AWS vs Azure vs Google Cloud
Summary: three players concentrate about two thirds of the worldwide market. AWS offers the widest catalogue, Azure dominates in companies already equipped with Microsoft, Google Cloud excels at data and Kubernetes. This lesson compares them honestly, presents the European players, and treats the sovereignty question.
1. Overview of the three giants
2. AWS in detail
3. Microsoft Azure in detail
Azure's decisive factor, rarely technical: many companies choose Azure because they already negotiate a global Microsoft contract including Office 365, Windows Server and SQL Server. Adding Azure to that agreement produces a substantial discount. It is a commercial decision before being a technical one, and it is perfectly rational.
4. Google Cloud in detail
Two Google Cloud services deserve to be known even if you do not use GCP: BigQuery, often described as having no true equivalent for analysing very large volumes, and Cloud Run, regularly cited as the simplest way to put a container into production on the internet.
5. The service correspondence table
This table is extremely useful in practice: it lets you transpose a skill from one provider to another.
| Need | AWS | Azure | Google Cloud |
|---|---|---|---|
| Virtual machines | EC2 | Virtual Machines | Compute Engine |
| Object storage | S3 | Blob Storage | Cloud Storage |
| Managed relational database | RDS / Aurora | Azure SQL / Database for PostgreSQL | Cloud SQL / AlloyDB |
| NoSQL database | DynamoDB | Cosmos DB | Firestore / Bigtable |
| Serverless (functions) | Lambda | Functions | Cloud Functions |
| Serverless containers | Fargate / App Runner | Container Apps | Cloud Run |
| Managed Kubernetes | EKS | AKS | GKE |
| Virtual private network | VPC | Virtual Network | VPC |
| Content delivery | CloudFront | Front Door / CDN | Cloud CDN |
| Identity management | IAM | Entra ID | Cloud IAM |
| Data warehouse | Redshift | Synapse Analytics | BigQuery |
| Message queues | SQS / SNS | Service Bus | Pub/Sub |
| Secrets | Secrets Manager | Key Vault | Secret Manager |
| Monitoring | CloudWatch | Monitor | Cloud Monitoring |
| Native infrastructure as code | CloudFormation | ARM / Bicep | Deployment Manager |
What this table reveals: the concepts are nearly identical from one provider to another. Someone who masters AWS becomes operational on Azure in a few weeks, not a few months. The fundamental skills — networking, identities, storage, scaling — are transferable.
Consequence for your career: do not worry about "choosing the right cloud to learn". Learn one seriously, the others will follow easily.
6. The European players
An honest arbitration about the European players: if your need boils down to virtual machines, storage and Kubernetes, Hetzner or Scaleway cost significantly less than AWS for an entirely comparable service. On the other hand, if you need twenty interconnected managed services, the catalogue gap becomes decisive again.
7. The sovereignty question
The point to remember, often misunderstood: sovereignty does not boil down to the geographical location of the datacenter. A company subject to American law can be compelled to hand over data stored in France. That is why the French state pushed the SecNumCloud qualification from ANSSI and joint ventures like Bleu (Microsoft + Orange + Capgemini) and S3ns (Google + Thales).
8. Vendor lock-in
A nuanced view of lock-in: it is not necessarily to be avoided. Using DynamoDB or BigQuery saves you considerable development time, and you may never need to migrate. Refusing every proprietary service "on principle" often amounts to rebuilding it yourself, worse and slower.
The right question is not "how to avoid dependence?" but "what migration cost am I willing to accept, in exchange for what productivity gain today?"
9. Multi-cloud — what to know
Direct recommendation: for the great majority of organizations, one main provider is the right choice. Use a second player only for a precise and justified need. Multi-cloud "as a precaution" costs dearly and protects from nothing in practice.
10. How to choose — synthesis
The most decisive criterion in reality, and the most often neglected: your team's skills. A team that already masters Azure will be more effective on Azure than on AWS, even if AWS is "objectively better" on some technical point. The best cloud is the one your team knows how to operate.
11. The certifications
| Provider | Fundamental level | Associate level | Expert level |
|---|---|---|---|
| AWS | Cloud Practitioner | Solutions Architect Associate | Solutions Architect Professional |
| Azure | Fundamentals (AZ-900) | Administrator (AZ-104) | Solutions Architect Expert (AZ-305) |
| Google Cloud | Cloud Digital Leader | Associate Cloud Engineer | Professional Cloud Architect |
Path advice: start with the fundamental certification of the provider you use or aim for. It is accessible with this course plus a few weeks of practice, costs about 100 dollars, and brings real credibility on a career-change CV.
The most recognized on the job market remains the AWS Solutions Architect Associate, but it requires several months of serious preparation.
Remember in 30 seconds
- AWS: the widest catalogue, the largest community, the safest default choice.
- Azure: wins through Microsoft integration and bundled discounts — often a commercial decision.
- Google Cloud: the best at data (BigQuery), AI and Kubernetes (GKE, Cloud Run).
- The concepts are transferable: mastering one cloud makes the others accessible in a few weeks.
- European players: Hetzner and Scaleway are significantly cheaper on raw compute, with a narrower catalogue.
- Sovereignty is not just a matter of geography — the company's nationality counts (CLOUD Act).
- Vendor lock-in is not to be fled blindly — it is a trade-off between exit cost and immediate productivity.
- Multi-cloud "as a precaution" is almost always a mistake.
- The most decisive criterion: your team's skills.