely why it works. It solves a specific market problem: reducing uncertainty in screening. A recruiter facing two hundred résumés uses objective filters, and the certification is one of them.

The real value concentrates in three situations:

1. Career transitions. With no infrastructure track record, the certification is the available evidence that you know the vocabulary and mental model of cloud computing.
2. Provider partner companies. Consultancies and system integrators must maintain a minimum number of certified professionals to keep their partnership tier. In those shops, certification is a contractual requirement, not a preference.
3. Forced breadth. Studying for the exam pushes you through services you would never use day to day — and that broad sweep has genuine learning value.

What it does not do: replace experience. A multiple-choice exam measures recognition, not the ability to debug a misconfigured network at eleven at night. People who treat the certification as an end rather than as a study structure typically pass the exam and stall in the technical interview.

Practical example: two candidates compete for a junior cloud role. One holds the Solutions Architect Associate; the other has a public project using Terraform to provision a real application, with a README explaining the decisions. The first passes screening more easily; the second usually wins the interview. Whoever has both rarely loses.

AWS, Azure or Google Cloud: how do you decide?

The three providers dominate the market and the decision weighs more on your career than on the technology. An honest summary:

The decision method that works is simple and takes fifteen minutes: open three job boards, search for the roles you want, in your city or remote, and count how many mention each provider. That number is worth more than any technical comparison, including this one.

Two common shortcuts: if you already work at a company using a specific cloud, certify in that one — the studying immediately turns into better work. If you are targeting financial services or large industry in Brazil, Azure tends to appear more often than global market share would suggest, because of the Microsoft environment those companies already maintain.

Practical example: a support analyst at a bank planning to certify in AWS "because it is the biggest" redid the search on internal job postings: eighteen positions asked for Azure, two for AWS. He took the AZ-104 and changed areas within four months, without changing employers.

What are each provider's tracks?

All three follow the same level logic — foundational, intermediate, advanced and specialties:

AWS

• Cloud Practitioner (CLF-C02) — foundational, US$ 100. Concepts, core services, the shared responsibility model and cost model.
• Solutions Architect Associate (SAA-C03) — US$ 150. The most recognized certification on the market; covers designing resilient and cost-effective architectures.
• Developer Associate and SysOps Administrator Associate — US$ 150. Focused on people who write code in the cloud and those who operate it.
• Solutions Architect Professional and DevOps Engineer Professional — US$ 300. Advanced level, with long and ambiguous scenarios.
• Specialty — Security, Advanced Networking, Machine Learning — US$ 300.

Microsoft Azure

• AZ-900 (Azure Fundamentals) — the entry point, and the most accessible foundational exam of the three.
• AZ-104 (Administrator) — resource administration, identity and networking. The most requested in Brazilian infrastructure roles.
• AZ-204 (Developer) — building solutions on the platform.
• AZ-305 (Solutions Architect Expert) — architect level, requires AZ-104 as a prerequisite.
• AI-900, DP-900 and SC-900 — foundations of AI, data and security, short and useful as complements.

Google Cloud

• Cloud Digital Leader — foundational with a business angle.
• Associate Cloud Engineer — day-to-day operations and deployment.
• Professional Cloud Architect — historically one of the best-paid certifications in the industry.
• Professional Data Engineer — the most valued on the track for data professionals, largely because of BigQuery.

One cost detail many people miss: AWS grants a 50% discount voucher for your next exam after each pass. A Cloud Practitioner → Solutions Architect Associate → Solutions Architect Professional track comes out at US$325 instead of US$ 550, as long as you use the vouchers in sequence.

Practical example: someone who already administers servers and knows networking can skip the foundational certification and go straight to associate level — saving time and money. Someone coming from outside the field gains more by taking the foundational exam first, because it organizes vocabulary the rest of the studying assumes.

How do you study efficiently for the exam?

A method that works across all three tracks, in five steps:

1. Start with the official exam guide. Every certification has a document listing the domains and the weight of each. It is the index for your studying and prevents spending weeks on a topic worth 5% of the exam.
2. Use the free tier and build something. All three providers offer a free tier. Deploy an application, configure networking, break it and fix it. A service you never touched becomes a question you guess.
3. Practice tests only after the foundation. Taking practice tests too early teaches you to memorize answers. Study the domain, then test — and treat each mistake as a topic to review, not an item to memorize.
4. Study costs seriously. It is the most tested and most ignored topic. Understanding when to use each storage class or instance type appears in many questions and generates real value at work — the same FinOps reasoning.
5. Book the date before you feel ready. A fixed deadline focuses the studying. Without a date, preparation stretches indefinitely.

A frequent trap: studying only through video. Video courses create a comfortable feeling of progress without consolidating anything. Alternate with hands-on console work and reading the official documentation — which, for all three providers, is excellent and is where the questions come from.

Practical example: a candidate failed the SAA-C03 after watching forty hours of course material. On the second attempt, he spent half the time deploying architectures in the free tier — VPC, load balancer, auto scaling — and passed comfortably. The content was the same; what changed was the kind of contact with it.

When is a certification NOT the best investment?

Four scenarios where the money and the weeks pay off better elsewhere:

1. You already have proven cloud experience. For someone operating real environments for years, the certification adds little beyond the badge — and the same time invested in a specialization or open source contribution yields more.
2. You are at the start of learning to program. Certifying in cloud without a foundation in networking, Linux and logic produces a badge with nothing under it. Fundamentals first; the guide on how to study programming covers that base.
3. You want to be a product developer. For front-end or application development roles, a portfolio and language mastery weigh far more than an infrastructure certification.
4. You are collecting badges. Three foundational certifications across different providers signal scattered focus, not breadth. A coherent track tells a better story on a résumé.

It is also worth calibrating salary expectations. A certification usually helps you enter a band, not jump a level on its own — what the data on technology salaries in Brazil shows is that seniority and architectural command weigh more than any single credential.

Practical example: a professional with five years of cloud operations swapped the Solutions Architect Professional for three months studying Kubernetes in depth and took over the company's internal platform. No recruiter ever asked about the certification he did not take.

Conclusion

A cloud certification is a positioning tool with a real effect and clear limits: it opens screening, organizes a study plan and is a requirement at provider partner companies — but it does not prove practical competence and does not replace experience. The choice between AWS, Azure and Google Cloud should come from counting jobs in the market you are targeting, not from technical comparisons; and the most efficient track is usually foundational, then associate-level architecture, and only then specialization. Study with the free tier open, treat cost as a central topic and book the date before you feel ready. Above all, pair the badge with something you built and can explain — because in the technical interview it is always the project, not the credential, that carries the conversation.