If your organisation is evaluating public cloud providers, the field narrows quickly to three dominant players: Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP). Each has a distinct heritage, a different target customer, and real trade-offs worth understanding before you commit a workload. This article walks through all three at a strategic level, covering what each platform is, where it excels, and where it falls short.
The Market Landscape
As of the period this analysis was originally written (mid-2019), AWS held roughly one-third of the public cloud infrastructure market, according to Synergy Research data cited in the original source. Azure held approximately 16 percent and GCP around 8 percent. IBM Cloud and Alibaba Cloud were also growing, and a long tail of Tier 2 and Tier 3 providers—Rackspace, Oracle Cloud, Tencent Cloud, and others—were competing for the remainder.
At a rough, generalised level, the three leaders tend to attract different buyers:
- AWS is used across the widest range of organisations, from start-ups to large enterprises, and across almost every industry vertical.
- Azure is a natural fit for enterprises already running Microsoft software stacks, including Windows Server, Active Directory, SQL Server, and Office 365.
- Google Cloud is particularly popular with developer-centric teams, data engineering workloads, and organisations that are open-source or Kubernetes-first.
These are broad generalisations. The right platform depends on your specific workloads, existing tooling, team expertise, and cost model—not on market share alone.
Pricing Models: An Important Differentiator
One concrete difference between the three platforms—at least as documented at the time of the original article—is how they meter compute time:
- AWS historically billed per hour for EC2 instances, meaning a job that ran for two minutes was billed as a full hour. (AWS later introduced per-second billing for many instance types; verify current billing behaviour for the specific services you use.)
- Azure bills per minute for its virtual machine compute, which can reduce waste for short-lived workloads.
- Google Cloud also bills per minute (with per-second granularity on some products), and applies sustained-use discounts automatically without requiring reserved instance commitments.
Beyond per-unit rates, enterprises should model total cost of ownership including data egress fees, support tiers, reserved capacity discounts, and licensing bundling (particularly relevant for Azure if you hold existing Microsoft Enterprise Agreements).
Amazon Web Services (AWS)
What Is AWS?
Amazon Web Services launched in 2006 and is the commercial cloud computing division of Amazon. It offers a broad portfolio spanning compute (EC2, Lambda), storage (S3, EBS, Glacier), databases (RDS, DynamoDB, Aurora), networking (VPC, CloudFront, Route 53), machine learning, security, and much more. Services are available as IaaS, PaaS, and SaaS offerings and are designed to compose with one another, allowing teams to build scalable, distributed architectures.
AWS Strengths
- First-mover advantage and maturity. AWS launched before any serious competitor and used that head start to build the broadest service catalogue and the largest global network of regions and availability zones.
- Depth of services. No other provider matches the raw number of distinct managed services AWS offers, which is an advantage for teams that want to avoid running their own infrastructure for niche requirements.
- Ecosystem and community. AWS has the largest pool of third-party tooling, training resources, certified professionals, and marketplace offerings.
- Enterprise credibility. Gartner has repeatedly recognised AWS as the most mature enterprise-ready cloud IaaS provider in its Magic Quadrant reports.
AWS Weaknesses
- Cost complexity. AWS pricing is notoriously difficult to model. The large number of services, instance types, and pricing dimensions can lead to bill shock if workloads are not carefully monitored and optimised.
- Data egress costs. Extracting large volumes of data out of AWS storage can be expensive. Teams should understand egress pricing before committing to an architecture that requires frequent large data transfers.
- Service sprawl. The breadth of the catalogue is also a source of cognitive overhead. Choosing between overlapping services requires deep expertise.
Microsoft Azure
What Is Azure?
Microsoft Azure launched in 2010 and is Microsoft’s public cloud platform. It provides IaaS, PaaS, and SaaS services spanning virtual machines, Kubernetes (AKS), serverless functions (Azure Functions), SQL and NoSQL databases, Active Directory integration, DevOps pipelines (Azure DevOps), analytics, and AI services. A defining characteristic of Azure is its tight integration with the rest of the Microsoft ecosystem: Windows Server, Active Directory, SQL Server, SharePoint, Dynamics 365, and Microsoft 365.
Azure Strengths
- Microsoft stack integration. Organisations running Windows Server, Active Directory, SQL Server, or .NET applications on-premises can lift workloads to Azure with less friction than to competing platforms. Hybrid connectivity via Azure Arc and Azure Stack further supports incremental migration.
- Licensing economics. Existing Microsoft Enterprise Agreement customers can apply Azure Hybrid Benefit licences, which can significantly reduce compute costs for Windows and SQL Server workloads.
- Identity and security. Azure Active Directory (now Entra ID) is the industry’s dominant cloud identity platform and integrates natively across Azure services, making it straightforward to enforce role-based access control at scale.
- DevOps tooling. Azure DevOps and GitHub (also owned by Microsoft) offer end-to-end CI/CD, repository management, and project tracking that integrate tightly with Azure deployments.
Azure Weaknesses
- Enterprise support experience. Gartner has noted that despite Microsoft’s long enterprise software history, some Azure customers report that the support, documentation, and partner ecosystem feel less polished than expected.
- Linux and open-source maturity. While Azure has significantly improved its Linux and open-source support over time, teams running predominantly Linux workloads may find AWS or GCP more natural environments.
- Service reliability history. Azure has experienced notable outages, some of which have affected Active Directory and dependent services broadly. Verify current SLA commitments and design for multi-region resilience.
Google Cloud Platform (GCP)
What Is GCP?
Google Cloud Platform launched in 2008 and is Google’s public cloud offering. It provides compute (Compute Engine, GKE, Cloud Run, Cloud Functions), storage (Cloud Storage, Persistent Disk), managed databases (Cloud SQL, Cloud Spanner, Bigtable, Firestore), big data and analytics (BigQuery, Dataflow, Pub/Sub), and AI/ML services (Vertex AI, Vision API, Speech-to-Text). GCP is also the origin of Kubernetes—Google open-sourced the project in 2014—which gives it particular credibility in container orchestration.
GCP Strengths
- Kubernetes and container expertise. Google invented Kubernetes and runs it at scale internally. Google Kubernetes Engine (GKE) is widely regarded as the most feature-complete managed Kubernetes offering.
- Data analytics and ML. BigQuery, Dataflow, and Vertex AI reflect decades of internal tooling that Google built to process its own data at global scale. These services are a strong fit for data engineering and machine learning workloads.
- Network performance. GCP runs on Google’s private global fibre network, which can deliver low-latency data transfer between regions and high throughput for distributed applications.
- Pricing transparency and sustained-use discounts. GCP automatically applies sustained-use discounts when VMs run for a significant portion of the month, without requiring upfront reserved instance commitments.
- Open-source and DevOps alignment. GCP is a popular choice for organisations that standardise on open-source tooling and DevOps-first workflows, partly because it avoids the Microsoft lock-in that Azure implies and offers a different competitive dynamic than AWS.
GCP Weaknesses
- Smaller service catalogue. GCP offers fewer distinct managed services than AWS and, in some categories, fewer configuration options.
- Fewer global regions. At the time of the original article, GCP had fewer data centre regions than AWS or Azure, though it has expanded significantly since then.
- Enterprise sales and support maturity. GCP has historically been seen as less enterprise-sales-focused than AWS or Microsoft. Gartner noted that clients typically chose GCP as a secondary rather than primary strategic provider.
- Product discontinuation history. Google has a documented history of discontinuing products and services. Engineering teams should evaluate the long-term strategic commitment to any GCP service before building a dependency on it.
Choosing the Right Platform: Practical Guidance
No single cloud is best for all workloads. A practical evaluation framework should consider:
- Existing technology stack. Heavy Microsoft workloads (Windows Server, SQL Server, Active Directory) favour Azure. Java, Python, and Linux-native workloads run well on all three; GCP and AWS may require less friction.
- Data and analytics requirements. BigQuery and GCP’s analytics ecosystem are strong differentiators for data-intensive use cases.
- Compliance and sovereignty. Verify which providers hold the certifications relevant to your industry (SOC 2, ISO 27001, FedRAMP, HIPAA BAA, PCI DSS) and whether they have regions in the geographies you need.
- Team expertise. The cloud your engineers already know is often the fastest to deliver value on, regardless of theoretical feature advantages elsewhere.
- Cost modelling. Model total cost across compute, storage, egress, and support for your specific anticipated workload profile—not just headline instance prices.
- Multi-cloud strategy. Many large enterprises use more than one provider. This adds operational complexity but reduces lock-in risk and lets you match workloads to the platform that fits them best.
Summary Comparison
| Dimension | AWS | Azure | Google Cloud |
|---|---|---|---|
| Market position | Largest (≈33% share, 2019) | Second (≈16% share, 2019) | Third (≈8% share, 2019) |
| Launched | 2006 | 2010 | 2008 |
| Service breadth | Widest | Very broad | Narrower but growing |
| Best fit | Broad enterprise and start-up | Microsoft-stack enterprises | DevOps, data, ML teams |
| Compute billing granularity | Per-hour (historically; verify current) | Per-minute | Per-minute / per-second |
| Kubernetes origin | EKS (managed) | AKS (managed) | GKE (origin platform) |
| Key weakness | Cost complexity, egress fees | Support experience, Linux maturity | Smaller catalogue, product continuity risk |
The public cloud market continues to evolve rapidly. Market share figures, service catalogues, pricing structures, and regional availability all change frequently. Treat this comparison as a starting framework and validate specifics against current provider documentation before making architectural decisions.