Comparing cloud services has become a routine part of strategic technology decisions. However, most comparisons start in the wrong place: virtual machine pricing, performance benchmarks, or the number of available services.

The truly relevant question is: which one costs more when the architecture begins to scale?

This article analyzes Amazon Web Services, Microsoft Azure, and Google Cloud Platform from a perspective that normally stays out of sales presentations: data transfer (egress), discount models, Spot instances, price predictability, and billing complexity.

AWS: the leader's strength

AWS is the global leader in market share and service maturity. Its portfolio is the broadest among the three providers, offering hundreds of solutions ranging from basic infrastructure to highly specialized services.

For many, it is the reference standard in public cloud.

ADVANTAGES:

• Greatest diversity of services and integrations.

• Consolidated ecosystem and broad technical community.

• Globally distributed infrastructure.

• High level of granularity in configurations.

For organizations requiring flexibility and technical depth, AWS offers practically every possible architectural combination.

POINTS OF ATTENTION

• High egress fees: data outbound to the internet or other regions generates significant charges.

• Inter-region transfer is paid.

• Complex billing: the billing structure has multiple layers and can make predictability difficult.

• Spot Instances: although they offer aggressive savings, they require architectures resilient to sudden availability interruptions.

In practice, AWS is usually not the cheapest when the application relies heavily on external traffic.

AWS offers robustness and flexibility. However, without a well-structured cost governance, the invoice can grow in a hard-to-predict manner, especially due to networking and data transfer.

AZURE: the power of the Microsoft ecosystem

Azure has established itself as the main competitor to AWS, driven by the Microsoft ecosystem and a strong corporate presence.

Companies already using Windows Server, Active Directory, or Microsoft 365 frequently consider Azure as a natural extension of their infrastructure.

ADVANTAGES

• Native integration with Microsoft products.

• Financial benefits for those who already have Windows licensing.

• Strong performance in hybrid environments.

• Broad global presence.

For organizations with a consolidated Microsoft stack, Azure can deliver significant operational efficiency.

POINTS OF ATTENTION

• Complex licensing model, which can make predictability difficult.

• Not all instance series offer comprehensive Spot versions.

• Billing structure that requires a detailed understanding of contracts and discounts.

• Outside the Microsoft ecosystem, the cost tends to approach that of AWS.

Azure is financially strategic when there is prior integration with the Microsoft universe. Outside of this scenario, the economic advantage decreases and the comparison relies once again on architecture and network usage.

GOOGLE CLOUD: aggressive discounts, but not always the lowest total

Google Cloud has gained ground supported by open source culture, Kubernetes, and data solutions. It is frequently associated with modern workloads and container-oriented architectures.

ADVANTAGES

• Automatic discounts for sustained use.

• Competitive Committed Use Discounts.

• Per-second billing widely applied.

• Strong positioning in analytics and data.

The automatic discount model reduces the need for complex commitments in some scenarios.

POINTS OF ATTENTION

• Recent changes in storage prices show that increases can occur.

• Not always the cheapest in total cost, especially depending on the instance profile.

• Comparisons require attention to equivalent memory and CPU configurations.

Google Cloud reduces the unit cost, but TCO (Total Cost of Ownership) depends on the efficiency of the orchestration.

A relatively transparent pricing structure and competitive discounts. Still, the actual cost depends on the consumption pattern and the intensity of external traffic.

The cost that changes everything

Egress is the charge applied when data leaves the provider's infrastructure, whether to the internet, another region, or another cloud.

Among Amazon Web Services, Microsoft Azure, and Google Cloud Platform, the model is similar: inbound data is generally free; outbound data is charged by transferred volume.

This cost becomes relevant in scenarios such as:

• Public APIs with a high volume of requests.

• SaaS platforms with many downloads.

• Globally distributed applications.

• Replication between regions.

• Multi-cloud strategies.

Traffic-intensive applications can spend more on network than on compute. At scale, a few cents per gigabyte make a significant difference in the monthly budget.

Any strategic cloud decision needs to consider the traffic pattern before looking only at instance pricing.

Long-term discounts

AWS, Azure, and Google Cloud offer discounts for those who make a commitment to use for one or three years. The principle is to have predictability for the provider in exchange for a reduction in unit price.

These models appear as:

• Reserved Instances and Savings Plans (AWS).

• Reserved VM Instances (Azure).

• Committed Use Discounts (Google Cloud).

In stable workloads, savings can be significant. The challenge arises when the environment is not predictable.

Architectural changes, product evolution, or demand variation can make the commitment less efficient than it initially seemed.

The central point is this: a high discount percentage does not guarantee savings. The deciding factor is the stability of consumption.

Spot and Preemptible Instances

Another cost reduction alternative is using the provider's idle capacity. These instances offer deep discounts but can be interrupted at any moment.

The model is mainly recommended for:

• Batch processing.

• Testing environments.

• Data pipelines.

• Non-critical tasks.

Discounts can reach up to 80% or 90% of the on-demand price. However, they require resilient architecture and constant monitoring.

There is also a relevant operational difference: AWS presents greater dynamic price variation, while Azure and Google Cloud tend to demonstrate greater relative stability.

Governance and FinOps

Even when using the same provider, two companies can have completely different invoices. The difference is usually in management.

Much of the waste in cloud comes from internal decisions, such as:

• Oversized resources.

• Forgotten active environments.

• Lack of automatic shutdown.

• Absence of monitoring by team or product.

• Lack of periodic architectural reviews.

Cloud is a variable expense model. This requires continuous tracking and alignment between technology and finance.

In practice, this level of control rarely happens entirely manually. That is why many companies structure dedicated FinOps operations combined with continuous infrastructure management.

An example is CloudOps Fly, a CodeBit service that operates directly in the 24/7 management of AWS environments. The proposal is to centralize monitoring, optimization, and governance in a continuous operational layer, reducing waste and keeping the infrastructure aligned with actual usage patterns.

This type of approach allows identifying inefficiencies in real time, applying recurring adjustments, and preventing invisible costs from accumulating over time.

In addition to technical optimization, the model includes direct financial benefits, such as cashback applied to cloud consumption. In this format, part of the amount spent returns to the company on its AWS invoice, reducing the effective cost of the operation and expanding the impact of FinOps practices.

Learn more: The master strategy to cut up to 90% of your AWS costs

Without governance, any cloud becomes expensive. With optimization discipline, differences between providers become strategic rather than emergency-driven.

Where the difference really matters

Comparing Amazon Web Services, Microsoft Azure, and Google Cloud Platform requires looking beyond the virtual machine price lists.

The real difference usually appears in three main dimensions:

• Data transfer and external traffic.

• Long-term commitments and contractual flexibility.

• Internal maturity of governance and FinOps.

AWS tends to require more attention regarding billing and network complexity.

Azure is financially strategic for those already within the Microsoft ecosystem.

Google Cloud offers competitive discounts and relatively predictable models, but this does not guarantee the lowest total cost.

In the end, the cost is not determined solely by the chosen provider, but rather by the combination of architecture, usage patterns, and management capability.

And it is in this combination that the difference really shows.